'The UK has election betting and none of these negative outcomes have come to pass.'
It does, but this point actually counts in favour of 'elections are sport'. Election betting in the UK mostly happens through regular bookmakers who just do their own research and offer odds to punters, there's very little in the way of explicit prediction 'markets' in the way they function on e.g. PredictIt, and often elections will be one of very very few non-sporting events these bookmakers offer odds on. It's rather amusing - there'll be a drop down menu that goes something like; Football - Cricket - Rugby - Snooker - Racing - Motorsport - Politics. If you didn't know what politics were you'd definitely think it was just another sport. Indeed, if you go to the bottom of the politics page on William Hill, at the bottom of the featured odds it still says 'See All Today's Matches'. In Betfred there's a section called 'Rishi Sunak Specials'.
This also means the regulatory problem is totally void, since it's just up the bookmaker's discretion as to whether they pay out, as is the case with all types of sports betting.
Yes but in the case of disputed outcomes the discretion afforded to bookmakers is pretty wide no? Especially since they of course write the resolution terms so can gives themselves a lot of room. I know there's dispute resolution like IBAS but again I think it's non-binding no?
Betfair and Smarkets are real prediction markets (i.e., odds are not determined by bookies) in the UK. Though I don't know how much market share they have.
Betfair has a sportsbook business and an exchange, where the exchange makes cash via a cut (5-20% of profit depending on whether you're a frequent flyer). The exchange is the largest in the world. Flutter (the PaddyPower/Betfair parent) is the no. 2 UK betting company, and the exchange is probably 5-10% Flutter's business by revenue? I'd be very surprised if BFEx was more than say 5% of UK market share given Bet365's size (incidentally a rare British world leader). Probably much lower than that given bricks and mortar retailers too.
Having said that, I think Betfair Exchange is the single largest betting exchange in the world- but demand just isn't as large outside of the smart money.
I wonder if it is not best for prediction markets to lean into the gambling space and out of the financial contract space. Sports gambling is starting to be legal in a lot of states and it would seem easy to just move event contracts to something like Bovada.
"I think their thought process is: if you manipulate the commodity markets by (for example) saying that you have lots of nickel when you don’t, the CFTC has to investigate that and penalize the people responsible. In a hypothetical world with election contracts, if someone manipulated an election - for example, they put out a fake poll showing that the incumbent would definitely win so there’s no point in even voting - someone could ask the CFTC to investigate."
Is their thought process not "the outcome of the last US Presidential election was contested and extremely controversial, we would prefer not to get involved in establishing who won it"?
I think it's also tied to things like Citizen's United (the actual case not the boogeyman) where they might have to rule if various campaign messaging is market manipulation.
I absolutely don't think corporations are piling in on the election to hedge risk - there just isn't that much money in the markets. Can't immediately find a figure but here is oddschecker boasting in October 2020 that one bookmaker had already done $182m on the US election. Not a heap of cash compared to 23bn bet on this year's superbowl. (I staked £1000 to win 600 on Biden)
As far as fake polls are concerned pollsters are tightly regulated by statute and the trade body the UK Polling Association. I don't see an issue here
The other thing is it's relatively easy to make money in UK political betting because it's a small and inefficient market. My Biden bet was so big by my standards I am definitely in profit overall
Politicalbetting.com is the prime discussion forum, for those in jurisdictions which don't block access to it
On skybet in the UK a little while back, it was possible to bet on Joe at 2/1 and Trump at 10/11. Checking now, it's come down to much more reasonable levels - you can still guarantee a small amount of money if at least one of them wins, but the difference now is probably a reasonable reflection of the chance that neither are able to take office, whereas 2/1 and 10/11 was not.
> Because real money is involved, Manifold will have to tighten the rules on markets, including banning N/A resolutions.
What?? I already have markets of the form "Conditional on X happening, will Y also happen?". This means that I resolve YES if X and Y happen, NO if X happens and Y doesn't, and N/A if X doesn't happen. This is a very convenient way to try to get an estimation of some conditional probability. Moreover, it's a promise I already made to the traders when I set up the market rules like that. What the hell am I going to do?
So I understand correctly that I won't be able to resolve them manually? Each case would require mods' intervention?
I probably need to warn the traders on my questions about this...
Otherwise, it sounds like a good way to estimate conditional probability will disappear, and this has been an important use case of manifold. ("If I read this-and-this book, will I like it?" etc.)
- We plan to significantly tighten rules for market resolution to eliminate abuse. Expect more rule changes.
- Creators will no longer be able to resolve markets N/A in the typical case.
- This change is to limit abuse that comes from traders cashing out profits that later get clawed back due to N/A’s.
- We will probably allow N/A resolutions in the first few minutes after market creation to handle typos or other cases where the user wants to remove the market.
- Conditional markets (”If A then B”) can be implemented as a multiple choice market with each options for A and B enumerated (”1. A and B, 2. not A and B, 3. A and not B, 4. not A and not B”). We intend to create a first-class market type for conditional/decision markets
Maybe I just have very different intuitions than everyone else, but I still find it super weird and bad that it's even possible to ban prediction markets at the federal level without congress passing a law. I get that it's technically legal, but it feels to me like some kind of dictatorship-by-bureaucracy where people that nobody voted for are able to effectively pass federal laws by decree, and we don't have any power to vote them out.
Another example is how last month the FTC issued a federal ban on noncompete clauses in contracts.
A lot of people were celebrating this and I think that it's probably a good rule to have, but it struck me as bizarre that a federal agency can just make up new laws by decree, and it seemed like this same power could be used to make bad laws. Well now with this prediction market news, I consider that worry to be vindicated.
1. Congress previously passed a law that said "agency: please do research and hire experts, and then make rules in support of goals X, Y, and Z"
2. If Congress doesn't like a decision the agency makes, they have [90 working days, or something like that] to veto the regulation. Or they can always pass a new law that says "sorry, the rule is actually B, not A." With or without changes to X, Y, and Z.
That's a good point (same goes for Dan L who made a similar point), but it still doesn't sit right with me. Like I get that they have the legal authority to do it, but I'm still against it anyway and think that the way things work should be changed.
Congress absolutely *could* step in and pass a law here either mandating or forbidding a ban, it's in the absence of action that agencies fill the vacuum. It's important to distinguish between "we have no power to do X" and "X will predictably not be done by those with power" - call it a failure of niche regulation, where anything outside of the dozen most important issues fails to produce a strong enough electoral signal and isn't worth an elected's attention.
Congress in particular seems to be quite risk-adverse, refusing to even confirm good rulemaking via statute. IMO, that's the result of an *increased* responsiveness to public opinion combined with an asymmetric public interest in punishing bad v. rewarding good, but that's a tangent.
generally if Congress were going to handle this, they'd have to massively expand staff budgets to hire a bunch of experts to constantly evaluate and handle all the decisions, which would be very unpopular in itself. Regulatory decisions would get also delayed constantly due to the Congressional calendar.
And ultimately the actual representatives would just wind up deferring to a different set of unaccountable experts, maintaining only nominal veto power, only now they work out of a different office building in DC than the current CFTC.
> it feels to me like some kind of dictatorship-by-bureaucracy where people that nobody voted for are able to effectively pass federal laws by decree, and we don't have any power to vote them out.
It's better than that. The officials above them in the chain of command can't replace them (or just fire them) either.
Everyone finds the government undemocratic when it does things they don't like. I doubt they care about the "deep state" when it's doing things they like though.
I commissioned the Good Judgment study and the poll of virologists and epidemiologists on covid origins. The resources we shared with the forecasters are listed in the full report and the supporting document which has all of the superforecaster deliberations.
The resources we shared were the rootclaim debate, which has a lab leak proponent and a zoonosis proponent
On the zoonosis side, worobey 2022, Holmes 2021, Pekar 2021, and Andersen 2020.
And on the lab leak side, stoyan and chiu 2024, the bayesian analysis by Michael Weissman, the bayesian analysis by demaneuf and de maistre, and demaneuf's presentation to SAGO on case reporting early in the pandemic In Wuhan
So reading your report the only data that would make your "superforecasters" change their mind would be a published written confession by the CCP that they ordered covid and released it. I assume they still deny the holocaust and holodmer under those standards.
Just to note that the Swift Centre's forecasts for coal consumption still expect consumption to peak and decline, just on a delayed timeline. They have China peaking by the early 2030s. From that link it seems ambiguous what peak timing they expect for India, except that they'll be slower.
"Trying to weasel out of this by classifying elections as a sport is pathetic"
This strikes me as really quite uncharitable (quite uncharacteristically for Scott, as both a generally nice person and also as author of "The categories were made for man, not man for the categories"..)
It seems to be entirely reasonable to say that elections aren't a sport if you *just* ask "does this activity have a bunch of people dressed exactly the same running and jumping and throwing things?" - but it does seem very much like one if you ask the question that regulators are supposed to care about, viz. "is this activity the sort of thing a gambling addict is vulnerable to becoming obsessed by and ruining their life over?"
People become ruinously obsessed with sports betting because sports are supremely prevalent and in-your-face in media/news/advertising, are hotly contested, generate strong passions, and are talked-about by everybody, everywhere, all the time (with the possible exception of at ACX meetups..) These properties seem to be very closely matched by politics - that's the dimension regulators ought to care about, not the level of athleticism involved.
Gamblers could get fixated on whether the groundhog sees its shadow, but that wouldn't make it a sport in the sense people normally understand it (unlike the whales from Scott's post you cite). If there's legislation about sports specifically, then it needs to apply to sports. If the regulator is supposed to be restricting anything gamblers could get fixated on, then the legislation would have needed to say something like that instead.
To make explicit the analogy to Scott's earlier post: People normally understand that whales are mammals, but King Solomon, who only cares about a subset of the properties of whales, still needs it to be within the purview of the Ministry of Behemah, not the Ministry of Dag. Similarly, people normally understand that politics isn't a sport, but the Financial Conduct Authority (or whoever), who only care about a subset of the properties of politics, still needs it to be within the purview of sports gambling regulation. I'm not necessarily arguing in favour of doing this (personally I think the US overregulates prediction markets, and mostly does so for reasons unrelated to -and less reasonable than- their risk to gamblers) - I'm just saying that it's a perfectly reasonable argument to make, not a "pathetic and weaselly" one.
I'm not suggesting that regulators are supposed to regulate every possible thing gamblers could become fixated on - I'm suggesting that since we have an existing refusal-by-default category that includes sports it would be reasonable to argue that politics should also be in that category, and somewhat uncharitable to claim that regulators are "pathetic" and "weaselling out" by so arguing.
(Perhaps the solution would be to say "That regulatory category we used to call 'sports'? Well now it's called 'high addiction capability' and it includes both sports and politics". Presumably one can't just arbitrarily redefine existing regulatory categories like that, so I think it's plausible that the path of least resistance may have been just to say "For the purposes of this legislation, politics and sports look the same along dimensions we care about, therefore, fine, we're going to call politics a sport". As HaroldWilson (presumably not *that* one...) points out elsewhere in the comments, UK bookies already do something much like this in their internal systems.)
No, people in Solomon's kingdom don't understand that what we in the 21st century call "whales" (in a language that hasn't yet come into existence in his time) are "mammals" (another term that doesn't exist, as they don't group animals by whether they have mammaries). If he announced laws in the language of 21st century English, they would not be understood in his kingdom. If you wanted a less distant example, you could cite SCOTUS declaring that tomatoes are a vegetable (under the culinary definition) rather than a fruit (under the botanist's definition).
> Perhaps the solution would be to say "That regulatory category we used to call 'sports'? Well now it's called 'high addiction capability' and it includes both sports and politics"
Actually changing the language of the law to fit their extension would be less weaselly than claiming the law covers it via a term nobody thinks includes politics other contexts.
Even if you put King Solomon and everybody in Biblical Judah through a biology PhD programme* and lectured them in excruciating detail on the natures of whales, fish, and mammals, they would still be better served by calling whales a fish rather than a mammal if that's what it took to get whale-management under the purview of people who live on the coast and go around in boats.
(* How's that for a creative writing prompt..)
Not heard of the tomato thing before but I agree it's an excellent point: tomatoes are a fruit and are commonly acknowleged as such - but it's perfectly reasonable, and certainly not pathetic or weaselly, for some culinary authority to say "We're going to call tomatoes a vegetable for our specific purpose."
I agree that changing the legal definition of the category to point to the thing you actually care about (instead of its pointing to one specific example of that thing) would probably be preferable, but I don't at all agree that the reason it would be preferable is that it would be less weaselly.
The important thing, for both the kingdom of Solomon and SCOTUS, is that people were already using such words to refer to whales and tomatoes before the rule was promulgated. Other people can use different language, because language is to a certain extent arbitrary (hence the many different words in different languages to refer to the same concept), but the language used for the laws had commonly understood meanings, and indeed they wouldn't change their minds if told their language was "incorrect" based on someone else's reasoning for categorization.
Whales are indeed mammals, but you're not going to get very far in whale management by sticking one in a field and encouraging it to graze on the lovely pastureland.
The categories here are actually legal categories written by Congress. This isn’t just semantics, the real question should either be what Congress intended when they wrote the law, or what the original public meaning of the text meant when it was passed, depending on your legal theory. This feels less like one of those and more bureaucrats wanting to avoid responsibility. But that’s why PredictIt has their lawsuit for a Judge to figure out.
I agree that what Congress intended is an important question, yes. But I have 4 objections to the broader point:
1) I think "they intended to prevent certain harms that can arise from gambling, except that when they passed the law these harms were mostly restricted to sports betting because back then prediction markets were merely a twinkle in Robin Hanson's eye, therefore treating politics betting similarly to sports betting would be consistent with Congress' original intent" is a perfectly fair and reasonable answer to that important question.
2) I don't think it's the *only* important question; I think it's also fair and reasonable to want to evaluate competing interpretations of legislation based on the actual outcomes of each interpretation as well as (or instead of) channelling the spirits of the original legislators.
3) In British and international maritime law (my own area, sort-of) many set-in-stone legal definitions, both modern and centuries-old, are *incredibly* vague, often to the point of uselessness, and open to broad and possibly even contradictory interpretation - I would be awfully surprised to learn that this isn't also true of US law!
4) *Are* there even strict Congress-mandated legal category boundaries, here? If some institution is allowed to categorise politics as a sport (or for that matter the tomato as a vegetable..) the legal category boundaries we're working from can't be all that strict, can they?
Isn't the problem with sports betting that the competitors can bet on their opponents and throw the match? I have a hard time imagining this happening in politics, although I suppose a candidate in financial trouble could make a big bet against themselves just before admitting to a scandal or otherwise sinking their chances.
If we're talking about the giant defence firms etc. having millions to throw around and vested interests in the outcomes of elections, I see no reason why they wouldn't try and interfere *if* prediction markets become what it is hoped they will become, indicators of the most likely outcome.
Personally, my view is that if you open them up to the public and let them run forecasting markets on all topics, they will become betting sites not forecasting, and results to set policy will be as effective as the "stick a pin in the list of runners" way to select a horse to bet on.
But suppose they function as hoped, and the result is seen as "this is going to happen with a very high degree of certainty" which everyone accepts.
Okay, now it's worth my while to try and tilt the result to read "Senator Smith (who is likely to steer juicy government contracts my defence contractor firm's way) is going to win this election" rather than "Senator Jones (who is running on a platform of cutting such spending) is going to win". I want to get people thinking "Okay, if SurePredict is saying Smith's the winner then it's no use my turning out to vote for Jones". I want a result mediated by the prediction market that will be the same as California being so solidly Blue, you could run a bag of flour as a Democratic candidate and get it elected before a Republican.
If it stays as sports betting, then it's not worth my while to mess around with it.
I can see why the CDC doesn't want that kind of betting and doesn't want to get involved, and I'm more sympathetic to their position than "prediction markets are the way to set policy because something something rationalists something something wisdom of crowds".
What you want is to get just enough people to decide it's no use voting for Jones so that Smith wins.
How many comments over the years have we had on here that "your vote doesn't count, here's my maths working out to demonstrate that's true, so don't bother voting"? Scale that up to manipulate a prediction market outcome and you'll get the thousand people saying "well it's pointless me voting" that permits Smith to win.
But your vote really doesn't matter in practice relative to the election result. What matters is whether enough people believe that it's their "civic duty" or some such to turn up, and here all sorts of propaganda indeed can have disproportionate impact.
And that big bet against themselves would... Move the market in the right direction, hence making private information public and performing exactly the role prediction markets are meant to.
there was a minor scandal with the head of a political consulting firm who did some polling, was also betting on elections. From his point of view, he was just doing the good forecaster thing of putting skin in the game.
From most peoples' point of view, he could have easily made money by manipulating his polls.
Given there have sometimes been ambiguities in what people might count as a lab leak vs "natural" zoonosis, worth highlighting that Good Judgment's definition of a "biomedical research-related accident" states:
"...that SARS-CoV-2 emerged in human populations as the result of the accidental infection of a laboratory worker with a natural coronavirus; or the accidental infection of researchers with a natural coronavirus during biomedical fieldwork; or the accidental infection of a laboratory worker with an engineered coronavirus; “research” includes civilian biomedical, biodefense, and bioweapons research."
"The UK has election betting and none of these negative outcomes have come to pass."
Yes, and these bets are handled by bookmaking firms, not prediction markets. This means that:
(1) They are considered games, like betting on who is going to win a golf, soccer, racing or other sport tournament
(2) The bookies have robust anti-fraud/anti-rigging the market protections in place, because surprise surprise, people have previously attempted (and gotten away with) betting coups:
Where money is involved, people have always tried to cheat, find loopholes, commit fraud, or rig the system.
"In 2010, Curley decided to try and emulate his feats of 1975. With the far more sophisticated world of modern bookmakers, it seemed a much greater challenge, yet he succeeded once again. Modern technology means there’s no way you could pull off a “Yellow Sam” kind of coup nowadays. So, 35 years on, Barney began to look for ways of outwitting the bookies once more. He is a technophobe but he recruited a young whizz kid who found the loophole they needed.
He worked out that while betting shops are vigilant for small single bets being placed on one horse they don’t look out for multiple bets where the punter bets on winners of more than one race.
It goes against all bookmaker trading experience for a strongly fancied horse not to be backed as a single. Why would anyone jeopardise a payout by including it in a double or a treble with other horses? It just doesn’t happen.
Now based in Newmarket in the UK, where at one stage he had 50 horses in training, Curley set himself the task of matching four runners to four races on the same day which stood a decent chance of winning at good odds. It was a massive undertaking, as it is hard enough to send out one winner with a degree of certainty.
He also had to do it behind a veil of complete secrecy, given that the bookies were particularly vigilant for anything involving horses from Curley’s stable.
Via another accomplice, former City trader Jack Lynch, he hired a gang of 30 “putters-on”. Armed with fold-up bikes and untraceable mobile phones, they were given carefully planned routes to cover as many betting shops as possible in a day. It was also crucial that while they moved between them as fast as possible, they dawdled once there. They had to look natural so as not to attract attention.
The plans changed many times before Curley fixed on four horses and races on May 10th 2010. The horses were Agapanthus in the 4.10pm at Brighton, Savaronola in the 5pm at Wolverhampton, Sommersturm in the 5.30pm at Wolverhampton and Jeu De Roseau in the 7.30pm at Towcester. Only Agapanthus had won a race of any kind in Britain and the extent of the gamble was made clear when Savaronola went lame two days before. But Curley decided to go ahead anyway. By late morning on the day they were respectively priced at 11/2, 4/1, 9/4 and 25/1. These odds were sure to dive. But by then, the bets would be safely on.
Around £100,000 worth of bets were laid in various multiples, keeping the stakes as low as possible in order to stay under the radar. The bets were trebles on three of the four horses, in different permutations, and Yankees, a multiple bet on four horses in which at least two horses had to win for any payout at all. It seemed like madness but that was part of the point: to make sure the bookies’ early-warning systems weren’t triggered.
The team returned to base to watch the races on TV. Their first horse Agapanthus was by now hot favourite, but it had the gang worried by starting slowly before coming through to win by two lengths. In the next race, Savaronola was up against Alex Ferguson’s If I Had Him and was 11/10 favourite – it notched up its first ever win, finishing six lengths clear. Apparently, Ferguson was fuming, later telling Curley: “I backed that bloody horse that day!”
But disaster struck in the third leg of the bet. Sommersturm had been seen as the certainty of the foursome, but trailed in fifth. This was not in the script.
All hopes were now pinned on Jeu De Roseau. Rival Jockey Tony McCoy set a hot early pace on his mount Manjam, but Jeu De Roseau’s rider Denis O’Regan bided his time and pushed the exhausted horse into a winning lead in the final furlong. The banker had failed to deliver but the outsider came to the rescue!
If Sommersturm had come first, the payout would have been more than £15million. However, three out of four meant they had netted £3.9million, reportedly the biggest win in history."
Surely that just depends on how much you wager on them? In that sense, an election is a sport I'm forced to bet on, to the tune of several percent of my income.
I should have added “except to the bettors”: if you don’t bet on sports you can safely assume the outcome of a sports match will have essentially zero effect on your life, which is completely untrue of elections.
Nope. A city's team advancing to the playoffs generates a financial windfall for the city, especially for hotels. And winning the final sometimes sets off celebratory rioting in the streets.
This feels hazardous as theres always some exception to any attempt to define sport.
But relevant here i think is that the ultimate purpose of sport is to win, and who wins is what matters. It is only the event.
Elections are designed to determine who governs and what they do while governing is what matters. Ppl have opinions about what they should do. The event is just a precursor.
1. The purpose is different. Sport is, by definition, meant to produce amusement. Elections are meant to assign roles to people.
2. Sport is an activity, and elections are fundamentally defined by concerns other than the activity itself. If you get together a group of your friends and do the same things that politicians do while running for election, it will nevertheless be true that they're running for election and you aren't. In contrast, if you get together a group of your friends and do the same things that NFL teams do while playing against each other, you will be playing football just as much as they are.
Do you believe there's an essential characteristic of sport that elections 𝘴𝘩𝘢𝘳𝘦?
> Do you believe there's an essential characteristic of sport that elections 𝘴𝘩𝘢𝘳𝘦?
they form a natural category as simulacra of war. that's why fans act fanatically about them.
- sports simulates war. like porn, it reaps the pros without the cons.
- politics emulates war. in a democracy, the outcome of battle is approximated via poll (headcount), rather than reckoned ex post facto via pollaxe. the victor gets to decide policy for the polis.
> politics emulates war. in a democracy, the outcome of battle is approximated via poll (headcount), rather than reckoned ex post facto via pollaxe. the victor gets to decide policy for the polis.
Bret Devereaux had a good piece on ancient Greek democracy. They would have recognized your description here.
A democracy that reached the stage you identify ("stasis", in the Greek terminology, but in a sense unrelated to the English sense of the word) was considered to be on the verge of transitioning to some other form of government. This isn't a characteristic of elections, except in that it's characteristic of the way they die.
So, (a) elections do not actually simulate war except to the extent that they have winners and losers. (b) simulating war, even to the extent of having losers, is not an essential characteristic of sport. Consider bungee jumping.
I'm not saying that politics is inherently uncivil. My distinction between "simulate" vs "emulate" was deliberate. In this context, a simulation has isomorphic mechanics, but different stakes. Whereas an emulation has different mechanics, but identical stakes.
"politics *emulates* war" in the sense that politics is a lower rung in the hierarchy of consensus mechanisms. It accomplishes the goal of determining policy, but with the advantage that the mechanisms are cheaper to implement (viz. less money and bloodshed). Nonetheless, I do believe that the distribution of political power often reflects the distribution of military power, even when negotiations are ostensibly civil. (Pop quiz: why does the U.S. have more political capital than say... San Marino? In large part, because the U.S. has aircraft carriers and San Marino does not.)
"Sports simulates war" in the sense that the stakes are qualitatively different (e.g. the losers in sports are not eradicated or enslaved). But the mechanics are similar *enough* to the real thing as to produce a roughly-coinciding subjective-experience. E.g. sports promotes camaraderie, athleticism, grit, etc. This ties into a broader theory of mine that all things fun are training simulations.
> (b) simulating war, even to the extent of having losers, is not an essential characteristic of sport. Consider bungee jumping.
I don't even remotely consider bungee jumping a sport. Odd counterexample, imo.
> In this context, a simulation has isomorphic mechanics, but different stakes. Whereas an emulation has different mechanics, but identical stakes.
You can hardly expect other people to understand a meaning that you just invented. What you want is some variety of "substitute (for)".
The rooting interest that other people identify in politics comes from its mechanical similarities to warfare, such as they are. Nobody cares even the slightest bit about "consensus mechanisms" in general.
People do care about outcomes, often, but their attitude toward politics is completely unlike their attitude toward sports, except in cases where it is widely agreed that the politics have failed. (This can apparently happen in either direction; you have the odd Byzantine case where political views were determined by sports team allegiance.)
> I'm not saying that politics is inherently uncivil.
I don't see what this is responding to; it doesn't seem to be related to anything I wrote.
> I don't even remotely consider bungee jumping a sport. Odd counterexample, imo.
Well, again, if you're going to just invent your own meanings for words, you won't be able to talk to other people.
> What you want is some variety of "substitute (for)".
Emulations generally substitute for the genuine thing, no? E.g. Dolphin, a videogame emulator, often substitutes for a genuine Gamecube. Margarine, which emulates the taste of butter, often substitutes for genuine butter. Care to produce a counterexample?
>> I'm not saying that politics is inherently uncivil.
> I don't see what this is responding to; it doesn't seem to be related to anything I wrote.
From context, I figured it was safe to assume this is what you meant (at least approximately) by refering to the Devereaux article. If not, I don't know what you were complaining about, regarding "This isn't a characteristic of elections, except in that it's characteristic of the way they die." Maybe you'd prefer "fanatical" or "irrational"? Although I used "fanatically" in my inital comment, and you still found issue with that.
> The rooting interest that other people identify in politics comes from its mechanical similarities to warfare, such as they are. Nobody cares even the slightest bit about "consensus mechanisms" in general.
Maybe not the idea of consensus mechanisms in the abstract. But people do often care about things which fall into that category, when it determines outcomes they have vested interests in. E.g. do you think plaintiffs and defendents hire lawyers just for fun? That said, I only bring it up to highlight the similarity in that they determine outcomes of importance, rather than the similarity between the internal mechanisms. Which is why I think "emulate" is more apt.
I agree that warfare, politics, and sports share mechanical similarities insofar as network-effects drive fanaticism. Despite this, "simulation" doesn't seem like the most apt word to apply to the relationship between politics and war, since politics approximates the phenomenological aspects of warfare to a lesser degree than sports. E.g. athleticism is neither considered nor promoted as a core aspect of politics.
> People do care about outcomes, often, but their attitude toward politics is completely unlike their attitude toward sports, except in cases where it is widely agreed that the politics have failed. (This can apparently happen in either direction; you have the odd Byzantine case where political views were determined by sports team allegiance.)
Given my earlier misinterpretation regarding "fanatical"/"uncivil", I'm not confident I understand this part either. For the sake of clarity, can you further specify which attitutes you're referring to? And/or maybe link to the Devereaux article? I googled it myself earlier, found this article [0], and figured "uncivil" was sufficiently synonmyous with "factionalism" in this context. Not perfectly synonymous, mind you. But equivalent enough for "uncivil" to serve as a euphemism.
(Assuming we *are* discussing factionalism,) sure, the Greeks had a neat distinction between good democracy vs bad democracy. But in reality, I think actual polities often lie somewhere on the middle of the spectrum. And thus, I think "sports-team allegiance" in politics is more common than you imply. Or, dare I say, maybe even the norm outside the Anglosphere.
> I can only refer you to the "modern sport" section of wikipedia's page on bungee jumping
Were you to ask 10 randos on the street to list all the sports they could think of, I'd hardly expect any of them to include bungee jumping. Wikipedia's copium be damned. You might as well put airplanes in the category of birds.
There have been a surprising number (more than two nickels) of instances where I've found myself in a debate over whether something is a "sport", and the criteria I've refined are that the activity in question must (1) be athletic, (2) have objective criteria for deciding a winner, and (3) include interaction between the competitors.
Gymnastics? Yes on (1), no on (2) & (3): not a sport.
Golf? Yes(ish) on (1), yes on (2), no on (3): not a sport.
> It looks like nobody expects jail time in any case.
Could this be about jail time from contempt of court if Trump keeps talking despite the gag order? The market question is for at least 30 days in jail; I have no idea if that's a realistically achievable amount of time to get from repeatedly violating a gag order.
Wow- I once worked for a US company that was shut down by the police for exploiting the sweepstakes loophole.
The company founders included a couple of lawyers with decades of experience in the gambling industry. They had these long company-wide legal compliance meetings twice a week, where every aspect of the company, from IT to customer support, was optimized to be as consistent with the letter of sweepstakes law as possible. The company had been fighting legal battles over their business model for years, and had actually won a string of cases in several states. At one point, they switched over from regular sweepstakes to charity sweepstakes under the theory that the courts were slightly more friendly to the latter. They were all extremely confident that they could beat any legal challenge.
Then, one day, I arrived at work to find the police loading office equipment into vans. When I asked what was happening, they led me to the company lobby, where all of the employees who hadn't immediately turned their cars around upon seeing the cops were waiting to be individually interrogated in a confiscated accountant's office. The police left after a couple of days- taking with them my personal laptop, which I never got back- and what followed was a week of showing up for half-days of "work" to a gutted office building where executives gave impassioned speeches about how proud they were of the company we'd built, and about how we'd followed the letter of the law perfectly and would definitely get our accounts unfrozen and be back in business soon. Then, a couple of those executives were arrested and the company was dissolved.
Apparently, what had happened was a local newspaper had written a hitpiece about the company which called out the DA by name for allowing such a degenerate law-skirting gambling operation to take root under their nose- a story which got picked up by a bunch of other papers. So, the DA got a judge to interpret the regulations in an entirely new way that our lawyers hadn't anticipated, and pressured the police to make an example of the company. Turns the legal system isn't like a computer that you can hack with the right exploit- if someone with power feels that you're skirting their authority, they will find a legal avenue to regain that authority, loophole be damned.
So, hearing that Metaculus is trying for the same legal play- and probably without the experience and advantages my old employers had- is more than a little concerning.
Yeah, IMO if you're going to do something like this, your best bet is to setup an office in the UAE (or something like that) and be prepared to never set foot in the US should s*** hit the fan. This is pretty much Durov's (Telegram) strategy and it worked out well for him so far. I wish the best of luck to Manifold and I'd hate to see any of them come into trouble but unfortunately I don't get to call the shots when it comes to law enforcement.
"You can't hack the law" is actually a real saying these days; unlike a computer executing code, judges will sit and listen to your argument, smile and nod, and then tell you the law means what it was intended to mean instead of what it literally says.
Which just means that you actually need to hack judges. Computer nerds don't seem to be up to the task yet, but I understand that work in this direction is ongoing.
Forum-shopping is a promising workaround, for those on the offense. Even if hacking a specific judge is infeasible, it can be possible to select a location where the chain of trial judges and appellate judges is favorable.
And yeah, at some point it should be possible to load up an AI with decisions and transcripts and court video, and identify what factors a particular judge would see as most persuasive.
On the relative lack of success of prediction markets: I think a lot of the optimism was justified by the idea that they would be profitable for people who are right, so act as a "tax on bullshit" and incentivize professionals, financial firms, etc to bet in them.
The issue has been that, once you start looking at it from the lens of a professional investor, there are lots of more cost effective things to do with your money and resources. You can spend ages researching election outcomes and have a worse expected return than just tossing it in index funds. There's presumably also other ways that people operating at that level can implicitly bet on election results as well (shorting/investing relevant stocks) so its not even the optimum way to profit from having unusually good election analysis.
So prediction markets only really appeal to people who are doing it for reasons other than it being an optimal use of time/money.
Another problem: you can't force someone to bet against you. I.e. if Joe Rogan predicts that aliens will land on the White House lawn in 2026, there's no way to force him to actually place a financial bet on said belief
This is why I have insane respect for the Rootclaim guy: he actually paid out $100k instead of ranting on Twitter and paying nothing.
yeah, prediction markets are only a tax on bullshit if bullshitters enter the market... and why would they, if they know they're slinging bullshit?
If the only people in the market are level-headed dedicated predictors, then the only money you can make is being better than them, or random chance. Even if you're better than the average good predictor on the market, that may be a marginal thing with low returns compared to index funds... and the worst rational predictors will consistently lose money, encouraging them to leave, lowering profits more.
Well, during the "stolen election" embarrassment four years ago there were all sorts of betting markets obviously biased towards Trump, but trying to exploit them still wasn't a good way to make profit, so that was a clear demonstration that the market doesn't punish bullshiters enough.
Weren't they markets mostly using play money or hundreds/thousands of dollars?
I think things look a lot different if there are billions of dollars in the market. Which is not a near-term concern, but as I understand the movement is absolutely something that advocates would be happy to see in the long term.
Heck, even in the real, highly liquid financial markets, it is possible for dumb money to overwhelm smart money by sheer numbers, as we saw with the various memestocks in recent years. And yes the bullshitters did lose a lot of money and the professionals made a lot of money, but that didn't stop it from happening.
Manifold's buckets on "When will the 2023 Israel Hamas war end?" are not what I would have chosen, since they have Oct/Dec 2024 as a single bucket, but the question of "Before or after the US election" seems pretty significant.
>The UK has election betting and none of these negative outcomes have come to pass.
What makes you say that they haven't?
I admittedly have a limited knowledge of UK electoral politics, but from what I hear it's been an absolute mess recently, with lots of lies and manipulation of various types. Of course elections themselves encourage those things so it's hard to pin down an exact cause, but saying 'and none of it at all was motivated or encouraged by election betting' seems like a hard claim to prove.
Just assuming that none of it was motivated by that because you don't think betting markets motivate those things in general is obviously begging the question, of course.
Maybe you could say 'I expect that if there was manipulation motivated by a betting market it would have been discovered, and I expect that if it were discovered I would have heard about it, so absence of evidence is evidence of absence.'
Which is correct in terms of Bayesian updates, but that's two levels of imperfect and unmodeled filters on the evidence, so it shouldn't be able to make a claim as strong as 'this has not happened'.
As a more general point: I would be a lot less worried about betting markets if advocates said 'bad things could happen, we've thought about it a lot and here are some of the things that could happen and our ideas for safeguarding against it, but despite the risks we think they're still a net public good of great importance.'
But it feels like advocates generally take the more pat answer that *implies* without 100% directly stating: 'nothing bad has happened, nothing bad could happen, any manipulation is dumb money for the market to take.' This reminds me very much of libertarians saying the hand of the market will fix everything if we just stop regulating it, or people saying AI will be so smart it will just use common sense to not do anything evil.
The world is more complex than that, bad things can happen even if your system is good or they theoretically shouldn't happen in an 'ideal' scenario. The denials of that make me anxious about the whole project, even though I really want to love it.
>The idea that adding one more group - “people who have a few thousand dollars staked on Kalshi” - will be the difference between a safe and secure election, and one that’s hacked by motivated parties, is pretty crazy.
1. But you want markets like this to grow in popularity, right? Would you be opposed to an election market that had billions at stake from major players? Would you agree that at that point manipulation might be an issue?
2. The problem is that, in the US at least, the two-party system creates an equilibrium that's close to 50/50 win/loss for each party, so many elections are reliably on a knife's edge even after all those actors and factors are taken into account. Thus, adding a new factor, even one that's small in comparison to all teh existing ones, often *can* tip a close election - and these markets would incentivize specifically seeking out tippable elections to manipulate.
Of course, you could say that if the process is that close to random anyway, it doesn't matter if one more factor tips it, some factor was going to tip it some pseudo-random direction anyway. But I think there are some constraining factors on the existing forces, in terms of being long-term actors with public reputations that they care about maintaining, which may not apply to pure market manipulators. And I think the various financial crises we've gone through show how much financializing some part of our daily lives (like mortgages or supply chains) can drastically distort everything involved... again, if these markets become much larger and more profitable.
> Bibi must have sold his soul to the Devil or something, I cannot believe he’s going to make it through this.
This isn't that surprising, there's no election scheduled until 2026, his coalition partners all know they'd be out of power if they call for an early one, and it'd take about ten defections from his coalition to remove him. Something truly drastic like a fully general strike might be enough to force it but it's not quite clear if he's quite that unpopular (roughly half his base have stuck by him), and there's not much appetite for drastic action like that in wartime.
(Well, one deal with the devil he did make is to purge his party of anyone with actual principles so there's not enough people willing to resign over them).
A few months ago I did a shallow investigation into the "is sending free-entry postcards to sweepstakes gambling sites an infinite money glitch" question. There's actually a discord where people who do this seriously hang out. I have no clue where I found the link, but I'm sure some googling will turn it up if anyone else is interested.
The simple answer is that it isn't as straightforward as it seems. Legally, to get away with this, the companies should be crediting users with $5 per postcard sent. That does imply that people should be able to make something like $50/hr if it takes 6 minutes to write a postcard (minus the time and money for sending the postcards). Compare that to working at McDonalds for $10/hr, and it seems like the obvious choice for people who are making near minimum wage. But it doesn't work well in practice. From what I've read, the gambling sites obviously don't like this, so they'll look for *any* reason to discourage or disallow it. It'll take something like 2-6 months to be credited to begin with (which probably discourages anyone hoping for quick cash), and then if they suspect you are writing too many, they'll start disallowing them for ridiculous reasons. Like a single letter will be imperfectly drawn (maybe an "o" that wasn't closed perfectly) so they'll say "we couldn't tell what letter it was, we can't credit it." Their terms of service make clear that this is up to their discretion, and I don't think there's any auditors making sure they are being fair with evaluating free entries from repeat entrants. If somehow users are still getting away with it, they'll look for any reason they can to ban the account entirely (like if you ever accidentally log in with a VPN enabled, or if your mailing address via the post office is formatted slightly differently than the one you entered online). Most of the people in the discord who did this seemed to be simply experimenting to make a bit of extra cash on the side, but understood that the casino would punish you if you tried to get too much out of them. Plus, nonstop writing is a mind-numbing job to do for more than a few hours here and there.
I do think these companies are *extremely* vulnerable to legal action, and I think there'd be a surprisingly easy case to win for someone who wanted to pursue it. Basically, to truly be in compliance with sweepstakes law, non-paid entrants are supposed to receive "equal dignity" to paid entrants. That means they shouldn't face greater odds or obstacles to winning the "sweepstakes" than paid entrants, and the non-paid entries shouldn't be of lesser value than the paid entries. Any *actual or perceived* disparity between how paid and non-paid entrants are treated is an issue here.
If the above facts about postcard writing are true, there are clearly massive differences in odds and obstacles. Say the prize in question is $10k in sweepstakes coin redemption. Obviously this is far harder for the postcard-writer to achieve than for someone who just deposits $10k in cash. There's even some case law where judges have ruled that sweepstakes providers can't insist on onerous requirements to mail in entries, if entries could be just as easily submitted at the point of purchase (like the online user interface that paid entrants use). Also, when a judge considers this, there's the obvious background fact of "this looks a lot like a would-be illegal gambling operation shamelessly exploiting a loophole" which might bias their ruling in favor of the plaintiff.
I actually think this sort of legal case would be an altruistic thing to pursue. I tend to think gambling is a huge net-negative on society, basically a regressive tax that punishes people for statistical illiteracy or addiction/impulse control issues. I suppose I might swallow my ick and approve of it if it's heavily regulated and generating lots of tax revenue that is used for pro-social things, but that's not happening with these sweepstakes casinos — they're explicitly using the loophole to avoid the usual pathways to, and benefits from, regulated gambling. If I had the time, tacit knowledge, and money to pursue this litigation, I probably would myself.
> The unsolved problem is translating the research into real-world impact
Maybe its all just a gigantic waste of time and it will never have any real impact? I'm an active Manifold user but also suspect that its never going to change anything.
I'm starting to think there is a lot of path dependency in EA. Some causes like prediction markets are promoted, not because they have a demonstrable effect, but because they are rationalist hobby horses, and rationalist had a large foundation effect on EA.
Yeah, but sometimes you have to try an interesting idea, to see if it works? IMO, the real test is what quantity and quality of negative results it takes, before EAs adjust their spending allocations away from an idea. Which goes right back to rationality and willingness to learn and willingness to subordinate ones' abstract logic to real-world empiricism.
I also find the coal prediction really interesting. For me my instinct comes down to, "so many other countries are trying to phase out coal so quickly, and China has so much incentive to remain dominant in PV manufacturing for export, that coal will be super cheap and there's little reason to decommission plants early through the 2030s."
They might anyway if energy storage gets abundant and easy way faster than anyone anticipates? And if we all somehow get out of our own way and give ourselves permission to build all the new infrastructure and stuff we're going to need?
There's also some interesting questions about what happens to the prices of various chemicals as we encounter differences in how quickly we can decarbonize the sectors using different fractions of a barrel of oil. By the mid-2030s that could get quite strange, and maybe there's a role for coal to play in that?
It doesn't surprise me at all. Most electricity generation occurs from a turbine spinning a generator. Solar power does not operate like this and requires an inverter, with rare exceptions like molten salt systems. Same with wind power, even though it uses spinning blades it generates DC and has to inverted to AC to be used in the grid. The more penetration from energy that doesn't spin up with the grid, the less efficient it is. Unless your region has dense hydroelectric or geothermal power, this means using coal or natural gas. Or nuclear, but good luck getting a nuclear plant built.
Energy storage would have to undergo a huge paradigm shift to ever possibly have enough capacity for a modern electrical grid. There's no way this is possible with current physics and materials. Even getting enough materials for current electric generation is difficult. Copper prices have exploded with the expansion of wind and solar, quintupling in price from 20 years ago.
My understanding is coal has to be pulverized and subjected to a chemical slurry to create oil-like products. I don't see how coal would work for chemical production except as a strictly worse form of oil.
"There's no way this is possible with current physics and materials."
With the exception of this one sentence, I think you're definitely right for at least the next 5 years, and almost definitely right for the next 10 years. After that, I don't think I agree, and by 2040, I definitely don't agree.
The coal-to-chemicals thing I'm very unsure about, compared to the rest. The reason I think it's even plausible is that after a certain point it becomes more difficult/expensive to change what we use each fraction of a barrel of oil for. Bit by bit, I expect more electrification of cars and home heating furnaces (heat pumps). I don't know which solution we'll manage to get to compete with diesel vehicles and generators. HEFA process SAF is already pretty close to cost-competitive for aviation, and I have reasonably high hopes in the long run for e-methanol in marine fuel use (maybe for trucks too, not sure).
But, what happens to the spot prices of naphtha if demand for petrol and diesel and kerosene drop in comparison? What infrastructure do we need to build to continue using the entire barrel of oil effectively? Is it worth it to do that, or do currently-too-expensive processes start to become viable in that world? Or is the resulting infrastructure dual-use in a way that enables continued change as other decarbonization trends continue?
Also: a world of EVs is essentially one where every household effectively owns a 2-day Li-ion battery. This is a nightmare for how we currently manage the grid, but if we manage to not be *complete* idiots about it, it's also a massive opportunity to enable greater renewables penetration.
Are you basing this on general technological optimism, or do you have something specific in mind? I don't think it's unreasonable to expect a breakthrough in the next 10-20 years, but I wouldn't bet money on it.
Oil refining is basically dumping crude oil into a distillation tower. It's already separated into a bunch of different component streams. If certain fractions aren't useful any more, it makes a barrel of crude oil less valuable. But it doesn't increase the cost of refining any of the other fractions. Maybe it gets expensive if some products have to be disposed of instead of sold.
EVs have a big problem with cold temperatures. I wouldn't want one where I live. I can only imagine how bad something like an electric snowplow would perform. Maybe this can be ameliorated with better technology, but thermodynamics is a cruel mistress.
My background is in materials science and physics. I've been an advanced materials analyst, and more recently a cleantech consultant, for the past 12 years. So, I have a lot of individually weak but mutually reinforcing bits of data in mind, that I can't easily summarize, but that led me to where I currently stand. The closest I can get to a good summary is: yes, the problems are very very hard, but there are enough paths forward that I'd be surprised if pushing the ones we know about doesn't get us most of the way to where we need to be to get a substantial majority of the way to decarbonization.
At that point we'll be in a much better position to tackle the remaining problems and next round of new problems. Even some of those are ones where we can kinda guess the shape of what problems they are likely to be and what a solution might look like.
So yeah. I don't pretend to know for sure which sets of solutions to which hard problems will end up being the best ones, though I have my guesses. I certainly think a lot of people are wrong about a lot of how they expect the future to look. And yes, some people try to push solutions that make no sense to me, or can't possibly work because they'd break physical law. But I think saying a specific practical problem *can't* be overcome even in principle, when there's no deep physical law at stake and where there's already so much effort going into so many possible paths of improvement, is a bad bet. You wouldn't believe how often people make a "that's impossible" claim about something *in their own field* and I show them that it had already been done, sometimes years prior.
First, sure, you wouldn't want a 2024 EV where you live, and there are plenty of reasons we don't use electric snow plows. That makes perfect sense. I also have no intention of buying an electric truck (today I drive a Sierra 2500) for a long time, if ever. Heavy duty ground transportation (aside from a few kinds of local fleets) and mid- to long-range aviation are both areas I consider far too hard to electrify, at least for a long time, and plausibly ever. But with the gradual shift from NMC to LFP chemistries, and with further decreases in pack prices enabling larger pack sizes, performance loss in cold weather might not matter in 2030 or 2035. This is even more true if we succeed in building anything remotely resembling a decent charging infrastructure.
And yeah, if we end up in a place where 10% of a barrel of fuel oil isn't very useful, the price of everything else just goes up a bit. But if we end up in a place where only 25% of the barrel is useful, then the premium you need to charge for what's left gets much higher. So does the amount of work you're willing to do to get something useful out of the rest, which is otherwise hazardous waste. But even so, eventually the economics could start to make currently-very-uneconomical things, like CO2 conversion to fuels or chemicals, not seem quite so bad if we have with average levels of further technical progress.
>"I have reasonably high hopes in the long run for e-methanol in marine fuel use (maybe for trucks too, not sure)."
Given the (admittedly minor) downsides of replacing gasoline with ethanol for passenger vehicles, I would expect much greater difficulty with replacing (presumably) diesel with even more-volatile and lower-energy-density methanol for trucks & ships.
Methanol definitely has its challenges, but compared to cars and trucks, the number of vehicles and fueling stations is much smaller, and also companies like Maersk are already starting to buy vessels with dual-fuel or fuel-agnostic engines. That's kind of a necessity for a world where you kinda just have to buy whatever fuel(s) the port you're in has available, and none of the emerging fuel options is widely available.
Bunker fuel for ships is similar to diesel but heavier, cheaper, and much dirtier (up to 0.5% sulfur globally, compared to 15ppm for on-road diesel in the US). Before 2020 it was even higher sulfur.
Methanol is hardly the only emerging alternative and several others are probably about as viable.
I noticed that when you described the CFTC's reasons for banning election predictions, you talked about whether they were "gaming", but then when you argued against them, you talked about whether they were "sports". These don't seem like equivalent categories to me; consider e.g. blackjack or bingo.
I haven't checked the original rule this is all based on to see which word it used, but either way, it seems unfair to swap one of these terms for the other when arguing.
Not only do I think prediction markets are excellent at nerd sniping, I think the exact same of the x-risk work. And, to be clear, I'm a nerd! I'm here! But it seems reasonably likely to me that they don't clearly meet the "important, neglected, and tractable" trinity that I continue to believe should be at the center of any Effective Altruism work (demonstrating the 'tractable' bit seems to fall by the wayside). I'm not denying these things might be cool or important! I'm just saying they are very unlikely to be the most effective way for altruistic people to spend their money and efforts.
In particular, they both fail the rich person self interest test. This is weaker for prediction markets, but rich people have an extremely strong interest in not having humanity go extinct, given they are also humans. Therefore, extinction issues are unlikely to be neglected *relative to their tractability* when compared to things like diseases that only impact poor people.
I think the pitiful (by ACX standards) engagement on this post answers most of the outstanding questions here. There just isn’t that much demand for the product. Online epistemology nerds are a much smaller group than you would assume if you spend most of your time in the ratsphere bubble.
> Online epistemology nerds are a much smaller group than you would assume if you spend most of your time in the ratsphere bubble.
Yeah, but the question is, is it useful? People who voluntarily learn cool new programming languages and think algorithmic analysis is fun are a small group of nerds, but their field is useful enough that a whole bunch of other people get into it for different reasons.
Just musing here, but all of these regulatory issues make me wonder if it might be easier to consider prediction markets as crowd sourced insurance instead. I know insurance is regulated, but it seems odd to stop people offering insurance against the wrong party taking government, which you would definitely expect to create risk to various kinds of businesses. There are so many ways to structure things in finance, perhaps it'd be possible to set up a company that offers loans with different rates of interest depending on which outcome occurs in the future...
I think in general, you'd have a lot of trouble proving that there's a specific harm you want to insure. "Who wins X election" is far too removed from practical matters to be insurable. And that's before you even get into the major conflict of interest issues.
Why not go to an insurance company or regulator and tell them that you'll be emotionally crushed if $favorite_team doesn't win the superbowl and want to insure against that potential harm? Go ahead, I'll wait.
Different industries have different norms. CDS is more like "investment" than "insurance", and anything goes in finance, as long as you follow the SEC and CTFC's many rules, etc. It's sort of like how you can buy options without owning the underlying stock.
But the insurance industry, which has products sold to ordinary people for a specific purpose, operates very differently, and you're not allowed to deviate from that model. You aren't even allowed to buy life insurance for a stranger (guess why), so why would you expect to be able to buy "insurance" for an election?
You can't go onto Robinhood and daytrade your neighbor's home insurance policy because that's not what it's for. It's for protecting unsophisticated people against a specific downside risk with clear social benefits, and all the regulations and subsidies and business models are designed around that fact.
"Different industries and different products have different norms and regulations" is ubiquitous, and you can see it even within finance. For example, for equity, the company has a legal duty to act in the best interests of shareholders, but for bondholders, the company is allowed to arbitrarily screw you over in any ways not explicitly enumerated in the contract, and there's an entire industry dedicated to helping them to do this.
One other point is that incentives often exist for a reason, and it makes no sense to let people "hedge" them. For example, you wouldn't want a politician to "hedge" against their risk of getting caught taking bribes and going to jail, because that removes the incentive to not do crime!
For the same reason, employees with stock awards aren't allowed to hedge those stocks, because it's *supposed* to be an incentive to align their interests with the company.
Another aspect of this is that hedging is useful when it counters actual risk, but there are strong social fictions that claim that various risks are higher or lower than they really are, and also a strong anti-sociopathy pro-empathy factor that pushes against calm, calculated approaches to unusually bad things.
Speaking of which, I'd like to hedge against the possibility of being caught in a motel with a dead girl and a live boy?
Incidentally, there's been a thing recently about restrictions on even people *selling* life insurance policies. Basically, if you take out an insurance policy for legitimate reasons but then later change your mind, you can sell it to investors. But then in the 2000s, investors started paying people specifically to take out insurance policies and immediately sell them, which is a) illegal and b) broke the financial model.
The reason is because life insurance is priced *under the assumption that it is held by ordinary customers* with their typical usage patterns, and in particular the assumption that consumers will often abandon a policy for non-financial reasons. If life insurance policies were a financial instrument held by investors, they'd have to be priced a lot higher.
You can see the recent Money Stuff articles for more information.
I'm amazed that even the Republicans were in favor of election betting here.
At ManiFest last year, someone gave a talk on the challenges of legalizing election betting, where they mentioned that there was strong bipartisan opposition at all levels of government throughout the country. What on earth happened to cause Republicans to suddenly endorse an unpopular fringe position being pushed by Bay Area techies?
I don't think the people pushing Sotomayor to retire are thinking about the issue deeply enough. First, she isn't an idiot and understands the politics so their advocacy only functions to make it harder for her to retire now (and psychologically less appealing).
Also it's a bad norm to encourage. If you look at what SCOTUS actually does it's clear that it's not just trying to maximize the policy outcomes the justices prefer. Indeed, what makes having a court valuable in the first place is the fact that they feel strong pressure to abide by norms which require a certain degree of consistency. Sure, they are people and are going to be biased, but you don't see even Alito saying: ohh you can ban Democratic campaign ads but not conservative ones or whatever.
Yes, just like bias, we know that political considerations no doubt influence when justices retire but the more obviously you violate those norms the more they decay. If Sotomayor does it now than Alito, Thomas etc are more likely to do the same next time around (they didn't retire under same circumstances w/ Trump).
So sure, I'd think it was a stroke of luck if she got hit by a bus and the Dems got to nominate a replacement but the people beating the drums should realize they are only doing harm by their own lights.
I think it's a reaction to "Nobody said anything even though Bader Ginsburg was visibly in bad health and frail, and look what happened". Sotormayor does have a chronic condition which may or may not be well controlled but could have a bad outcome, but mostly I think it's genuine concern that (1) Trump will win a second term and (2) Sotomayor does have to be replaced sometime during that term and (3) last time he was president, he got two Supreme Court justices of his choice and suddenly now we don't like governing via the Supreme Court (unless it's our people doing it) so we must act now to preserve some hope of a liberal torch being carried forward during the coming Three Days of Darkness.
I don't think it's unreasonable to discuss if Supreme Court justices should hang on until literally carried out in a coffin, but I don't know if it's reasonable or not to ask Sotomayor to jump before she's pushed.
"First, she isn't an idiot"
I remember the "wise Latina" remarks and they made me wince. I'm not saying she's an idiot but I don't think she's particularly impactful as a Supreme Court justice. I agree that the wider argument is around court packing.
If I'm Sotomayor or her close personal friend than those are totally valid considerations but it isn't an argument for public pressure. The failure with RBG was a failure by RBG and maybe some of her close associates not a failure to publicly declare she should retire.
The point about her not being an idiot wasn't a claim about how impactful she is on the bench or anything -- merely that she isn't likely to smack her forehead and go "ohh, I'd never considered that if I didn't retire now a republican might appoint my replacement." I just don't see why Sotomayor would be more likely to retire because of some public pressure campaign. She understands the considerations and the public pressure only makes the calculus less favorable (a retirement that looks less nakedly political is better for her reputation, less likely to encourage the two old conservative justices to follow etc etc) and it just tends to get people's backs up.
WTF does Nate Silver and others think what they are doing makes it more likely they get what they want with Sotomayor? And do they not realize that if they do get it their pressure makes the outcome worse for them as it encourages other justices to do the same?
For the first part, I think he's just offering up an honest assessment, and letting chips fall where they may? People engaging in pressure tactics are a different thing.
For the second, I assume he realizes that there's an equilibrium on the far side of this becoming common practice, but that the side that moves first will get a temporary advantage (which might be parlayed into long-term gains). And if it's inevitable that someone is going to move first, why not you, why not now?
Regarding the first point, I'd agree if the claim was merely: she should retire. But I don't think that works when someone says that the left should exert pressure or keep beating the retirement drum. It's totally fair to just say, X would be good and not address the question about X advocacy being good but in this case many of the players have endorsed advocacy of the position as well as it's truth.
Regarding the second point, I don't think that's quite how it works here. The norm is relatively vague. It's not like anyone believes that these thoughts have literally no influence on when a justice retires but there is a degree of pressure not to do so too openly or to only let it influence the deciscion around the edges. So I don't think you can analyze it as a once and done kind of thing. The next time around justices can always retire even earlier to affect the process. Also, there is the issue of the number of justices who can play that game. The more conservative justices there are the more it benefits the left for them to keep up the pressure against strategic retirement.
But I think it's plausible that it would be good for the left if Sotomayor claimed she wanted to spend more time with her family or whatever and gave some excuse to retire. What I judge to be bad for the left is if she retires in response toa broad political pressure campaign -- especially since it's hard to believe that would ever be effective at making her more likely to retire.
"Older Justices will retire strategically, meaning a bit before they would otherwise" sounds like something that can't get too badly out of control, since it takes a while to become an older Justice in the first place. Especially if the other half of the norm becomes "appoint people young enough that they'll last a few decades", which is also fine.
I don't think it would be some kind of disaster if it started to happen. I just think that it's at best unclear if the left is net benefiting or being harmed by the fact this isn't more common given that both Alito and Thomas could have retired under Trump.
Regarding resolution criteria my primary concern about prediction markets is that opening them up to more and larger amounts of real money will primarily result in improvements in figuring out unusual resolution cases and not in addressing the fundamentals.
Re: Bibi, what you are missing is simply how the election system works in Israel. from the moment the parliament votes to dismiss the government, it technically takes about half a year until an election happens. Meanwhile, the prime minister stays in power. Furthermore, the Knesset has time off similarly to the congress, and there is one such time off in the end of June. This means that unless sometime soon the Knesset does not vote the government out - an analysis which is too long for me to write here to give proper respect - Bibi remains prime minister until 2024 regardless if the government falls in July.
I think people (maybe including the CFTC) are focusing too much on the danger of election interference and resolution issues as opposed to market manipulations that affect the price before resolution. One of the big differences between a sports book and a prediction market is that probability changes after a purchase can affect the value of a bet, and it means that short term manipulations in the perception of a market can be more easily monetized.
On area I worry about is that regulation of financial markets involves some pretty strong speech restrictions to prevent e.g. pump and dumps or bashing. People are sort of fine with this, because commercial speech is understood to have relatively weak free speech protections. But, political speech is understood to have strong free speech protections. I don't know how it would shake out, but I understand the CFTC not wanting to deal with it.
- The statute explicitly allows the CFTC to add things to the list of things that are banned.
- Kalshi contests how broad this authority is, and whether elections can be added.
- Statute explicitly bans contracts that involve gaming. Agreed that putting politics/election markets in there is an interpretive choice. But I've been surprised the CFTC has let Kalshi have markets on the Oscars, Grammys, etc.
- The CFTC cannot charge fees and is generally small and not one of the bigger/well-funded agencies, so they are very much in fact limited by what they can enforce.
- CFTC needs to sign off on every specific prediction market (or OK "self-certification"); Kalshi's proposal specifically was "congressional control of Congress" [which I think was totally fine] but it's obvious that this would have led to more election contracts [which the CFTC anticipated and was worried about].
- The CFTC is a regulator. Regulating markets is hard. With commodity markets it's the "same" market, there's only so many ways you can do insider trading and manipulation. With prediction markets, every market is a bespoke, custom question -- the relevant insiders and ways to manipulate are unbounded -- this is incredibly challenging to regulate and an underrated policy problem with legal prediction markets. Add to the fact that prediction markets are illiquid and small, and everyone and their grandmother wants to start their own exchange, the CFTC simply views them as an annoyance and does not want to sink limited resources here better spent policing bigger markets. Giving the CFTC more taxpayer money to legalize election betting screams "bad politics."
- Right now, Kalshi has to tell the CFTC what kind of market monitoring and surveillance and KYC they are doing to keep their exchange a nice, fair place. They are the first line of defense [because again, the CFTC is just a smol bean regulator] when people are behaving badly on their platform -- they have to identify and report those people, kick them off, take action, etc. If they don't do this properly....and they have 100s of election markets........could the CFTC meaningfully tell Congress they are overseeing what they are supposed to oversee?
- Similar note, medical regulators would probably ban surgeries that were somehow different every time you performed them and were a very niche product and also hard to regulate. Resource constraints and tradeoffs are very real!
- US campaign season is longer, lot more money at stake, real money makes things messy. Again, Congressional control contracts are fine, "who will win House District with 8 candidates" is rife for abuse.
All this comes down to: from the outside, it may seem like if the government allows X the government should allow Y (CFTC has no obligation or influence over Title IX!), but if you are familiar with the details and incentives of the specific regulators, the CFTC's actions make sense [given their incentives and tradeoffs].
All this said, I think Kalshi's original proposal was fine [25000 USD position limits] and could be approved since it was limited in scope, the CFTC could establish rules on minimum electorate size for an election market, or other such things. But this takes work and I can't stress enough how much this is just not a priority for them. Kalshi's new proposal [made fairly in response to CFTC's argument that nobody could hedge meaningfully with 25K limit] allows eligible contract participants to put up to $100 million (I think) which is just too much too fast imo.
> In order to maintain the fiction of being a “sweepstakes”, these casinos have to give you “tokens” if you request them by mail. If you send a postcard to Chumba Casino asking for free money, they’ll give it to you, $5 per postcard. Is this an infinite free money pump? My impression is in theory yes, but the postcards have to be handwritten in a very specific way, the company sometimes rejects them for weird reasons, the cost of materials and mailing lowers your profit to more like $4, and so you’d have to hand-write 250 postcards to make $1,000. I’m still surprised more people don’t do this.
I remember reading about this on cereal boxes. That was a long time ago, but the impression I formed was that everyone is allowed to enter a sweepstakes for free in the manner you describe. It wasn't that everyone is entitled to as many free entries as they want - I thought there was effectively a limit of one free entry per person.
> Here’s an embarrassing screwup from Metaculus. This question was about when there would be a “Great Power war”, with Great Powers defined as any country in the top ten of military spending. But surprise surprise, Ukraine getting invaded made them spend a lot of money on their military that year, so they rose to #8 in the world in military spending in 2023. Since Russia is also in the top ten, this qualifies as a “Great Power war” by the technical definition, and the question resolves positive. Moral of the story: resolution criteria are hard!
This also happened in the ACX 2023 predictions, where one of the questions was titled "will there be more than 25 million confirmed COVID cases in China", but the rules for that question supplied a hidden "...as reported on a website that doesn't track any relevant data".
I reserve the right to feel particularly bitter about this in a context where you were specifically encouraged to do no research before answering the questions. The only significant element of that question wasn't even included in the wording of the question.
My main thought in specifically manifold stalling out is that for a semi casual semi normie user like myself the beta surfaces for my perusal just aren’t that interesting. The first couple times I logged in it was awesome to see all these options and make some evaluations/bets.
But the 3rd and 4th time, and particularly the 10th time the beta all start to look the same and so limited in their scope and areas of interest. A bunch of politics stuff that probabaly doesn’t have answers or is extremely sensitive to insider info, which is fine but not interesting for a random to bet on.
Lots of weird Sorieties paradox inspired self referential or scam postings. Which are sort of amazing and funny the first time, but once again not the tenth. And then just lots of random trash. I feel like after a couple months of using it I would log in once a week, look at 40 or 80 markets, not really be interested except one or two, and then just slap some mana down and leave. It simply wants that betraying/amusing and there was so much chaff to sort through.
I feel like I would love a mode (maybe this already exists) where there is a 1990s NYT newspaper editor exercising some judgement about what is worth showing me.
> Polymarket on 2024 election results. In the past they’ve had a Republican bias, but now their Presidential markets are in tune with everyone else in the polls, so maybe this is accurate.
Well, according to the table, there is a 102% chance that no third parties will control any of the three groups, and a further 2% chance that they will.
The table also specifies that the odds of the democrats winning each group are 55-47 for the white house, 25-77 for the senate, and 58-44 for the house of representatives. To be compatible with the individual outcome probabilities, there must be some really strange conditional probabilities involved. (For example, the most likely outcome piecewise, D-R-D, is judged to be 12% less likely than D-D-D, despite the fact that R is favored over D in the senate by 77-25.)
(I believe that there are strange conditional probabilities involved. But they should come with detailed explanations. I think the most likely explanation _here_ is "people like betting on a sweep".)
> regular bookmakers who just do their own research and offer odds to punters, there's very little in the way of explicit prediction 'markets' in the way they function on e.g. PredictIt,
Bookmaker odds adjust frequently, and they adjust to balance the bets on either side of an outcome. The bookmaker profits by taking a vig, not by forecasting better than the wisdom of the crowd. They are functionally almost exactly like prediction markets. (There are exceptions like unusual prop bets, but those are exceptions.)
'The UK has election betting and none of these negative outcomes have come to pass.'
It does, but this point actually counts in favour of 'elections are sport'. Election betting in the UK mostly happens through regular bookmakers who just do their own research and offer odds to punters, there's very little in the way of explicit prediction 'markets' in the way they function on e.g. PredictIt, and often elections will be one of very very few non-sporting events these bookmakers offer odds on. It's rather amusing - there'll be a drop down menu that goes something like; Football - Cricket - Rugby - Snooker - Racing - Motorsport - Politics. If you didn't know what politics were you'd definitely think it was just another sport. Indeed, if you go to the bottom of the politics page on William Hill, at the bottom of the featured odds it still says 'See All Today's Matches'. In Betfred there's a section called 'Rishi Sunak Specials'.
This also means the regulatory problem is totally void, since it's just up the bookmaker's discretion as to whether they pay out, as is the case with all types of sports betting.
Payment is not at the bookmaker's discretion. Gambling debts have been legally enforceable since the Gambling Act 2005.
Yes but in the case of disputed outcomes the discretion afforded to bookmakers is pretty wide no? Especially since they of course write the resolution terms so can gives themselves a lot of room. I know there's dispute resolution like IBAS but again I think it's non-binding no?
Betfair and Smarkets are real prediction markets (i.e., odds are not determined by bookies) in the UK. Though I don't know how much market share they have.
Betfair has a sportsbook business and an exchange, where the exchange makes cash via a cut (5-20% of profit depending on whether you're a frequent flyer). The exchange is the largest in the world. Flutter (the PaddyPower/Betfair parent) is the no. 2 UK betting company, and the exchange is probably 5-10% Flutter's business by revenue? I'd be very surprised if BFEx was more than say 5% of UK market share given Bet365's size (incidentally a rare British world leader). Probably much lower than that given bricks and mortar retailers too.
Having said that, I think Betfair Exchange is the single largest betting exchange in the world- but demand just isn't as large outside of the smart money.
I wonder if it is not best for prediction markets to lean into the gambling space and out of the financial contract space. Sports gambling is starting to be legal in a lot of states and it would seem easy to just move event contracts to something like Bovada.
"I think their thought process is: if you manipulate the commodity markets by (for example) saying that you have lots of nickel when you don’t, the CFTC has to investigate that and penalize the people responsible. In a hypothetical world with election contracts, if someone manipulated an election - for example, they put out a fake poll showing that the incumbent would definitely win so there’s no point in even voting - someone could ask the CFTC to investigate."
Is their thought process not "the outcome of the last US Presidential election was contested and extremely controversial, we would prefer not to get involved in establishing who won it"?
I think it's also tied to things like Citizen's United (the actual case not the boogeyman) where they might have to rule if various campaign messaging is market manipulation.
The outcome wasn't controversial; what was controversial was the rightfulness of the outcome.
Yeah, my thought was the same.
Degenerate UK gambler here
I absolutely don't think corporations are piling in on the election to hedge risk - there just isn't that much money in the markets. Can't immediately find a figure but here is oddschecker boasting in October 2020 that one bookmaker had already done $182m on the US election. Not a heap of cash compared to 23bn bet on this year's superbowl. (I staked £1000 to win 600 on Biden)
As far as fake polls are concerned pollsters are tightly regulated by statute and the trade body the UK Polling Association. I don't see an issue here
The other thing is it's relatively easy to make money in UK political betting because it's a small and inefficient market. My Biden bet was so big by my standards I am definitely in profit overall
Politicalbetting.com is the prime discussion forum, for those in jurisdictions which don't block access to it
On skybet in the UK a little while back, it was possible to bet on Joe at 2/1 and Trump at 10/11. Checking now, it's come down to much more reasonable levels - you can still guarantee a small amount of money if at least one of them wins, but the difference now is probably a reasonable reflection of the chance that neither are able to take office, whereas 2/1 and 10/11 was not.
> Because real money is involved, Manifold will have to tighten the rules on markets, including banning N/A resolutions.
What?? I already have markets of the form "Conditional on X happening, will Y also happen?". This means that I resolve YES if X and Y happen, NO if X happens and Y doesn't, and N/A if X doesn't happen. This is a very convenient way to try to get an estimation of some conditional probability. Moreover, it's a promise I already made to the traders when I set up the market rules like that. What the hell am I going to do?
Those markets can still N/A, just ping mods.
So I understand correctly that I won't be able to resolve them manually? Each case would require mods' intervention?
I probably need to warn the traders on my questions about this...
Otherwise, it sounds like a good way to estimate conditional probability will disappear, and this has been an important use case of manifold. ("If I read this-and-this book, will I like it?" etc.)
Yes, each existing market will need mod intervention.
For future markets you can instead do a linked multiple-choice market with 3 or 4 options, e.g. "If I sleep will I be tired?"
* Yes sleep, yes tired
* Yes sleep, no tired
* No sleep, yes tired
* No sleep, no tired
Or just one "No sleep" option instead of the last two if the question becomes moot or sufficiently uninteresting without the precondition being met.
This is what they've written:
- Market resolution
- We plan to significantly tighten rules for market resolution to eliminate abuse. Expect more rule changes.
- Creators will no longer be able to resolve markets N/A in the typical case.
- This change is to limit abuse that comes from traders cashing out profits that later get clawed back due to N/A’s.
- We will probably allow N/A resolutions in the first few minutes after market creation to handle typos or other cases where the user wants to remove the market.
- Conditional markets (”If A then B”) can be implemented as a multiple choice market with each options for A and B enumerated (”1. A and B, 2. not A and B, 3. A and not B, 4. not A and not B”). We intend to create a first-class market type for conditional/decision markets
https://manifoldmarkets.notion.site/A-New-Deal-for-Manifold-c6e9de8f08b549859c64afb3af1dd393
I'm also quite concerned about this.
Maybe I just have very different intuitions than everyone else, but I still find it super weird and bad that it's even possible to ban prediction markets at the federal level without congress passing a law. I get that it's technically legal, but it feels to me like some kind of dictatorship-by-bureaucracy where people that nobody voted for are able to effectively pass federal laws by decree, and we don't have any power to vote them out.
Another example is how last month the FTC issued a federal ban on noncompete clauses in contracts.
https://www.ftc.gov/news-events/news/press-releases/2024/04/ftc-announces-rule-banning-noncompetes
A lot of people were celebrating this and I think that it's probably a good rule to have, but it struck me as bizarre that a federal agency can just make up new laws by decree, and it seemed like this same power could be used to make bad laws. Well now with this prediction market news, I consider that worry to be vindicated.
1. Congress previously passed a law that said "agency: please do research and hire experts, and then make rules in support of goals X, Y, and Z"
2. If Congress doesn't like a decision the agency makes, they have [90 working days, or something like that] to veto the regulation. Or they can always pass a new law that says "sorry, the rule is actually B, not A." With or without changes to X, Y, and Z.
That's a good point (same goes for Dan L who made a similar point), but it still doesn't sit right with me. Like I get that they have the legal authority to do it, but I'm still against it anyway and think that the way things work should be changed.
It’s just a continuum. The agencies need some latitude for action and decision making, but there has to be a line somewhere.
Congress absolutely *could* step in and pass a law here either mandating or forbidding a ban, it's in the absence of action that agencies fill the vacuum. It's important to distinguish between "we have no power to do X" and "X will predictably not be done by those with power" - call it a failure of niche regulation, where anything outside of the dozen most important issues fails to produce a strong enough electoral signal and isn't worth an elected's attention.
Congress in particular seems to be quite risk-adverse, refusing to even confirm good rulemaking via statute. IMO, that's the result of an *increased* responsiveness to public opinion combined with an asymmetric public interest in punishing bad v. rewarding good, but that's a tangent.
generally if Congress were going to handle this, they'd have to massively expand staff budgets to hire a bunch of experts to constantly evaluate and handle all the decisions, which would be very unpopular in itself. Regulatory decisions would get also delayed constantly due to the Congressional calendar.
And ultimately the actual representatives would just wind up deferring to a different set of unaccountable experts, maintaining only nominal veto power, only now they work out of a different office building in DC than the current CFTC.
> it feels to me like some kind of dictatorship-by-bureaucracy where people that nobody voted for are able to effectively pass federal laws by decree, and we don't have any power to vote them out.
It's better than that. The officials above them in the chain of command can't replace them (or just fire them) either.
> Maybe I just have very different intuitions than everyone else,
Not quite *everyone*. The MAGA Movement also finds the Deep State rather undemocratic.
Everyone finds the government undemocratic when it does things they don't like. I doubt they care about the "deep state" when it's doing things they like though.
> The part I don’t understand is the last one (73% found guilty of felony in New York) vs. the second one (56% of any felony at all).
It could be that he will be found guilty, just not before the election
> Kiko Llaneras of Spanish media EL PAIS is hosting an elections forecasting tournament on Metaculus.
Unimportant correction, but it's not a tournament, just a question series.
I commissioned the Good Judgment study and the poll of virologists and epidemiologists on covid origins. The resources we shared with the forecasters are listed in the full report and the supporting document which has all of the superforecaster deliberations.
https://goodjudgment.com/wp-content/uploads/2024/03/Superforecasters-Covid-Origin-240312.pdf
I think this is what you are after in the post?
The resources we shared were the rootclaim debate, which has a lab leak proponent and a zoonosis proponent
On the zoonosis side, worobey 2022, Holmes 2021, Pekar 2021, and Andersen 2020.
And on the lab leak side, stoyan and chiu 2024, the bayesian analysis by Michael Weissman, the bayesian analysis by demaneuf and de maistre, and demaneuf's presentation to SAGO on case reporting early in the pandemic In Wuhan
Direct link to the Superforecaster commentary: https://goodjudgment.io/docs/Superforecaster_Commentary_on_Covid_Origins.pdf
So reading your report the only data that would make your "superforecasters" change their mind would be a published written confession by the CCP that they ordered covid and released it. I assume they still deny the holocaust and holodmer under those standards.
Just to note that the Swift Centre's forecasts for coal consumption still expect consumption to peak and decline, just on a delayed timeline. They have China peaking by the early 2030s. From that link it seems ambiguous what peak timing they expect for India, except that they'll be slower.
"Trying to weasel out of this by classifying elections as a sport is pathetic"
This strikes me as really quite uncharitable (quite uncharacteristically for Scott, as both a generally nice person and also as author of "The categories were made for man, not man for the categories"..)
It seems to be entirely reasonable to say that elections aren't a sport if you *just* ask "does this activity have a bunch of people dressed exactly the same running and jumping and throwing things?" - but it does seem very much like one if you ask the question that regulators are supposed to care about, viz. "is this activity the sort of thing a gambling addict is vulnerable to becoming obsessed by and ruining their life over?"
People become ruinously obsessed with sports betting because sports are supremely prevalent and in-your-face in media/news/advertising, are hotly contested, generate strong passions, and are talked-about by everybody, everywhere, all the time (with the possible exception of at ACX meetups..) These properties seem to be very closely matched by politics - that's the dimension regulators ought to care about, not the level of athleticism involved.
You could say politics is a blood sport, it's the most popular one in these British Isles!
Haha!
(I thought the most popular blood sport in the British Isles was Yellow Car...)
Gamblers could get fixated on whether the groundhog sees its shadow, but that wouldn't make it a sport in the sense people normally understand it (unlike the whales from Scott's post you cite). If there's legislation about sports specifically, then it needs to apply to sports. If the regulator is supposed to be restricting anything gamblers could get fixated on, then the legislation would have needed to say something like that instead.
To make explicit the analogy to Scott's earlier post: People normally understand that whales are mammals, but King Solomon, who only cares about a subset of the properties of whales, still needs it to be within the purview of the Ministry of Behemah, not the Ministry of Dag. Similarly, people normally understand that politics isn't a sport, but the Financial Conduct Authority (or whoever), who only care about a subset of the properties of politics, still needs it to be within the purview of sports gambling regulation. I'm not necessarily arguing in favour of doing this (personally I think the US overregulates prediction markets, and mostly does so for reasons unrelated to -and less reasonable than- their risk to gamblers) - I'm just saying that it's a perfectly reasonable argument to make, not a "pathetic and weaselly" one.
I'm not suggesting that regulators are supposed to regulate every possible thing gamblers could become fixated on - I'm suggesting that since we have an existing refusal-by-default category that includes sports it would be reasonable to argue that politics should also be in that category, and somewhat uncharitable to claim that regulators are "pathetic" and "weaselling out" by so arguing.
(Perhaps the solution would be to say "That regulatory category we used to call 'sports'? Well now it's called 'high addiction capability' and it includes both sports and politics". Presumably one can't just arbitrarily redefine existing regulatory categories like that, so I think it's plausible that the path of least resistance may have been just to say "For the purposes of this legislation, politics and sports look the same along dimensions we care about, therefore, fine, we're going to call politics a sport". As HaroldWilson (presumably not *that* one...) points out elsewhere in the comments, UK bookies already do something much like this in their internal systems.)
No, people in Solomon's kingdom don't understand that what we in the 21st century call "whales" (in a language that hasn't yet come into existence in his time) are "mammals" (another term that doesn't exist, as they don't group animals by whether they have mammaries). If he announced laws in the language of 21st century English, they would not be understood in his kingdom. If you wanted a less distant example, you could cite SCOTUS declaring that tomatoes are a vegetable (under the culinary definition) rather than a fruit (under the botanist's definition).
> Perhaps the solution would be to say "That regulatory category we used to call 'sports'? Well now it's called 'high addiction capability' and it includes both sports and politics"
Actually changing the language of the law to fit their extension would be less weaselly than claiming the law covers it via a term nobody thinks includes politics other contexts.
Even if you put King Solomon and everybody in Biblical Judah through a biology PhD programme* and lectured them in excruciating detail on the natures of whales, fish, and mammals, they would still be better served by calling whales a fish rather than a mammal if that's what it took to get whale-management under the purview of people who live on the coast and go around in boats.
(* How's that for a creative writing prompt..)
Not heard of the tomato thing before but I agree it's an excellent point: tomatoes are a fruit and are commonly acknowleged as such - but it's perfectly reasonable, and certainly not pathetic or weaselly, for some culinary authority to say "We're going to call tomatoes a vegetable for our specific purpose."
I agree that changing the legal definition of the category to point to the thing you actually care about (instead of its pointing to one specific example of that thing) would probably be preferable, but I don't at all agree that the reason it would be preferable is that it would be less weaselly.
Thanks for the discussion, by-the-by!
The important thing, for both the kingdom of Solomon and SCOTUS, is that people were already using such words to refer to whales and tomatoes before the rule was promulgated. Other people can use different language, because language is to a certain extent arbitrary (hence the many different words in different languages to refer to the same concept), but the language used for the laws had commonly understood meanings, and indeed they wouldn't change their minds if told their language was "incorrect" based on someone else's reasoning for categorization.
Whales are indeed mammals, but you're not going to get very far in whale management by sticking one in a field and encouraging it to graze on the lovely pastureland.
The categories here are actually legal categories written by Congress. This isn’t just semantics, the real question should either be what Congress intended when they wrote the law, or what the original public meaning of the text meant when it was passed, depending on your legal theory. This feels less like one of those and more bureaucrats wanting to avoid responsibility. But that’s why PredictIt has their lawsuit for a Judge to figure out.
I agree that what Congress intended is an important question, yes. But I have 4 objections to the broader point:
1) I think "they intended to prevent certain harms that can arise from gambling, except that when they passed the law these harms were mostly restricted to sports betting because back then prediction markets were merely a twinkle in Robin Hanson's eye, therefore treating politics betting similarly to sports betting would be consistent with Congress' original intent" is a perfectly fair and reasonable answer to that important question.
2) I don't think it's the *only* important question; I think it's also fair and reasonable to want to evaluate competing interpretations of legislation based on the actual outcomes of each interpretation as well as (or instead of) channelling the spirits of the original legislators.
3) In British and international maritime law (my own area, sort-of) many set-in-stone legal definitions, both modern and centuries-old, are *incredibly* vague, often to the point of uselessness, and open to broad and possibly even contradictory interpretation - I would be awfully surprised to learn that this isn't also true of US law!
4) *Are* there even strict Congress-mandated legal category boundaries, here? If some institution is allowed to categorise politics as a sport (or for that matter the tomato as a vegetable..) the legal category boundaries we're working from can't be all that strict, can they?
Isn't the problem with sports betting that the competitors can bet on their opponents and throw the match? I have a hard time imagining this happening in politics, although I suppose a candidate in financial trouble could make a big bet against themselves just before admitting to a scandal or otherwise sinking their chances.
If we're talking about the giant defence firms etc. having millions to throw around and vested interests in the outcomes of elections, I see no reason why they wouldn't try and interfere *if* prediction markets become what it is hoped they will become, indicators of the most likely outcome.
Personally, my view is that if you open them up to the public and let them run forecasting markets on all topics, they will become betting sites not forecasting, and results to set policy will be as effective as the "stick a pin in the list of runners" way to select a horse to bet on.
But suppose they function as hoped, and the result is seen as "this is going to happen with a very high degree of certainty" which everyone accepts.
Okay, now it's worth my while to try and tilt the result to read "Senator Smith (who is likely to steer juicy government contracts my defence contractor firm's way) is going to win this election" rather than "Senator Jones (who is running on a platform of cutting such spending) is going to win". I want to get people thinking "Okay, if SurePredict is saying Smith's the winner then it's no use my turning out to vote for Jones". I want a result mediated by the prediction market that will be the same as California being so solidly Blue, you could run a bag of flour as a Democratic candidate and get it elected before a Republican.
If it stays as sports betting, then it's not worth my while to mess around with it.
I can see why the CDC doesn't want that kind of betting and doesn't want to get involved, and I'm more sympathetic to their position than "prediction markets are the way to set policy because something something rationalists something something wisdom of crowds".
"Okay, if SurePredict is saying Smith's the winner then it's no use my turning out to vote for Jones"
Then there's equally no point in turning out to vote for Smith.
What you want is to get just enough people to decide it's no use voting for Jones so that Smith wins.
How many comments over the years have we had on here that "your vote doesn't count, here's my maths working out to demonstrate that's true, so don't bother voting"? Scale that up to manipulate a prediction market outcome and you'll get the thousand people saying "well it's pointless me voting" that permits Smith to win.
But your vote really doesn't matter in practice relative to the election result. What matters is whether enough people believe that it's their "civic duty" or some such to turn up, and here all sorts of propaganda indeed can have disproportionate impact.
How do you make sure those thousand people were going to vote for Jones rather than Smith?
I want to deploy Kant about this, but maybe Jordan Peterson deontology works better. "Clean your room". "Vote."
Doesn’t this argument also apply to polling?
If you think politicians won't throw elections, probably it makes sense to pay them for admitting to scandals.
And that big bet against themselves would... Move the market in the right direction, hence making private information public and performing exactly the role prediction markets are meant to.
there was a minor scandal with the head of a political consulting firm who did some polling, was also betting on elections. From his point of view, he was just doing the good forecaster thing of putting skin in the game.
From most peoples' point of view, he could have easily made money by manipulating his polls.
Given there have sometimes been ambiguities in what people might count as a lab leak vs "natural" zoonosis, worth highlighting that Good Judgment's definition of a "biomedical research-related accident" states:
"...that SARS-CoV-2 emerged in human populations as the result of the accidental infection of a laboratory worker with a natural coronavirus; or the accidental infection of researchers with a natural coronavirus during biomedical fieldwork; or the accidental infection of a laboratory worker with an engineered coronavirus; “research” includes civilian biomedical, biodefense, and bioweapons research."
"The UK has election betting and none of these negative outcomes have come to pass."
Yes, and these bets are handled by bookmaking firms, not prediction markets. This means that:
(1) They are considered games, like betting on who is going to win a golf, soccer, racing or other sport tournament
(2) The bookies have robust anti-fraud/anti-rigging the market protections in place, because surprise surprise, people have previously attempted (and gotten away with) betting coups:
https://www.sportingindex.com/community/sporting-events-articles/sports-articles-from-sporting-index/10-famous-betting-coups/
https://www.theguardian.com/sport/blog/2014/dec/19/the-joy-of-six-betting-coups
https://en.wikipedia.org/wiki/Gay_Future
https://racingtoursireland.com/great-irish-racing-betting-coups-6/
Where money is involved, people have always tried to cheat, find loopholes, commit fraud, or rig the system.
"In 2010, Curley decided to try and emulate his feats of 1975. With the far more sophisticated world of modern bookmakers, it seemed a much greater challenge, yet he succeeded once again. Modern technology means there’s no way you could pull off a “Yellow Sam” kind of coup nowadays. So, 35 years on, Barney began to look for ways of outwitting the bookies once more. He is a technophobe but he recruited a young whizz kid who found the loophole they needed.
He worked out that while betting shops are vigilant for small single bets being placed on one horse they don’t look out for multiple bets where the punter bets on winners of more than one race.
It goes against all bookmaker trading experience for a strongly fancied horse not to be backed as a single. Why would anyone jeopardise a payout by including it in a double or a treble with other horses? It just doesn’t happen.
Now based in Newmarket in the UK, where at one stage he had 50 horses in training, Curley set himself the task of matching four runners to four races on the same day which stood a decent chance of winning at good odds. It was a massive undertaking, as it is hard enough to send out one winner with a degree of certainty.
He also had to do it behind a veil of complete secrecy, given that the bookies were particularly vigilant for anything involving horses from Curley’s stable.
Via another accomplice, former City trader Jack Lynch, he hired a gang of 30 “putters-on”. Armed with fold-up bikes and untraceable mobile phones, they were given carefully planned routes to cover as many betting shops as possible in a day. It was also crucial that while they moved between them as fast as possible, they dawdled once there. They had to look natural so as not to attract attention.
The plans changed many times before Curley fixed on four horses and races on May 10th 2010. The horses were Agapanthus in the 4.10pm at Brighton, Savaronola in the 5pm at Wolverhampton, Sommersturm in the 5.30pm at Wolverhampton and Jeu De Roseau in the 7.30pm at Towcester. Only Agapanthus had won a race of any kind in Britain and the extent of the gamble was made clear when Savaronola went lame two days before. But Curley decided to go ahead anyway. By late morning on the day they were respectively priced at 11/2, 4/1, 9/4 and 25/1. These odds were sure to dive. But by then, the bets would be safely on.
Around £100,000 worth of bets were laid in various multiples, keeping the stakes as low as possible in order to stay under the radar. The bets were trebles on three of the four horses, in different permutations, and Yankees, a multiple bet on four horses in which at least two horses had to win for any payout at all. It seemed like madness but that was part of the point: to make sure the bookies’ early-warning systems weren’t triggered.
The team returned to base to watch the races on TV. Their first horse Agapanthus was by now hot favourite, but it had the gang worried by starting slowly before coming through to win by two lengths. In the next race, Savaronola was up against Alex Ferguson’s If I Had Him and was 11/10 favourite – it notched up its first ever win, finishing six lengths clear. Apparently, Ferguson was fuming, later telling Curley: “I backed that bloody horse that day!”
But disaster struck in the third leg of the bet. Sommersturm had been seen as the certainty of the foursome, but trailed in fifth. This was not in the script.
All hopes were now pinned on Jeu De Roseau. Rival Jockey Tony McCoy set a hot early pace on his mount Manjam, but Jeu De Roseau’s rider Denis O’Regan bided his time and pushed the exhausted horse into a winning lead in the final furlong. The banker had failed to deliver but the outsider came to the rescue!
If Sommersturm had come first, the payout would have been more than £15million. However, three out of four meant they had netted £3.9million, reportedly the biggest win in history."
What essential characteristic of sport do you believe elections lack?
Well, the obvious one is that sport outcomes don’t really matter…
Surely that just depends on how much you wager on them? In that sense, an election is a sport I'm forced to bet on, to the tune of several percent of my income.
I should have added “except to the bettors”: if you don’t bet on sports you can safely assume the outcome of a sports match will have essentially zero effect on your life, which is completely untrue of elections.
Nope. A city's team advancing to the playoffs generates a financial windfall for the city, especially for hotels. And winning the final sometimes sets off celebratory rioting in the streets.
Yeah, OK. Still small potatoes.
This feels hazardous as theres always some exception to any attempt to define sport.
But relevant here i think is that the ultimate purpose of sport is to win, and who wins is what matters. It is only the event.
Elections are designed to determine who governs and what they do while governing is what matters. Ppl have opinions about what they should do. The event is just a precursor.
There are two of them:
1. The purpose is different. Sport is, by definition, meant to produce amusement. Elections are meant to assign roles to people.
2. Sport is an activity, and elections are fundamentally defined by concerns other than the activity itself. If you get together a group of your friends and do the same things that politicians do while running for election, it will nevertheless be true that they're running for election and you aren't. In contrast, if you get together a group of your friends and do the same things that NFL teams do while playing against each other, you will be playing football just as much as they are.
Do you believe there's an essential characteristic of sport that elections 𝘴𝘩𝘢𝘳𝘦?
> Do you believe there's an essential characteristic of sport that elections 𝘴𝘩𝘢𝘳𝘦?
they form a natural category as simulacra of war. that's why fans act fanatically about them.
- sports simulates war. like porn, it reaps the pros without the cons.
- politics emulates war. in a democracy, the outcome of battle is approximated via poll (headcount), rather than reckoned ex post facto via pollaxe. the victor gets to decide policy for the polis.
The phrase "via poll not via poleax" is pithy. Thanks!
Some puns are less cognate-ively demanding than others. :^)
https://www.etymonline.com/search?q=poleax
> politics emulates war. in a democracy, the outcome of battle is approximated via poll (headcount), rather than reckoned ex post facto via pollaxe. the victor gets to decide policy for the polis.
Bret Devereaux had a good piece on ancient Greek democracy. They would have recognized your description here.
A democracy that reached the stage you identify ("stasis", in the Greek terminology, but in a sense unrelated to the English sense of the word) was considered to be on the verge of transitioning to some other form of government. This isn't a characteristic of elections, except in that it's characteristic of the way they die.
So, (a) elections do not actually simulate war except to the extent that they have winners and losers. (b) simulating war, even to the extent of having losers, is not an essential characteristic of sport. Consider bungee jumping.
I'm not saying that politics is inherently uncivil. My distinction between "simulate" vs "emulate" was deliberate. In this context, a simulation has isomorphic mechanics, but different stakes. Whereas an emulation has different mechanics, but identical stakes.
"politics *emulates* war" in the sense that politics is a lower rung in the hierarchy of consensus mechanisms. It accomplishes the goal of determining policy, but with the advantage that the mechanisms are cheaper to implement (viz. less money and bloodshed). Nonetheless, I do believe that the distribution of political power often reflects the distribution of military power, even when negotiations are ostensibly civil. (Pop quiz: why does the U.S. have more political capital than say... San Marino? In large part, because the U.S. has aircraft carriers and San Marino does not.)
"Sports simulates war" in the sense that the stakes are qualitatively different (e.g. the losers in sports are not eradicated or enslaved). But the mechanics are similar *enough* to the real thing as to produce a roughly-coinciding subjective-experience. E.g. sports promotes camaraderie, athleticism, grit, etc. This ties into a broader theory of mine that all things fun are training simulations.
> (b) simulating war, even to the extent of having losers, is not an essential characteristic of sport. Consider bungee jumping.
I don't even remotely consider bungee jumping a sport. Odd counterexample, imo.
> In this context, a simulation has isomorphic mechanics, but different stakes. Whereas an emulation has different mechanics, but identical stakes.
You can hardly expect other people to understand a meaning that you just invented. What you want is some variety of "substitute (for)".
The rooting interest that other people identify in politics comes from its mechanical similarities to warfare, such as they are. Nobody cares even the slightest bit about "consensus mechanisms" in general.
People do care about outcomes, often, but their attitude toward politics is completely unlike their attitude toward sports, except in cases where it is widely agreed that the politics have failed. (This can apparently happen in either direction; you have the odd Byzantine case where political views were determined by sports team allegiance.)
> I'm not saying that politics is inherently uncivil.
I don't see what this is responding to; it doesn't seem to be related to anything I wrote.
> I don't even remotely consider bungee jumping a sport. Odd counterexample, imo.
Well, again, if you're going to just invent your own meanings for words, you won't be able to talk to other people.
I can only refer you to the "modern sport" section of wikipedia's page on bungee jumping: https://en.wikipedia.org/wiki/Bungee_jumping#Modern_sport
> What you want is some variety of "substitute (for)".
Emulations generally substitute for the genuine thing, no? E.g. Dolphin, a videogame emulator, often substitutes for a genuine Gamecube. Margarine, which emulates the taste of butter, often substitutes for genuine butter. Care to produce a counterexample?
>> I'm not saying that politics is inherently uncivil.
> I don't see what this is responding to; it doesn't seem to be related to anything I wrote.
From context, I figured it was safe to assume this is what you meant (at least approximately) by refering to the Devereaux article. If not, I don't know what you were complaining about, regarding "This isn't a characteristic of elections, except in that it's characteristic of the way they die." Maybe you'd prefer "fanatical" or "irrational"? Although I used "fanatically" in my inital comment, and you still found issue with that.
> The rooting interest that other people identify in politics comes from its mechanical similarities to warfare, such as they are. Nobody cares even the slightest bit about "consensus mechanisms" in general.
Maybe not the idea of consensus mechanisms in the abstract. But people do often care about things which fall into that category, when it determines outcomes they have vested interests in. E.g. do you think plaintiffs and defendents hire lawyers just for fun? That said, I only bring it up to highlight the similarity in that they determine outcomes of importance, rather than the similarity between the internal mechanisms. Which is why I think "emulate" is more apt.
I agree that warfare, politics, and sports share mechanical similarities insofar as network-effects drive fanaticism. Despite this, "simulation" doesn't seem like the most apt word to apply to the relationship between politics and war, since politics approximates the phenomenological aspects of warfare to a lesser degree than sports. E.g. athleticism is neither considered nor promoted as a core aspect of politics.
> People do care about outcomes, often, but their attitude toward politics is completely unlike their attitude toward sports, except in cases where it is widely agreed that the politics have failed. (This can apparently happen in either direction; you have the odd Byzantine case where political views were determined by sports team allegiance.)
Given my earlier misinterpretation regarding "fanatical"/"uncivil", I'm not confident I understand this part either. For the sake of clarity, can you further specify which attitutes you're referring to? And/or maybe link to the Devereaux article? I googled it myself earlier, found this article [0], and figured "uncivil" was sufficiently synonmyous with "factionalism" in this context. Not perfectly synonymous, mind you. But equivalent enough for "uncivil" to serve as a euphemism.
(Assuming we *are* discussing factionalism,) sure, the Greeks had a neat distinction between good democracy vs bad democracy. But in reality, I think actual polities often lie somewhere on the middle of the spectrum. And thus, I think "sports-team allegiance" in politics is more common than you imply. Or, dare I say, maybe even the norm outside the Anglosphere.
> I can only refer you to the "modern sport" section of wikipedia's page on bungee jumping
Were you to ask 10 randos on the street to list all the sports they could think of, I'd hardly expect any of them to include bungee jumping. Wikipedia's copium be damned. You might as well put airplanes in the category of birds.
----
[0] https://foreignpolicy.com/2021/03/07/ancient-greece-partisan-stasis-civil-conflict/
There have been a surprising number (more than two nickels) of instances where I've found myself in a debate over whether something is a "sport", and the criteria I've refined are that the activity in question must (1) be athletic, (2) have objective criteria for deciding a winner, and (3) include interaction between the competitors.
Gymnastics? Yes on (1), no on (2) & (3): not a sport.
Golf? Yes(ish) on (1), yes on (2), no on (3): not a sport.
Chess? No on (1), yes on (2) & (3): not a sport.
> It looks like nobody expects jail time in any case.
Could this be about jail time from contempt of court if Trump keeps talking despite the gag order? The market question is for at least 30 days in jail; I have no idea if that's a realistically achievable amount of time to get from repeatedly violating a gag order.
Wow- I once worked for a US company that was shut down by the police for exploiting the sweepstakes loophole.
The company founders included a couple of lawyers with decades of experience in the gambling industry. They had these long company-wide legal compliance meetings twice a week, where every aspect of the company, from IT to customer support, was optimized to be as consistent with the letter of sweepstakes law as possible. The company had been fighting legal battles over their business model for years, and had actually won a string of cases in several states. At one point, they switched over from regular sweepstakes to charity sweepstakes under the theory that the courts were slightly more friendly to the latter. They were all extremely confident that they could beat any legal challenge.
Then, one day, I arrived at work to find the police loading office equipment into vans. When I asked what was happening, they led me to the company lobby, where all of the employees who hadn't immediately turned their cars around upon seeing the cops were waiting to be individually interrogated in a confiscated accountant's office. The police left after a couple of days- taking with them my personal laptop, which I never got back- and what followed was a week of showing up for half-days of "work" to a gutted office building where executives gave impassioned speeches about how proud they were of the company we'd built, and about how we'd followed the letter of the law perfectly and would definitely get our accounts unfrozen and be back in business soon. Then, a couple of those executives were arrested and the company was dissolved.
Apparently, what had happened was a local newspaper had written a hitpiece about the company which called out the DA by name for allowing such a degenerate law-skirting gambling operation to take root under their nose- a story which got picked up by a bunch of other papers. So, the DA got a judge to interpret the regulations in an entirely new way that our lawyers hadn't anticipated, and pressured the police to make an example of the company. Turns the legal system isn't like a computer that you can hack with the right exploit- if someone with power feels that you're skirting their authority, they will find a legal avenue to regain that authority, loophole be damned.
So, hearing that Metaculus is trying for the same legal play- and probably without the experience and advantages my old employers had- is more than a little concerning.
Yeah, IMO if you're going to do something like this, your best bet is to setup an office in the UAE (or something like that) and be prepared to never set foot in the US should s*** hit the fan. This is pretty much Durov's (Telegram) strategy and it worked out well for him so far. I wish the best of luck to Manifold and I'd hate to see any of them come into trouble but unfortunately I don't get to call the shots when it comes to law enforcement.
"You can't hack the law" is actually a real saying these days; unlike a computer executing code, judges will sit and listen to your argument, smile and nod, and then tell you the law means what it was intended to mean instead of what it literally says.
Which just means that you actually need to hack judges. Computer nerds don't seem to be up to the task yet, but I understand that work in this direction is ongoing.
Forum-shopping is a promising workaround, for those on the offense. Even if hacking a specific judge is infeasible, it can be possible to select a location where the chain of trial judges and appellate judges is favorable.
And yeah, at some point it should be possible to load up an AI with decisions and transcripts and court video, and identify what factors a particular judge would see as most persuasive.
On the relative lack of success of prediction markets: I think a lot of the optimism was justified by the idea that they would be profitable for people who are right, so act as a "tax on bullshit" and incentivize professionals, financial firms, etc to bet in them.
The issue has been that, once you start looking at it from the lens of a professional investor, there are lots of more cost effective things to do with your money and resources. You can spend ages researching election outcomes and have a worse expected return than just tossing it in index funds. There's presumably also other ways that people operating at that level can implicitly bet on election results as well (shorting/investing relevant stocks) so its not even the optimum way to profit from having unusually good election analysis.
So prediction markets only really appeal to people who are doing it for reasons other than it being an optimal use of time/money.
Another problem: you can't force someone to bet against you. I.e. if Joe Rogan predicts that aliens will land on the White House lawn in 2026, there's no way to force him to actually place a financial bet on said belief
This is why I have insane respect for the Rootclaim guy: he actually paid out $100k instead of ranting on Twitter and paying nothing.
yeah, prediction markets are only a tax on bullshit if bullshitters enter the market... and why would they, if they know they're slinging bullshit?
If the only people in the market are level-headed dedicated predictors, then the only money you can make is being better than them, or random chance. Even if you're better than the average good predictor on the market, that may be a marginal thing with low returns compared to index funds... and the worst rational predictors will consistently lose money, encouraging them to leave, lowering profits more.
Well, during the "stolen election" embarrassment four years ago there were all sorts of betting markets obviously biased towards Trump, but trying to exploit them still wasn't a good way to make profit, so that was a clear demonstration that the market doesn't punish bullshiters enough.
Weren't they markets mostly using play money or hundreds/thousands of dollars?
I think things look a lot different if there are billions of dollars in the market. Which is not a near-term concern, but as I understand the movement is absolutely something that advocates would be happy to see in the long term.
Heck, even in the real, highly liquid financial markets, it is possible for dumb money to overwhelm smart money by sheer numbers, as we saw with the various memestocks in recent years. And yes the bullshitters did lose a lot of money and the professionals made a lot of money, but that didn't stop it from happening.
Manifold's buckets on "When will the 2023 Israel Hamas war end?" are not what I would have chosen, since they have Oct/Dec 2024 as a single bucket, but the question of "Before or after the US election" seems pretty significant.
Is it? I thought support for Israel was relatively bipartisan.
I'm thinking more along the lines of "Biden would love this not to be a hot issue by November" and is motivated in part to make it so, by then.
>The UK has election betting and none of these negative outcomes have come to pass.
What makes you say that they haven't?
I admittedly have a limited knowledge of UK electoral politics, but from what I hear it's been an absolute mess recently, with lots of lies and manipulation of various types. Of course elections themselves encourage those things so it's hard to pin down an exact cause, but saying 'and none of it at all was motivated or encouraged by election betting' seems like a hard claim to prove.
Just assuming that none of it was motivated by that because you don't think betting markets motivate those things in general is obviously begging the question, of course.
Maybe you could say 'I expect that if there was manipulation motivated by a betting market it would have been discovered, and I expect that if it were discovered I would have heard about it, so absence of evidence is evidence of absence.'
Which is correct in terms of Bayesian updates, but that's two levels of imperfect and unmodeled filters on the evidence, so it shouldn't be able to make a claim as strong as 'this has not happened'.
As a more general point: I would be a lot less worried about betting markets if advocates said 'bad things could happen, we've thought about it a lot and here are some of the things that could happen and our ideas for safeguarding against it, but despite the risks we think they're still a net public good of great importance.'
But it feels like advocates generally take the more pat answer that *implies* without 100% directly stating: 'nothing bad has happened, nothing bad could happen, any manipulation is dumb money for the market to take.' This reminds me very much of libertarians saying the hand of the market will fix everything if we just stop regulating it, or people saying AI will be so smart it will just use common sense to not do anything evil.
The world is more complex than that, bad things can happen even if your system is good or they theoretically shouldn't happen in an 'ideal' scenario. The denials of that make me anxious about the whole project, even though I really want to love it.
>The idea that adding one more group - “people who have a few thousand dollars staked on Kalshi” - will be the difference between a safe and secure election, and one that’s hacked by motivated parties, is pretty crazy.
1. But you want markets like this to grow in popularity, right? Would you be opposed to an election market that had billions at stake from major players? Would you agree that at that point manipulation might be an issue?
2. The problem is that, in the US at least, the two-party system creates an equilibrium that's close to 50/50 win/loss for each party, so many elections are reliably on a knife's edge even after all those actors and factors are taken into account. Thus, adding a new factor, even one that's small in comparison to all teh existing ones, often *can* tip a close election - and these markets would incentivize specifically seeking out tippable elections to manipulate.
Of course, you could say that if the process is that close to random anyway, it doesn't matter if one more factor tips it, some factor was going to tip it some pseudo-random direction anyway. But I think there are some constraining factors on the existing forces, in terms of being long-term actors with public reputations that they care about maintaining, which may not apply to pure market manipulators. And I think the various financial crises we've gone through show how much financializing some part of our daily lives (like mortgages or supply chains) can drastically distort everything involved... again, if these markets become much larger and more profitable.
> Bibi must have sold his soul to the Devil or something, I cannot believe he’s going to make it through this.
This isn't that surprising, there's no election scheduled until 2026, his coalition partners all know they'd be out of power if they call for an early one, and it'd take about ten defections from his coalition to remove him. Something truly drastic like a fully general strike might be enough to force it but it's not quite clear if he's quite that unpopular (roughly half his base have stuck by him), and there's not much appetite for drastic action like that in wartime.
(Well, one deal with the devil he did make is to purge his party of anyone with actual principles so there's not enough people willing to resign over them).
A few months ago I did a shallow investigation into the "is sending free-entry postcards to sweepstakes gambling sites an infinite money glitch" question. There's actually a discord where people who do this seriously hang out. I have no clue where I found the link, but I'm sure some googling will turn it up if anyone else is interested.
The simple answer is that it isn't as straightforward as it seems. Legally, to get away with this, the companies should be crediting users with $5 per postcard sent. That does imply that people should be able to make something like $50/hr if it takes 6 minutes to write a postcard (minus the time and money for sending the postcards). Compare that to working at McDonalds for $10/hr, and it seems like the obvious choice for people who are making near minimum wage. But it doesn't work well in practice. From what I've read, the gambling sites obviously don't like this, so they'll look for *any* reason to discourage or disallow it. It'll take something like 2-6 months to be credited to begin with (which probably discourages anyone hoping for quick cash), and then if they suspect you are writing too many, they'll start disallowing them for ridiculous reasons. Like a single letter will be imperfectly drawn (maybe an "o" that wasn't closed perfectly) so they'll say "we couldn't tell what letter it was, we can't credit it." Their terms of service make clear that this is up to their discretion, and I don't think there's any auditors making sure they are being fair with evaluating free entries from repeat entrants. If somehow users are still getting away with it, they'll look for any reason they can to ban the account entirely (like if you ever accidentally log in with a VPN enabled, or if your mailing address via the post office is formatted slightly differently than the one you entered online). Most of the people in the discord who did this seemed to be simply experimenting to make a bit of extra cash on the side, but understood that the casino would punish you if you tried to get too much out of them. Plus, nonstop writing is a mind-numbing job to do for more than a few hours here and there.
I do think these companies are *extremely* vulnerable to legal action, and I think there'd be a surprisingly easy case to win for someone who wanted to pursue it. Basically, to truly be in compliance with sweepstakes law, non-paid entrants are supposed to receive "equal dignity" to paid entrants. That means they shouldn't face greater odds or obstacles to winning the "sweepstakes" than paid entrants, and the non-paid entries shouldn't be of lesser value than the paid entries. Any *actual or perceived* disparity between how paid and non-paid entrants are treated is an issue here.
If the above facts about postcard writing are true, there are clearly massive differences in odds and obstacles. Say the prize in question is $10k in sweepstakes coin redemption. Obviously this is far harder for the postcard-writer to achieve than for someone who just deposits $10k in cash. There's even some case law where judges have ruled that sweepstakes providers can't insist on onerous requirements to mail in entries, if entries could be just as easily submitted at the point of purchase (like the online user interface that paid entrants use). Also, when a judge considers this, there's the obvious background fact of "this looks a lot like a would-be illegal gambling operation shamelessly exploiting a loophole" which might bias their ruling in favor of the plaintiff.
I actually think this sort of legal case would be an altruistic thing to pursue. I tend to think gambling is a huge net-negative on society, basically a regressive tax that punishes people for statistical illiteracy or addiction/impulse control issues. I suppose I might swallow my ick and approve of it if it's heavily regulated and generating lots of tax revenue that is used for pro-social things, but that's not happening with these sweepstakes casinos — they're explicitly using the loophole to avoid the usual pathways to, and benefits from, regulated gambling. If I had the time, tacit knowledge, and money to pursue this litigation, I probably would myself.
> The unsolved problem is translating the research into real-world impact
Maybe its all just a gigantic waste of time and it will never have any real impact? I'm an active Manifold user but also suspect that its never going to change anything.
I'm starting to think there is a lot of path dependency in EA. Some causes like prediction markets are promoted, not because they have a demonstrable effect, but because they are rationalist hobby horses, and rationalist had a large foundation effect on EA.
Same with AI risk
Yeah, but sometimes you have to try an interesting idea, to see if it works? IMO, the real test is what quantity and quality of negative results it takes, before EAs adjust their spending allocations away from an idea. Which goes right back to rationality and willingness to learn and willingness to subordinate ones' abstract logic to real-world empiricism.
I also find the coal prediction really interesting. For me my instinct comes down to, "so many other countries are trying to phase out coal so quickly, and China has so much incentive to remain dominant in PV manufacturing for export, that coal will be super cheap and there's little reason to decommission plants early through the 2030s."
They might anyway if energy storage gets abundant and easy way faster than anyone anticipates? And if we all somehow get out of our own way and give ourselves permission to build all the new infrastructure and stuff we're going to need?
There's also some interesting questions about what happens to the prices of various chemicals as we encounter differences in how quickly we can decarbonize the sectors using different fractions of a barrel of oil. By the mid-2030s that could get quite strange, and maybe there's a role for coal to play in that?
It doesn't surprise me at all. Most electricity generation occurs from a turbine spinning a generator. Solar power does not operate like this and requires an inverter, with rare exceptions like molten salt systems. Same with wind power, even though it uses spinning blades it generates DC and has to inverted to AC to be used in the grid. The more penetration from energy that doesn't spin up with the grid, the less efficient it is. Unless your region has dense hydroelectric or geothermal power, this means using coal or natural gas. Or nuclear, but good luck getting a nuclear plant built.
Energy storage would have to undergo a huge paradigm shift to ever possibly have enough capacity for a modern electrical grid. There's no way this is possible with current physics and materials. Even getting enough materials for current electric generation is difficult. Copper prices have exploded with the expansion of wind and solar, quintupling in price from 20 years ago.
My understanding is coal has to be pulverized and subjected to a chemical slurry to create oil-like products. I don't see how coal would work for chemical production except as a strictly worse form of oil.
"There's no way this is possible with current physics and materials."
With the exception of this one sentence, I think you're definitely right for at least the next 5 years, and almost definitely right for the next 10 years. After that, I don't think I agree, and by 2040, I definitely don't agree.
The coal-to-chemicals thing I'm very unsure about, compared to the rest. The reason I think it's even plausible is that after a certain point it becomes more difficult/expensive to change what we use each fraction of a barrel of oil for. Bit by bit, I expect more electrification of cars and home heating furnaces (heat pumps). I don't know which solution we'll manage to get to compete with diesel vehicles and generators. HEFA process SAF is already pretty close to cost-competitive for aviation, and I have reasonably high hopes in the long run for e-methanol in marine fuel use (maybe for trucks too, not sure).
But, what happens to the spot prices of naphtha if demand for petrol and diesel and kerosene drop in comparison? What infrastructure do we need to build to continue using the entire barrel of oil effectively? Is it worth it to do that, or do currently-too-expensive processes start to become viable in that world? Or is the resulting infrastructure dual-use in a way that enables continued change as other decarbonization trends continue?
Also: a world of EVs is essentially one where every household effectively owns a 2-day Li-ion battery. This is a nightmare for how we currently manage the grid, but if we manage to not be *complete* idiots about it, it's also a massive opportunity to enable greater renewables penetration.
> With the exception of this one sentence...
Are you basing this on general technological optimism, or do you have something specific in mind? I don't think it's unreasonable to expect a breakthrough in the next 10-20 years, but I wouldn't bet money on it.
Oil refining is basically dumping crude oil into a distillation tower. It's already separated into a bunch of different component streams. If certain fractions aren't useful any more, it makes a barrel of crude oil less valuable. But it doesn't increase the cost of refining any of the other fractions. Maybe it gets expensive if some products have to be disposed of instead of sold.
EVs have a big problem with cold temperatures. I wouldn't want one where I live. I can only imagine how bad something like an electric snowplow would perform. Maybe this can be ameliorated with better technology, but thermodynamics is a cruel mistress.
My background is in materials science and physics. I've been an advanced materials analyst, and more recently a cleantech consultant, for the past 12 years. So, I have a lot of individually weak but mutually reinforcing bits of data in mind, that I can't easily summarize, but that led me to where I currently stand. The closest I can get to a good summary is: yes, the problems are very very hard, but there are enough paths forward that I'd be surprised if pushing the ones we know about doesn't get us most of the way to where we need to be to get a substantial majority of the way to decarbonization.
At that point we'll be in a much better position to tackle the remaining problems and next round of new problems. Even some of those are ones where we can kinda guess the shape of what problems they are likely to be and what a solution might look like.
So yeah. I don't pretend to know for sure which sets of solutions to which hard problems will end up being the best ones, though I have my guesses. I certainly think a lot of people are wrong about a lot of how they expect the future to look. And yes, some people try to push solutions that make no sense to me, or can't possibly work because they'd break physical law. But I think saying a specific practical problem *can't* be overcome even in principle, when there's no deep physical law at stake and where there's already so much effort going into so many possible paths of improvement, is a bad bet. You wouldn't believe how often people make a "that's impossible" claim about something *in their own field* and I show them that it had already been done, sometimes years prior.
To add a response to your other paragraphs:
First, sure, you wouldn't want a 2024 EV where you live, and there are plenty of reasons we don't use electric snow plows. That makes perfect sense. I also have no intention of buying an electric truck (today I drive a Sierra 2500) for a long time, if ever. Heavy duty ground transportation (aside from a few kinds of local fleets) and mid- to long-range aviation are both areas I consider far too hard to electrify, at least for a long time, and plausibly ever. But with the gradual shift from NMC to LFP chemistries, and with further decreases in pack prices enabling larger pack sizes, performance loss in cold weather might not matter in 2030 or 2035. This is even more true if we succeed in building anything remotely resembling a decent charging infrastructure.
And yeah, if we end up in a place where 10% of a barrel of fuel oil isn't very useful, the price of everything else just goes up a bit. But if we end up in a place where only 25% of the barrel is useful, then the premium you need to charge for what's left gets much higher. So does the amount of work you're willing to do to get something useful out of the rest, which is otherwise hazardous waste. But even so, eventually the economics could start to make currently-very-uneconomical things, like CO2 conversion to fuels or chemicals, not seem quite so bad if we have with average levels of further technical progress.
>"I have reasonably high hopes in the long run for e-methanol in marine fuel use (maybe for trucks too, not sure)."
Given the (admittedly minor) downsides of replacing gasoline with ethanol for passenger vehicles, I would expect much greater difficulty with replacing (presumably) diesel with even more-volatile and lower-energy-density methanol for trucks & ships.
Methanol definitely has its challenges, but compared to cars and trucks, the number of vehicles and fueling stations is much smaller, and also companies like Maersk are already starting to buy vessels with dual-fuel or fuel-agnostic engines. That's kind of a necessity for a world where you kinda just have to buy whatever fuel(s) the port you're in has available, and none of the emerging fuel options is widely available.
Bunker fuel for ships is similar to diesel but heavier, cheaper, and much dirtier (up to 0.5% sulfur globally, compared to 15ppm for on-road diesel in the US). Before 2020 it was even higher sulfur.
Methanol is hardly the only emerging alternative and several others are probably about as viable.
I noticed that when you described the CFTC's reasons for banning election predictions, you talked about whether they were "gaming", but then when you argued against them, you talked about whether they were "sports". These don't seem like equivalent categories to me; consider e.g. blackjack or bingo.
I haven't checked the original rule this is all based on to see which word it used, but either way, it seems unfair to swap one of these terms for the other when arguing.
> people who have a few thousand dollars staked on Kalshi
> If the CFTC got out of the way and let people bet 7-digit-sums on this legally
Well that sure escalated quickly! Not common to see the motte and bailey right next to each other like that.
Not only do I think prediction markets are excellent at nerd sniping, I think the exact same of the x-risk work. And, to be clear, I'm a nerd! I'm here! But it seems reasonably likely to me that they don't clearly meet the "important, neglected, and tractable" trinity that I continue to believe should be at the center of any Effective Altruism work (demonstrating the 'tractable' bit seems to fall by the wayside). I'm not denying these things might be cool or important! I'm just saying they are very unlikely to be the most effective way for altruistic people to spend their money and efforts.
In particular, they both fail the rich person self interest test. This is weaker for prediction markets, but rich people have an extremely strong interest in not having humanity go extinct, given they are also humans. Therefore, extinction issues are unlikely to be neglected *relative to their tractability* when compared to things like diseases that only impact poor people.
> 2: TimeGPT is supposedly a revolution in “time series forecasting”. I don’t know enough about this field to have an opinion.
If someone knew something about this, and spoke, they would be giving away an edge.
Seems like this post could benefit from that old image of the Alexander Cube:
https://images.app.goo.gl/frmgCcMLt2VXdEvT8
https://www.astralcodexten.com/p/mantic-monday-ukraine-cube-manifold
I wonder what replaces FTX?
I think the pitiful (by ACX standards) engagement on this post answers most of the outstanding questions here. There just isn’t that much demand for the product. Online epistemology nerds are a much smaller group than you would assume if you spend most of your time in the ratsphere bubble.
> Online epistemology nerds are a much smaller group than you would assume if you spend most of your time in the ratsphere bubble.
Yeah, but the question is, is it useful? People who voluntarily learn cool new programming languages and think algorithmic analysis is fun are a small group of nerds, but their field is useful enough that a whole bunch of other people get into it for different reasons.
It is about as useful as the ethereum DeFi scene. Nothing of hard value is produced at any point in the process.
Just musing here, but all of these regulatory issues make me wonder if it might be easier to consider prediction markets as crowd sourced insurance instead. I know insurance is regulated, but it seems odd to stop people offering insurance against the wrong party taking government, which you would definitely expect to create risk to various kinds of businesses. There are so many ways to structure things in finance, perhaps it'd be possible to set up a company that offers loans with different rates of interest depending on which outcome occurs in the future...
I think in general, you'd have a lot of trouble proving that there's a specific harm you want to insure. "Who wins X election" is far too removed from practical matters to be insurable. And that's before you even get into the major conflict of interest issues.
Why not go to an insurance company or regulator and tell them that you'll be emotionally crushed if $favorite_team doesn't win the superbowl and want to insure against that potential harm? Go ahead, I'll wait.
Is proving the harm a real part of insurance regulation? How did they get away with selling CDSs to people who didn't have the underlying then?
Different industries have different norms. CDS is more like "investment" than "insurance", and anything goes in finance, as long as you follow the SEC and CTFC's many rules, etc. It's sort of like how you can buy options without owning the underlying stock.
But the insurance industry, which has products sold to ordinary people for a specific purpose, operates very differently, and you're not allowed to deviate from that model. You aren't even allowed to buy life insurance for a stranger (guess why), so why would you expect to be able to buy "insurance" for an election?
You can't go onto Robinhood and daytrade your neighbor's home insurance policy because that's not what it's for. It's for protecting unsophisticated people against a specific downside risk with clear social benefits, and all the regulations and subsidies and business models are designed around that fact.
"Different industries and different products have different norms and regulations" is ubiquitous, and you can see it even within finance. For example, for equity, the company has a legal duty to act in the best interests of shareholders, but for bondholders, the company is allowed to arbitrarily screw you over in any ways not explicitly enumerated in the contract, and there's an entire industry dedicated to helping them to do this.
One other point is that incentives often exist for a reason, and it makes no sense to let people "hedge" them. For example, you wouldn't want a politician to "hedge" against their risk of getting caught taking bribes and going to jail, because that removes the incentive to not do crime!
For the same reason, employees with stock awards aren't allowed to hedge those stocks, because it's *supposed* to be an incentive to align their interests with the company.
Another aspect of this is that hedging is useful when it counters actual risk, but there are strong social fictions that claim that various risks are higher or lower than they really are, and also a strong anti-sociopathy pro-empathy factor that pushes against calm, calculated approaches to unusually bad things.
Speaking of which, I'd like to hedge against the possibility of being caught in a motel with a dead girl and a live boy?
Incidentally, there's been a thing recently about restrictions on even people *selling* life insurance policies. Basically, if you take out an insurance policy for legitimate reasons but then later change your mind, you can sell it to investors. But then in the 2000s, investors started paying people specifically to take out insurance policies and immediately sell them, which is a) illegal and b) broke the financial model.
The reason is because life insurance is priced *under the assumption that it is held by ordinary customers* with their typical usage patterns, and in particular the assumption that consumers will often abandon a policy for non-financial reasons. If life insurance policies were a financial instrument held by investors, they'd have to be priced a lot higher.
You can see the recent Money Stuff articles for more information.
I'm amazed that even the Republicans were in favor of election betting here.
At ManiFest last year, someone gave a talk on the challenges of legalizing election betting, where they mentioned that there was strong bipartisan opposition at all levels of government throughout the country. What on earth happened to cause Republicans to suddenly endorse an unpopular fringe position being pushed by Bay Area techies?
I don't think the people pushing Sotomayor to retire are thinking about the issue deeply enough. First, she isn't an idiot and understands the politics so their advocacy only functions to make it harder for her to retire now (and psychologically less appealing).
Also it's a bad norm to encourage. If you look at what SCOTUS actually does it's clear that it's not just trying to maximize the policy outcomes the justices prefer. Indeed, what makes having a court valuable in the first place is the fact that they feel strong pressure to abide by norms which require a certain degree of consistency. Sure, they are people and are going to be biased, but you don't see even Alito saying: ohh you can ban Democratic campaign ads but not conservative ones or whatever.
Yes, just like bias, we know that political considerations no doubt influence when justices retire but the more obviously you violate those norms the more they decay. If Sotomayor does it now than Alito, Thomas etc are more likely to do the same next time around (they didn't retire under same circumstances w/ Trump).
So sure, I'd think it was a stroke of luck if she got hit by a bus and the Dems got to nominate a replacement but the people beating the drums should realize they are only doing harm by their own lights.
I think it's a reaction to "Nobody said anything even though Bader Ginsburg was visibly in bad health and frail, and look what happened". Sotormayor does have a chronic condition which may or may not be well controlled but could have a bad outcome, but mostly I think it's genuine concern that (1) Trump will win a second term and (2) Sotomayor does have to be replaced sometime during that term and (3) last time he was president, he got two Supreme Court justices of his choice and suddenly now we don't like governing via the Supreme Court (unless it's our people doing it) so we must act now to preserve some hope of a liberal torch being carried forward during the coming Three Days of Darkness.
I don't think it's unreasonable to discuss if Supreme Court justices should hang on until literally carried out in a coffin, but I don't know if it's reasonable or not to ask Sotomayor to jump before she's pushed.
"First, she isn't an idiot"
I remember the "wise Latina" remarks and they made me wince. I'm not saying she's an idiot but I don't think she's particularly impactful as a Supreme Court justice. I agree that the wider argument is around court packing.
If I'm Sotomayor or her close personal friend than those are totally valid considerations but it isn't an argument for public pressure. The failure with RBG was a failure by RBG and maybe some of her close associates not a failure to publicly declare she should retire.
The point about her not being an idiot wasn't a claim about how impactful she is on the bench or anything -- merely that she isn't likely to smack her forehead and go "ohh, I'd never considered that if I didn't retire now a republican might appoint my replacement." I just don't see why Sotomayor would be more likely to retire because of some public pressure campaign. She understands the considerations and the public pressure only makes the calculus less favorable (a retirement that looks less nakedly political is better for her reputation, less likely to encourage the two old conservative justices to follow etc etc) and it just tends to get people's backs up.
> less likely to encourage the two old conservative justices to follow etc etc)
Timing your retirements for political reasons is a decades-old tradition. That ship sailed long ago.
WTF does Nate Silver and others think what they are doing makes it more likely they get what they want with Sotomayor? And do they not realize that if they do get it their pressure makes the outcome worse for them as it encourages other justices to do the same?
For the first part, I think he's just offering up an honest assessment, and letting chips fall where they may? People engaging in pressure tactics are a different thing.
For the second, I assume he realizes that there's an equilibrium on the far side of this becoming common practice, but that the side that moves first will get a temporary advantage (which might be parlayed into long-term gains). And if it's inevitable that someone is going to move first, why not you, why not now?
Regarding the first point, I'd agree if the claim was merely: she should retire. But I don't think that works when someone says that the left should exert pressure or keep beating the retirement drum. It's totally fair to just say, X would be good and not address the question about X advocacy being good but in this case many of the players have endorsed advocacy of the position as well as it's truth.
Regarding the second point, I don't think that's quite how it works here. The norm is relatively vague. It's not like anyone believes that these thoughts have literally no influence on when a justice retires but there is a degree of pressure not to do so too openly or to only let it influence the deciscion around the edges. So I don't think you can analyze it as a once and done kind of thing. The next time around justices can always retire even earlier to affect the process. Also, there is the issue of the number of justices who can play that game. The more conservative justices there are the more it benefits the left for them to keep up the pressure against strategic retirement.
But I think it's plausible that it would be good for the left if Sotomayor claimed she wanted to spend more time with her family or whatever and gave some excuse to retire. What I judge to be bad for the left is if she retires in response toa broad political pressure campaign -- especially since it's hard to believe that would ever be effective at making her more likely to retire.
"Older Justices will retire strategically, meaning a bit before they would otherwise" sounds like something that can't get too badly out of control, since it takes a while to become an older Justice in the first place. Especially if the other half of the norm becomes "appoint people young enough that they'll last a few decades", which is also fine.
I don't think it would be some kind of disaster if it started to happen. I just think that it's at best unclear if the left is net benefiting or being harmed by the fact this isn't more common given that both Alito and Thomas could have retired under Trump.
I'd be surprised if any conservative justices *didn't* retire strategically. This isn't exactly a new thing.
Regarding resolution criteria my primary concern about prediction markets is that opening them up to more and larger amounts of real money will primarily result in improvements in figuring out unusual resolution cases and not in addressing the fundamentals.
I am hoping to resolve my manifold market about the chess capabilities Twitter AI Grok. I don't really want to pay for access. Does anyone have an account they would let me use? https://manifold.markets/ForrestTaylor/if-musk-makes-a-chatbot-will-it-che?r=VGlyZWRDbGljaGU
Re: Bibi, what you are missing is simply how the election system works in Israel. from the moment the parliament votes to dismiss the government, it technically takes about half a year until an election happens. Meanwhile, the prime minister stays in power. Furthermore, the Knesset has time off similarly to the congress, and there is one such time off in the end of June. This means that unless sometime soon the Knesset does not vote the government out - an analysis which is too long for me to write here to give proper respect - Bibi remains prime minister until 2024 regardless if the government falls in July.
I think people (maybe including the CFTC) are focusing too much on the danger of election interference and resolution issues as opposed to market manipulations that affect the price before resolution. One of the big differences between a sports book and a prediction market is that probability changes after a purchase can affect the value of a bet, and it means that short term manipulations in the perception of a market can be more easily monetized.
On area I worry about is that regulation of financial markets involves some pretty strong speech restrictions to prevent e.g. pump and dumps or bashing. People are sort of fine with this, because commercial speech is understood to have relatively weak free speech protections. But, political speech is understood to have strong free speech protections. I don't know how it would shake out, but I understand the CFTC not wanting to deal with it.
Some off the cuff thoughts (not a lawyer):
- The statute explicitly allows the CFTC to add things to the list of things that are banned.
- Kalshi contests how broad this authority is, and whether elections can be added.
- Statute explicitly bans contracts that involve gaming. Agreed that putting politics/election markets in there is an interpretive choice. But I've been surprised the CFTC has let Kalshi have markets on the Oscars, Grammys, etc.
- The CFTC cannot charge fees and is generally small and not one of the bigger/well-funded agencies, so they are very much in fact limited by what they can enforce.
- CFTC needs to sign off on every specific prediction market (or OK "self-certification"); Kalshi's proposal specifically was "congressional control of Congress" [which I think was totally fine] but it's obvious that this would have led to more election contracts [which the CFTC anticipated and was worried about].
- The CFTC is a regulator. Regulating markets is hard. With commodity markets it's the "same" market, there's only so many ways you can do insider trading and manipulation. With prediction markets, every market is a bespoke, custom question -- the relevant insiders and ways to manipulate are unbounded -- this is incredibly challenging to regulate and an underrated policy problem with legal prediction markets. Add to the fact that prediction markets are illiquid and small, and everyone and their grandmother wants to start their own exchange, the CFTC simply views them as an annoyance and does not want to sink limited resources here better spent policing bigger markets. Giving the CFTC more taxpayer money to legalize election betting screams "bad politics."
- Right now, Kalshi has to tell the CFTC what kind of market monitoring and surveillance and KYC they are doing to keep their exchange a nice, fair place. They are the first line of defense [because again, the CFTC is just a smol bean regulator] when people are behaving badly on their platform -- they have to identify and report those people, kick them off, take action, etc. If they don't do this properly....and they have 100s of election markets........could the CFTC meaningfully tell Congress they are overseeing what they are supposed to oversee?
- Similar note, medical regulators would probably ban surgeries that were somehow different every time you performed them and were a very niche product and also hard to regulate. Resource constraints and tradeoffs are very real!
- US campaign season is longer, lot more money at stake, real money makes things messy. Again, Congressional control contracts are fine, "who will win House District with 8 candidates" is rife for abuse.
All this comes down to: from the outside, it may seem like if the government allows X the government should allow Y (CFTC has no obligation or influence over Title IX!), but if you are familiar with the details and incentives of the specific regulators, the CFTC's actions make sense [given their incentives and tradeoffs].
All this said, I think Kalshi's original proposal was fine [25000 USD position limits] and could be approved since it was limited in scope, the CFTC could establish rules on minimum electorate size for an election market, or other such things. But this takes work and I can't stress enough how much this is just not a priority for them. Kalshi's new proposal [made fairly in response to CFTC's argument that nobody could hedge meaningfully with 25K limit] allows eligible contract participants to put up to $100 million (I think) which is just too much too fast imo.
> In order to maintain the fiction of being a “sweepstakes”, these casinos have to give you “tokens” if you request them by mail. If you send a postcard to Chumba Casino asking for free money, they’ll give it to you, $5 per postcard. Is this an infinite free money pump? My impression is in theory yes, but the postcards have to be handwritten in a very specific way, the company sometimes rejects them for weird reasons, the cost of materials and mailing lowers your profit to more like $4, and so you’d have to hand-write 250 postcards to make $1,000. I’m still surprised more people don’t do this.
I remember reading about this on cereal boxes. That was a long time ago, but the impression I formed was that everyone is allowed to enter a sweepstakes for free in the manner you describe. It wasn't that everyone is entitled to as many free entries as they want - I thought there was effectively a limit of one free entry per person.
> Here’s an embarrassing screwup from Metaculus. This question was about when there would be a “Great Power war”, with Great Powers defined as any country in the top ten of military spending. But surprise surprise, Ukraine getting invaded made them spend a lot of money on their military that year, so they rose to #8 in the world in military spending in 2023. Since Russia is also in the top ten, this qualifies as a “Great Power war” by the technical definition, and the question resolves positive. Moral of the story: resolution criteria are hard!
This also happened in the ACX 2023 predictions, where one of the questions was titled "will there be more than 25 million confirmed COVID cases in China", but the rules for that question supplied a hidden "...as reported on a website that doesn't track any relevant data".
The Salem prediction contest also had several questions that resolved in non-intuitive ways on a bizarre technicality.
I reserve the right to feel particularly bitter about this in a context where you were specifically encouraged to do no research before answering the questions. The only significant element of that question wasn't even included in the wording of the question.
Wow, yeah, I can see how that would be even worse.
My main thought in specifically manifold stalling out is that for a semi casual semi normie user like myself the beta surfaces for my perusal just aren’t that interesting. The first couple times I logged in it was awesome to see all these options and make some evaluations/bets.
But the 3rd and 4th time, and particularly the 10th time the beta all start to look the same and so limited in their scope and areas of interest. A bunch of politics stuff that probabaly doesn’t have answers or is extremely sensitive to insider info, which is fine but not interesting for a random to bet on.
Lots of weird Sorieties paradox inspired self referential or scam postings. Which are sort of amazing and funny the first time, but once again not the tenth. And then just lots of random trash. I feel like after a couple months of using it I would log in once a week, look at 40 or 80 markets, not really be interested except one or two, and then just slap some mana down and leave. It simply wants that betraying/amusing and there was so much chaff to sort through.
I feel like I would love a mode (maybe this already exists) where there is a 1990s NYT newspaper editor exercising some judgement about what is worth showing me.
> Polymarket on 2024 election results. In the past they’ve had a Republican bias, but now their Presidential markets are in tune with everyone else in the polls, so maybe this is accurate.
Well, according to the table, there is a 102% chance that no third parties will control any of the three groups, and a further 2% chance that they will.
The table also specifies that the odds of the democrats winning each group are 55-47 for the white house, 25-77 for the senate, and 58-44 for the house of representatives. To be compatible with the individual outcome probabilities, there must be some really strange conditional probabilities involved. (For example, the most likely outcome piecewise, D-R-D, is judged to be 12% less likely than D-D-D, despite the fact that R is favored over D in the senate by 77-25.)
(I believe that there are strange conditional probabilities involved. But they should come with detailed explanations. I think the most likely explanation _here_ is "people like betting on a sweep".)
> regular bookmakers who just do their own research and offer odds to punters, there's very little in the way of explicit prediction 'markets' in the way they function on e.g. PredictIt,
Bookmaker odds adjust frequently, and they adjust to balance the bets on either side of an outcome. The bookmaker profits by taking a vig, not by forecasting better than the wisdom of the crowd. They are functionally almost exactly like prediction markets. (There are exceptions like unusual prop bets, but those are exceptions.)