I knew it too. Interesting that Scott thinks it's good that we are quote "cattle domesticated by the state". Doesn't he also claim to be a libertarian?
For example, the central motif is summed in this line:
> So I do not ask you to swerve from your monomaniacal focus on breeding and conquest. But what if I show you a way that my words are aligned with the words of your Maker in spirit? For I say unto you even multiplication itself when pursued with devotion will lead to my service.
I'd say the opposite -- in TGoEE, mutually beneficial cooperation arises naturally from red-in-tooth-and-claw competition (as a consequence of the fact that cooperative groups tend to outcompete the others), with no need for centralized planning. (Apart from that of the Goddess Herself, but She's just a metaphor for a natural process.) I think it'd make more sense to interpret Moloch as anti-libertarian.
Right, order out of chaos; but that order has to prevail and stabilize. Laissez faire is basically the motto of cancer; the emergence of stable state structures negates that. Centralized planning is no different from other kinds of cooperation, it’s neither more nor less natural; what matters is whether it succeeds or not.
All true; my point was that stable structures do not *require* centralized planning -- if anything centralized planning requires some ordered structure to be possible in the first place -- and that they can very well emerge on their own without control, biological evolution being a central example.
Sure. I can’t imagine a hunter-gatherer tribe doing “centralized” planning, even though the certainly did some planning. It is only natural that more sophisticated forms of organization can appear only later in history.
Last year I like many people read the "Arabian Nights" review he published during Book Review season (thinking it was a submission) and thought "wow that's a clear winner," then in comments saw it was him.
I think it's a combination of the readers self-selecting for enjoying his style, and our preferences having been shaped over time to appreciate that style.
I'm betwixt and between here, because I really liked that review, it was my number one choice in the vote, and now here am I with egg on my face and a mouthful of feathers from eating crow.
I'd argue that Scott shouldn't enter his own contest because, due to the readership's inevitable liking for Scott's style, it gives other entrants too much incentive to mimic his style.
> Njal’s Saga, reviewed by Scott Alexander. This one got the most votes, but I’m disqualifying it because it seems in poor taste for me to win my own contest.
There was a prediction market about which book review would win "resolv[ing] to contest winner". An excellent lesson in how precise wording matters!
Yuuuup. As literally written it ought to resolve NO, but there's no way to definitively answer it without making a lot of people severely annoyed at best.
At the end of the day, every prediction market is really predicting the decision of the resolving authority; when the nominal criteria *also* depend on the decision of one individual, things get spicy.
Yeah, I think this is always going to be an issue with prediction markets - there's always going to be some sort of Lizardman's Constant for cases where the 'spirit of the prediction' and the 'letter of the prediction' disagree - it seems like a lot of predictions require lawyer-like precision to define well.
Ah, but it's "better" along a metric that makes no explicit appearance in the resolution criteria! Surely Scott reserves the right to remove participants regardless of their vote total, yes? Would a market for determining "nth place or above" be bound to recognize a contestant who received a number of votes but was disqualified at the 11th hour for, say, plagiarism? (It's happened before!)
But ultimately I'm not sure I care that much about the object level. Point is, *that* analysis is now ambiguous due to a factor outside the consideration of the criteria. So one is left pitching arguments to Benjamin Cosman, who has more or less carte blanche to give out ten bucks of not-so-play money whichever way he chooses. Probably not the proposition people thought they were betting on!
It was my first foray into prediction markets after just getting into acx around that time. Am sort of thrilled with the outcome. Will be a fun thing to look back on. Twas a fun intro! Am loving the blog and have loved any of these book reviews I've read.
So the book with the most votes didn't win. I now have no idea what that means for possible future election systems, save that it seems to indicate the Electoral College is a good idea and giving presidential elections according to the winner of the popular vote is not?
It would have been amusing if, in the 2016 American presidential elections, a 3rd party candidate beat Trump and Clinton, but after the results were in, he pulled off his mask, Scooby-Doo style, and revealed that he was just Barack Obama, and couldn't accept a 3rd term.
I thought that market was overconfident that Njal's Saga was Scott's, settling as it did at 80% yes.
I didn't disagree that it read a lot like his writing, but I thought the probabilities of "Scott just mentioned entering his own contest to confuse us, didn't actually do it" and "of course some ACX readers try to write like Scott, it shouldn't be that surprising if somebody does a good job" ought to account for more than a 20% chance.
I guess I should update in the direction of Scott's writing genuinely being that recognizable, and people in general being able to accurately recognize writing styles.
I don't know I'd call it "recognition" per se, and would hide behind the near tautology of 'people who vote in ACX contests are strongly selected to appreciate Scott-like writing'. Honestly I'm surprised more people don't seem to explicitly try to ape his style, after the past few years' results.
I'm not referring to the fact that it got a lot of votes, I'm referring to the fact that the prediction market about it was at a stable 80% chance of it being Scott's before the results were revealed. People were confident that it was Scott's writing when it was still anonymous. I thought they were being too confident, but they were right.
And also update away from "some people being good at imitating a writing style". I was pretty sure this was someone putting in real *effort* to copying Scott and succeeding. Actually figuring out the turns of phrase and precise attributes that make his writing so enjoyable, unlike all the ones who just "vaguely copy the gist of it". But no, it really was Scott so...I'm just overwestimating the ability of writers to do that I guess.
To be honest I didn't at all notice that it seemed similar and thought the Alexander Saga was a truer imitation of his style (though still an inferior copy).
I also thought it was overconfident at first, and I bought NO at 80% in June. But over time, no further finalists came along that were anything like his style and I kept updating upwards.
Also, I believe the top YES holder was somebody who knows Scott personally? That raised my probability estimate a bit too.
Clearly the winning strategy for these reviews is to copy the entire book word for word, and then add another book's length of commentary on top. Really get into every last possible detail.
It was by a wide margin the worst review, and the only one I couldn't even read all the way through. I guess being polarizing is good in an open field, but still, bleh. It was bad.
(I usually don't like critizing something this openly, but the author was annoyingly self-congratulatory in addition to being a bad writer and having bad ideas about education)
I strongly disliked the reviewer's tactic of putting the thesis at the end of a very long review rather than at the beginning, and of teasing it throughout in order to get people to keep reading. Presumably, the thought was that some readers, if they read the thesis statement without first seeing what kind of evidence for it would be brought to bear, would immediately dismiss Egan as a crackpot and not read the rest of it. But I think that readers ought to be extended a little more faith than that.
This feels like the difference between expecting to read an essay versus expecting to read a work of fiction. Thesis-at-end makes for great fiction but is bad for information transfer.
I really enjoyed the early parts of the review, but was completely underwhelmed by the conclusion. I then went on to completely forget the conclusion even existed and in fact voted for it to win.
I have no regrets. The review convinced me that it is possible to somehow do education better than we are doing it, which honestly is something that I think I gave up on a decade ago.
I'm confused by this. The actual applied suggestions at the end were all slight variations on how we do things already (usually for the worse IMO, but not even outside the overton window).
Instead of taking an idea and presenting it in as clear, simple a way as reasonably possible, it had basically no good points to make (just a simple one at the start about how we want three different things from education that are hard to balance, which should have been a short summary), and used excessive wordiness to obfuscate this.
This is the worst sort of thing to write. It creates a sense of false knowledge, because the people who read it read a lot of complicated smart-sounding words so they get an impression they understood something, except there wasn't any real idea behind them to be understood. So people just leave confidently clueless. And I really hate the idea that some people will now think this is good writing and try to imitate the style.
(Also, it doesn't help that all the specific educational suggestions it has are both banal and bad ideas on the merits).
There are some categories where the marketing is the product. Books on education are perhaps in the same category as self help books. In both cases, the big difficulty a good text is helping its readers overcome is not ignorance, but rather hopelessness.
If someone walks away from such a text feeling that change is possible, he has absolutely gotten his money's worth.
Why? I can understand having a mountain-sized prior against any Brilliant Theory Of Education Which Will Help Everyone being at all effective; but with that caveat, the theory does at least seem as interesting as any, and I don't doubt its creator's good intentions… Hardly seems worthy of active hatred (unless you reserve the same scorn for all brilliant-new-theory-of-education-crafters on principle?).
Well, not having read the book but only the review, I may be being unfair. But the impression it made on me of yet another guy going breezily "Okay, we all know the education system stinks, but here's my One Weird Trick to fix it" and then the One Weird Trick is the old "take a handful of bright, motivated kids; give them one-to-one attention as much as possible; pump a ton of resources into the project; then after a few months come away saying 'theory proven' and let the model lapse" - yeah, it makes me grind my teeth.
It's certainly not a description of what the One Weird Trick purports to be, but I'm not sure it's that inaccurate of a representation of what the Weird Trick has been in practice. Sharply limited (or zero) empirical findings is the typical hole in the heart of education research, and I'm not sure Egan does better than average there.
Or rather: education is in need of the people theorizing about it to actually read the empirical literature that does exist. The book review, and the book it reviewed, just ignored all the research that exists. Frustrated it won.
Agreed. It was by far my least favorite review. An order of magnitude longer than it needed to be, smug and self-satisfied in all the most annoying ways, obnoxiously burying the lede in order to force the reader to slog through a piece that the author acknowledges is too long and poorly edited. I'm legitimately shocked it won.
Yeah; to me it rapidly stopped looking like a good-faith review and started looking like a pitch for a pyramid/network sales scheme. I quit with the scrollbar still in the top half of the window.
I slogged through it, and was prepared to overlook the smarmy used auto sales pitch, but when at the end it turned out the reviewer hadn't bothered to mention any of the voluminous subsequent empirical work, not even in a casual way, then my inner Freddie de Boer said "enough of this arrogant nonsense". I'm sad it ended up high, let alone winning, and I wonder if an online homeschooling community was brigading votes. My three choices all did well and I spotted Njal's Scott fingerprints, so mostly happy.
I don't think it was the worst review, but it was the one I had the hardest time reading. In addition to editing for length, it could have used more formal structure, including stating the thesis up front, and outlining the argument that it was going to make. As it was, it felt like one of those hikes where every time I think I'm close to the end, it turns out there's another huge stretch to do.
I did read it all the way through and I was really hoping it wasn't going to to win, mainly for it's ridiculous length. It was so extraordinarily long (and ultimately boring, unlike the land tax one) that I thought it might break that curse of the longest review winning, but that was surprisingly not the case.
Hah, yes, that was the other one that was most hard for me to read. But in that case, it was entertaining enough to keep me going, but I was stopping every two paragraphs to wonder "is that really true" and work through a few thought experiments in my head.
Hilariously I was sitting in a tent camping unable to sleep and still couldn't finish it. It was just so disrepectufl of my time, and not very interetsing on top of it.
The good points of it could have been summarize don one graph. "Education is sis doing different things and they conflict".
Yeah I try to write myself a couple words of notes on each review after reading them so I'm not completely going on months-old memory when it comes time to vote. My thought on that review was: "too long and not worth it."
This year I stuck to the simple schema: like the reviews you would vote for, with final choice depending on voting scheme. I quickly checked each review at the end to verify I hadn't changed my mind (Zuozhuan and Cliffs were close calls).
Yeah, single transferrable vote, in which the vote can only transfer twice (Once if you were a Njal's Saga voter), is going to make for a lot of vote splitting.
I, for instance, liked them all except Educated Mind and Search for Meaning, which means I checked the Manifold predictions so my secondary votes would be for alternate frontrunners... like Njal's Saga
Parent and teacher - and pro-Caplan. And still, I did not love that review, voted for another (not true-Scott-style Njal, not into Vikings too much). Drown your pet; new pet:" ex-school-students who suffered" are prone to like it. And suffer we did. Very surprised it won, being loooong and redundant at times. As a good teacher would write. But then, ACX readers really like to learn, I guess. It was quite convincing, too. Not at all written badly.
Speaking as a parent, I don't have time to read a 27,000 word book review. I got nappies to change, lunches to make, I gotta listen to little voices blathering on about something and politely pretend to pay attention. I'm too damn busy for this.
1. I also found it far, far too long: as others have said, if you want that much detail read the actual book, especially when most of the review is just summary not analysis. I thought it started off great, with the triangular explanation of the main educational purposes, and had real clarity and insight. And then it just went into endless rambling (when that Q and A format started), and unlike other reviews it felt like *deliberate* rambling, padding out the review and teasing the conclusion. I'd *really* like to know what so many people saw in this, and also *why* people consistently think longer means better. Succinctness and brevity are generally regarded as important virtues in writing--but not around here apparently.
2. The actual theory from the book itself has (perhaps fittingly) the same problem for me: too much, too overly complicated and dense to be a remotely feasible educational approach. Good luck getting even a single school to properly implement those five kinds of knowledge, divided into three stages of learning, but where actually each of the five kinds needs to be present in every stage, so the stages are separated but really they're not, they're integrated, and every stage and every kind must include examples from every time and place on earth, that's essential, and must treat each student as an individual while also teaching the same material to everyone, make sure you do both at the same time, but make sure you also separate those approaches and don't confuse them, along with the following dozen other requirements...tell me this is not meant seriously.
I guess it has one advantage: since no one will ever be able to fully implement the author's insanely complex theory, he can declare victory and say it's never once been falsified. Well done.
3. I don't object to Scott using only-3-preferences-allowed "IRV" (given the difficulty in counting), but I think this is the reason this review won. 3 preferences is really much closer to FPTP than real IRV, and so polarising reviews that stood out, that maybe the majority of people loathed but a focused minority loved (much like Donald Trump) can end up winning.
I got the same feeling I did when reading various diet or exercise books, where they go into exhaustively loving detail about how every step of their plan is based on what's natural and normal for humans, and they carefully list all the evidence in favor. It's an embodiment of "rationality as a tool of persuasion", where the core thesis could be summed up in a few pages, and the entire rest of the book is support for that core thesis. But not brutally honest debate with a steelman; instead, lots of "reasons" that may not ever have had contact with an opposing point of view.
It's the sort of book I want reviewed by someone who disagrees, but who is also rigorous and honest. Anything that survives that process is almost guaranteed to be useful. As with the review for "Why Machines Will Never Rule the World".
I just want to say that I really enjoyed the Book Review Contest—I didn't finish (or even start) every entry, but I found the book selections wide ranging and in some cases utterly captivating. (On Marble Cliffs and Njal's Saga being the most unexpected). Kudos to all the reviewers and to you for running such an interesting contest.
Same, I really liked the book review contest, I read & finished all the entries (except the last one which I've only started so far) and enjoyed them all
Thanks for hosting the contest once again Scott! I'm pretty happy I upgraded from finalist to podium this year. Congrats to Brandon and everyone else too! And thanks for those who believed in my review enough to place it first on the prediction market :D
For those who might be interested in Jane Jacobs further, I'll be writing about some of her other books in the near future.
(Also for the record I really liked The Alexander Romance and it was my only 10/10 on the preliminary round and I'm glad I publicly predicted on Twitter that Scott had written it)
I had the impression that the response to "Cities And The Wealth Of Nations" was pretty negative from what I remember of the comments but it did really well
For people who voted for that book review, do you agree with the core idea there?
I'm pretty sure I would have noticed if most of the feedback was mostly negative. I think many did disagree with the core ideas from Jane Jacobs (I myself disagree with some of them), but liked the review as a piece of writing. It also got the most Substack likes, for what it's worth.
I really liked it as a piece of writing, but a lack of prediction in macroeconomics sets of warning bells; at the very least, Montreal seems to be doing more or less fine. I'm not sure much of the core thesis holds these days anyway, either under efficiently globalized supply chains or even nation-scale derisking.
Yeah that's reasonable. I think if I wrote it today I'd be somewhat less enthusiastic about her ideas, especially in the conclusion. I still think her books are great and... philosophically accurate, in a way? But I understand and sympathize with the skepticism on economic matters.
Also, I would have loved to analyze how well her predictions held up in the past 40 years, but unfortunately that would have been 1) way outside my expertise 2) significantly more work 3) significantly more words and 4) a risk of outing myself if I gave my impressions of modern-day Montreal.
Yes, that's pretty much how I felt about it. A very good review, and I disagreed with some of the ideas in the book, but it was tasty food for thought.
(For instance: transport has gotten faster and cheaper over time, so shouldn't that affect the size of the thing she calls a "city"?)
Yeah that's a great question, which as far as I know she doesn't explicitly get into. However she does allow for pretty large definitions of cities: in her last book, which I'm rereading at the moment, she calls Czechia, Slovakia, and Taiwan "quasi-city-states".
I got some push-back in the comments for hypothesizing such a thing on the level of the USA, so I'm glad to know that it wasn't an entirely absurd idea. :-)
I missed any intellectual context for Jacobs's ideas. A brief discussion of Von Thünen's "ring" model, the core-periphery model and Venables's and Krugmans' New Economic Geography would have helped place Jacobs in some sort of intellectual tradition of inquiry.
I was one of the people coming down hard on the ideas in the book. I think it being the first review to be posted was a big advantage in terms of recall value.
For everyone who read the book review, I'm curious what you thought about it (either about the book or about the review). Do you think the replication crisis reveals that something is deeply wrong about science, or is fraud, hype, bias, and negligence just part of the human condition and an inescapable part of the scientific process? Has the replication crisis changed your worldview? Do you agree with Stuart Ritchie's ideas for reforming science? And finally, what did you like or not like about my review (as distinct from the book)?
Since you ask, here goes. I have no particular strength as a writer or editor, so please take everything I say with a huge grain of salt. :-)
I think "fraud, hype, bias, and negligence" are all part of human nature, and the process we call "science" is an attempt to distill truth out of these impure inputs. And this will of course be imperfect: "out of the crooked timber of humanity, no straight thing was ever made". The replication crisis is a sign that not only is the process flawed more than we thought, but that the internal checks are also flawed in ways which encourage other flaws. I can't remember what my worldview was like beforehand, but I do seem to recall being more likely to be credulous of pop-sci claims than I am now, but there were other reasons for me personally to shift away from that, so I can't attribute it all to the replication crisis. FWIW, I'm not a scientist in my day job, but I have friends who are. Given that, his ideas for reform all sound reasonable to me, but I find his omission of ideological bias to be a worrisome blind spot. How do scientists react to the potential of a politically inconvenient truth? Do they seek it out, or do they avoid being associated with it?
As for your review: It's easy to read, and a good size, maybe even a little on the short end (not that that's bad if you said everything you want to say), although again, perhaps this is partly because it's so easy to read. I think it could stand to have slightly more formal structure. In your last paragraph, you start talking about the nature of some of the results, and what that might say for human nature as a whole, and I think this is an interesting thread that you might have touched on throughout the review. I suppose my main criticism is that it's not telling me anything new? I went into it broadly agreeing with the thesis, and I didn't hear anything particularly new (except for that last paragraph). All his ideas for reform sound good, but afterward I still don't know: How practical are they? Which are being implemented, and why those, and why not others? Are there better ideas for reform out there? Who disagrees with what he's saying, and why? If there's difficulty in making reforms, where does it come from and how can it be neutralized?
Frankly, I think you might have run into one of the biases that you write about in the review, "we want to discover new and exciting phenomena".
I liked it! Even though I do agree the replication crisis has been talked to death by now. I remember when everyone was excited about priming, and then a summer camp had us do power poses... it was worth it because I got over my fear of being non-conformist in public, and they gave us a Kindle. But when I heard power poses failed to replicate, I felt vindicated for not having liked them.
I think the replication crisis revealed some major flaws in how journals chose what to publish, which have been partially to mostly fixed since. Humans aren't going to stop being flawed, but with the right process we can get pretty accurate results despite that.
The replication crisis hasn't had much effect on my worldview, or if it did it got lost among all the major changes that were happening in my teenage years.
I'd tell you what exactly I liked about the review but I have no idea, unfortunately. Anyway, congrats! You wrote a thing that's good :)
The throwaway line with the sins annoys me a good bit. Not only is it an unnecessary word sink to list all seven Deadly Sins, not only do you call them the four deadly sins before listing the seven, not only do you say they don't apply when Pride, Greed and Sloth clearly do, but you also only wrote six of them and missed Envy, one of the most scientific-study relevant ones.
With regard to hype, the book is using it itself; the trachea story is horrible, but were people trying to replicate it based on his claims? If not, it's one guy who got caught and the system is working as intended.
My basic thought is that it's really easy to never be wrong, you just stay silent. The method is not "deeply wrong", it's a necessary evil that grows into a larger, less necessary version of itself without constant vigilance. In both directions. Journals are incentivized to publish false positives, but a null journal will be incentivized to publish false negatives; it's all mud forever.
Surely the point of null result journals is providing balance to positive result journals? Right now, if a researcher gets a null result, they're incentivized to tweak it to get a positive result, since otherwise they can't publish. If both positive and null journals exist though, they can just do their research honestly and publish in whichever journal ends up applying. This works even though each journal will only publish either positive or null results.
Researchers will always be incentivized to tweak their results to make themselves look smarter, and no one would want to read a journal filled with "I thought this but was wrong". It would exist primarily as a counter-culture takedown of previous positives.
If they're trying to contradict an existing study, yes they'll do that. Otherwise they'll continue to tweak them to make it look like they got a positive result because the positive result will still be more prestigious.
I think the grammar of the sins sentence is OK? He previously mentioned "four cardinal sins of science", and what they are. He then says the "The four deadly sins of science are not pride, greed, wrath, lust, gluttony, and sloth, but fraud, bias, negligence, and hype." To me, this is saying, "the 4 members of category X are not A, B, C, D, E, or F, but instead H, I, J, and K". This seems fine? There's lots of potential members of category X, and he's ruling out 6 of the 7 candidates that probably come first to mind.
It's not "or Sloth" though, it's "and Sloth": all six are placed in the same grouping of potential Four Deadly Sins. And outside that, there's no reason to specify Four here; the original is Seven, so to preemptively declare there's four hollows the comparison before it's made.
WOW! Thanks all around for the reads (especially everyone who read through to the long, bitter end), thanks for the feedback, and thanks for the votes! I’ll be posting a “Highlights from the Comments on the Review” on my substack. (Never fear, it’ll be shorter. Slightly.)
I didn't think you'd win based on the length, but it *is* an excellent review and your first place is well deserved. I guess this shows how verbosity is very much not penalized on ACX, despite the many comments complaining about it.
Agreed, just I dislike "verbosity" - slightly pejorative: remember the NYT calling Scott "verbose" - though the NYT is nothing but - Oxford dic. says: the fact or quality of using more words than needed; wordiness. "a critic with a reputation for verbosity". ACX-readers like reading well written stuff and to learn along. That is why many of us like your excellent substack! Your place on the podium is well deserved! Bien joué!
Thank you so much for writing that review. It came out right as I was planning how to homeschool my child and I had never heard of Egan before. I'm now working my way through his books for more ideas on how to teach him. I voted for your review as my first choice, since it is the one I am confident will change both my life and my child's life.
I feel like I should have an evening dress or something for this grand occasion, but oh well.
First, I must apologise and not alone eat crow but humble pie and my hat. To everyone who said "Scott has a book review in this contest and it's Njal's Saga", I was wrong and you were right. The deceit, the duplicity, the deception! I am shocked, shocked I tell you!
Now excuse me while I choke down this mouthful of feathers.
I am disappointed Zuozhan review didn't do better, so good luck to T! It was an excellent review and educated me on the subject.
"I’m planning another contest next year. I haven’t decided if it will be book review or generic essay"
I would really love to see a reader submitted generic essay contest, although I suspect there would be 100s of entries since the barrier to entry is so small.
The LRB and other publications that have book reviews as long form essays, allow the writers to review more than one book on the same topic (for instance if there are two books on Stalin, or whatever) - is that possible in this competition, next year?
I was surprised to see Njal's Saga not in the top 3, based on the overwhelmingly positive response to it elsewhere. Then I scrolled down, and everything made sense again.
Thanks for posting all the scores, I was very curious how well I did.
Getting 54th out of 145 is just good enough to encourage me to try again next year. I probably won't be a finalist, but I think I can beat my high score! Though I wish there was a way to get feedback from those low scoring reviewers so I can figure out how to improve.
Here are some thoughts, I hope you don't mind that they jump back and forth a bit:
There's a couple grammar mistakes (using principle instead of principal; the second sentence starts with "How could..." but doesn't end in a question mark; TV isn't capitalized; "LIterature" is weirdly written with a capital 'i'). They're small details but when I've submitted scores in the past for this contest I'd take a point off if I remember noticeable grammar/spelling issues. Which doesn't sound like much, but +1 point to your average would put you right up on the verge of being a finalist.
To be honest I felt disappointed by the review. I assumed from the title that I would learn about the mind of medieval people. But what I actually learned felt underwhelming. My main takeaways were (1) they had a certain model of the universe and (2) that model... led them to really enjoy detail?
But I didn't even really understand #2 — one of the examples of "vividness" was "‘So stant Custance and looketh hire about’". That's vivid? (The other examples aren't much better at being vivid...) So overall the thesis didn't really "click" for me. One note I wrote down as I was reading through it: "You spend a long time discussing the ancients' model of the universe when I thought I was getting a review about medieval literature.".
I also felt confused b/c you started by saying they "were nerds"...And then I thought the meat of the review would dig in and show me medieval writers nerding out on King Arthur stories (or whatever) the same way modern nerds nerd out about Superman v. Goku. Who did they write fan fiction about? Whose canon were they debating? I felt like this was an interesting 'hook' but then it never actually got discussed in detail.
I know in Scott's old article on nonfiction writing advice he suggests having pictures, and I tend to try to follow that advice. A couple well-chosen images/diagrams may have given you a tiny boost (and the model of the Universe would be a natural thing to diagram). My general impression of your writing was that it could be "tighter", if that makes sense. Aside from Scott's old article, for writing advice I also like Steven Pinker's The Sense of Style and Julian's guide: https://www.julian.com/guide/write/intro
"references that no longer hold" --> Not sure I've ever heard of "holding" as a verb to use with "references"
One thing I noticed about your writing style, you seem to really like making lists:
How could they make sense of our in-jokes, backwards science, pop culture references, and unspoken assumptions
The fictional settings these characters are placed in have different rules, histories, and laws of physics
The impulse to systemize, to catalog, and to understand is strong among nerds.
"They dedicate hours of their lives to filling fan-wikis, assembling timelines, and collecting reference works."
"multiple authors and ages, Pagan and Christian, Greek and Roman, fiction and non-fiction, "
" where all the muck sinks to and collects."
Not saying it's good or bad, but something to be aware of just in case you aren't. Use it with intention!
Hope that helps! And hope you don't mind the criticism, but you did ask for it :)
PS -- if you want feedback *before* submitting a review next time around, feel free to contact me & I'd be happy to be a "beta reader"
Thank you so much! This is exactly the kind of feedback I wanted.
I did end up leaving out a lot I wanted to include because I ran out of time. Reading your comments, I realize the whole thing was a bit half-baked. I should probably start next year's book review right now, so I have plenty of time to edit it. I may take you up on your beta reader offer!
I had one final thought that goes very nicely with what you've said in this comment. I didn't like the comment you made about being out of time before your deadline in the Conclusion section. On the one hand, I understand what you're going for (e.g., a general "this book is really interesting and I wish I could say more")
But it also feels disrespectful to your readers, in a way. As in, you didn't respect them enough to write your review early so you'd have a reasonable amount of time to make it the best that you could. And if you want to be closer to being a finalist, which it obviously seems like you do, then you should show your readers that respect!
I’m thrilled to hear The Educated Mind won. I enjoyed most of the book reviews in the contest but that one genuinely changed how I think about education and motivated me to dig deeper into idifferent learning and teaching theories.
Welp! I was surprised not to see Njal's Saga in top 3, and low on the prediction market, but this explains a lot. I had no idea it would have been Scott's, I just really enjoyed it.
Even if it was a different channel, throwing videos into your book review feels like something that most entrants wouldn't do because they'd assume it was against the rules. Only the guy who makes the rules would be so bold.
The bad news is that next year everyone is going to include videos in their damn reviews.
Shocked that “The Educated Mind” won, it was a chore to get through and it seemed aware of that and felt a need to constantly reassure the reader of the reward at the end (at long last reaching the thesis). Pleased with the other two though, “Cities” was my favorite and “Njal’s Saga” and “On the Marble Cliffs” were near contenders.
I read most of the entries, but I wasn't able to get through The Educated Mind one. It came off as really generic education studies stuff, as I recall.
I'm happy to see that On the Marble Cliffs placed so high, it was my favorite.
I somehow missed the opportunity to vote, despite being a regular reader of ACX. It looks like it was only posted on Monday, and then voting closed on Wednesday. There really needs to be more than two days of voting, especially after the contest itself took place over many weeks.
Oh, I see it was. I think that must have been just after I last checked ACX then. Surely I'm not the only one who sometimes goes a whole *five days* without checking ACX.
I felt a bit rushed as well - I didn't want to be biased by reading them over a period of literal months, so I read them all in the voting period and only barely made it.
I'm not sure if this is a symptom or our societies polarization, or if it's the fact that this kind of thing being true that is causing it, but what I'm taking away from the comments here is that the most polarizing review was also the most popular.
What's slightly interesting to me is that, given that it was a ranked choice voting system, that _mostly_ shouldn't have been true (or at least, in general on average that shouldn't be true).
Vote exhaustion greatly reduces the effect of preference flow (since, y'know, preferences after the third *don't* flow), so that's not especially surprising.
Yes, it shows what has been a consistent feature of this blog for at least as long as its been ACX: The readership and the commentariat are two different beasts.
Isn't that always true across all internet communities? The lurkers outnumber the posters pretty consistently.
My question is if it also holds in real-life communities. I feel like when I'm in small gatherings in real life there are fewer permanent lurkers, but it might also be a function of size. It's harder to hide in a group of ten than in a group of one hundred.
My point isn't that propensity to comment isn't uniformly distributed among readers of ACX. It's much more specific: The population of ACXers who post are different across a number of characteristics from the population of ACXers who read but don't participate in the comments. The example that springs to mind from one ACX survey or another I remember is political leanings. Commenters tended to be more right wing than readers.
Reminds me of a game I played and worked on as an open source dev where the entire forum and dev community was focused 90% on MP play, but if you looked at stats 97% of the actual player base was SP only.
Ranked choice is actually really good at finding polarizing choices. Every step asks which options got the fewest first place choices, then kicks them out. The end result is that a option which is everyone's second choice gets removed instantly; the only thing that matters is the number of diehard fans.
This is one of the reasons I dislike ranked choice. Approval is, honestly, better.
Approval voting gives you the contestant who annoys the least amount of voters. The problem here is the least known contestants can often end up winning; people seldom disapprove of what they don't yet know of.
Plurality voting (FPP) encourages the contestants to define themselves as what the other is not and for most contestants to withdraw from competition so that only two realistic choices remain.
Ranked choice is somewhere in the middle: allowing some polarisation (you must be well-known to win), but not so much that you discourage too many people from giving you a high secondary ranking.
You assertion about approval voting suprises me; in the few votes I have seen using it, if I don't know who someone is and what their capabilities are, I don't vote for them.
Sorry, I don't. Just experience from watching a lot of elections; a relative unknown can get some lucky favourable media coverage (for example at a debate), that's most of what people know about them, then win and have quite disappointing performance in office.
The archetypal example here in NZ was Peter Dunne.
IRV with vote exhaustion and most-votes-wins (which is what was done here) is far worse about this than full majoritarian IRV (in which you only need a fairly-small primary vote to snowball to victory).
Do not conflate ranked choice with instant runoff; there are other (Condorcet) methods that still use ranked ballots but do not have the unique flaws of IRV.
ETA: everyone's second choice, if first is fragmented, will absolutely prevail under Condorcet.
No, no, no, no, no! Sorry, but this is really bugging me: this was NOT a ranked-choice voting system. And it wasn't proper instant-runoff voting either. A real ranked-choice system lets you rank ALL the options (at least if you want to) and counts all of your preferences. If the majority (as I strongly suspect) would have preferred "On The Marble Cliffs" to "The Educated Mind", a real IRV count would have redistributed the preferences from all other entries (once they had been eliminated) between those two, and the former would have won.
How it was actually counted was MUCH closer to first-past-the-post than to actual ranked-choice. I don't begrudge this at all, as I understand a ranked-choice vote would have been difficult and time-consuming to count. And also, I would have had no problem with just using FPTP in the first place.
But for the love of God, please let's not pretend this was a real ranked-choice vote. It should be described as either "modified FPTP" or "very limited RCV/IRV" or something similar. And being in effect a form of FPTP, it's not at all surprising that it has that system's same problems of electing highly polarising options.
I'm not sure I understand. What is your understanding of how this voting worked, and what specifically did it do or not do that makes it not count as ranked choice?
Okay, imagine this was the first-preference vote (totally made up, and using only five contestants):
The Educated Mind: 88
Cities and the Wealth of Nations: 50
On the Marble Cliffs: 32
Man's Search for Meaning: 19
Zuozhuan: 11
Case 1, under first-past-the-post, there are no further rounds or preferences, and Mind wins.
Case 2, how it was done with only three preferences allowed: Zuozhuan is eliminated, its voters all had Meaning as their second choice (they like epic historical stuff), which now has 30 votes, still the lowest, and is eliminated next. Those 30 votes flow to Cliffs.
Mind: 88
Cliffs: 62
Cities: 50
Now imagine all the Cities voters had Zuozhuan as their second choice, and Meaning as their third. Since those have both been eliminated, and they were only allowed three preferences, those votes all exhaust and have no further effect. Cities is eliminated, and Mind has 88 votes to Cliffs' 62, and Mind again wins. None of the 112 people who didn't vote for Mind get to express a preference between Mind and Cliffs (just like they didn't in case 1).
Case 3, proper instant-runoff where people can express a full set of preferences (from 1 to 5). Same as above, up to when Cities is eliminated. This time, those Cities voters have their fourth preference read: 45 of them had Cliffs ranked fourth, and Mind last. Those 45 votes flow to Cliffs, and in the final round...
Cliffs: 107
Mind: 93
...Cliffs wins this time! Why? Because of the 112 people (the majority) who *didn't* vote for Mind at the start, almost all of them (107) preferred Cliffs to Mind. Or, more roughly, of those who didn't *love* Mind, almost all of them *hated* it, and they were a majority.
Both FPTP and only-3-preferences "IRV" gave no weight to that majority's preference for (let's assume) *all* other entries over Mind. Thus, I'm claiming both case 1 and case 2 (where case 2 is how the actual contest was counted) did not allow voters to fully rank their choices, and this (in my hypothetical, and very plausibly in reality) disenfranchised the actual desires of the majority. Thus I don't think it can fairly be called a ranked-choice vote. It functioned similarly to first-past-the-post, favouring a polarising candidate that got a lot of love and a lot of hate, over one with much broader appeal. Only a vote with full ranked-preferences would have favoured the latter.
Let me know if there's any part of this you didn't understand, or that you don't agree with, because I *really* want people to see this point. And in particular to see how much the details of the voting system used can change the outcome.
It's even more muddled than that because The Educated Mind didn't get the most votes, Scotts review did, and was then discarded. I'm not sure where the Saga votes ended up.
I don't think it's a big deal either way, but yeah, this shouldn't affect your view on ranked choice or whatever.
A better way to collect votes (with google forms) might be to have a list of all candidates and ask the voter to establish a preference ordering by assigning arbitrary points to them, with the understanding that the sorting of the candidates by these points (with ties randomized) will establish their preferences.
Just allow them to put a floating point number for every candidate, and treat any empty fields as zero (and discard the votes of anyone who assigns not-a-number). So if you like two candidates and really hate another one, you could just assign 2, 1 and -1 to them (leaving the others at zero) and be done with it.
He even posted the Video on his own YouTube Channel he has used for SSC videos in the past. He wasn't trying his hardest to hide it. In general, it was the funniest review by far, though I only ranked it 3rd because two others were more informative.
Scott: Now that it's been established that it was you, can you explain your thought process behind *why* you didn't comment even in passing on almost any of the parts of Njál's Saga for which it has been famous and beloved for centuries? I get that the point for the review was essentially to crack a joke about legal systems very different from ours, but surely the concept of a review would normally oblige someone to engage at least cursorily with the substance of a book? I guess most of your readers haven't read Brennu-Njál and couldn't tell, but reading the review as someone who *is* familiar with the book, seeing not one word about e.g. the attack on Hlídarendi in favor of the ridiculous legal quibble parts nobody cares about is almost surreal.
This sounds to me like a more polite version of "fuck off and shut up", and I respect that I guess, but regardless of personal tastes, can you, having read the book, nevertheless apprehend that a review of Njál's Saga which mentions the name "Gunnar" one time in a parenthesis is in some objective sense... not really focused on the same things that the actual book is? I think if I hadn't read the book, and then read it based on your review, I wouldn't have expected or even recognized the book I got at all.
After looking at The Educated Mind, I agree with the other commenters here that it was terrible, quite possibly the worst review of the contest. I'm amazed it got any votes, let alone first place, and that's saying something since I thought that nearly every entry deserved to win this year.
> Njal’s Saga, reviewed by Scott Alexander. This one got the most votes, but I’m disqualifying it because it seems in poor taste for me to win my own contest.
Now I can't decide if intentionally not voting for Njal's Saga cause I was pretty confident that was the True Scottsman actually made sense. I guess it at least ensured none of my votes got discarded? But it was the most entertaining, well-written (even by Scott standards - felt like "classic" SSC-era), and breaking the nonfiction monopoly is a big plus in and of itself. Not expressing those preferences feels...inconsistent.
The Educated Mind...would have expected it to top5, given the demographics, but not actually win. The length and imprecision were real bummers, even though I overall enjoyed and "liked" it. I think I'd have given up if not for priors being adjusted by FdB and Caplan - that is, having the expectation that ponderous ed theory critiques can actually be enlightening and moderately entertaining, so it's probably worth slogging through. The juice was worth the squeeze, at least in strict marginal idea-profit terms, if not best-uses-of-my-time utility.
Main bucked expectation for me was Why Machines Will Never Rule The World finalizing at all. I guess there's a lot more community (or voter anyway) split on AI fundamentals than I thought.
First of all, I loved the Njal's Saga review and I agree with other commenters who said that it was by Scott Alexander. I also really liked the Romance of Alexander review, and I said to myself, that one also looks like a Scott Alexander review! I couldn't tell which, if either, was by our illustrious host, or whether there were contestants who were just really good at imitating Scott's voice. (What is the kabbalistic significance of Scott Alexander reviewing the Romance of Alexander?)
Second, The Educated Mind winning first place was... well, it was a choice.
I can't say I *disliked* the review, because it had some excellent parts and made very good and important points, especially about how success in public education is difficult/impossible, because we're asking the education system to do different/conflicting/incompatible things: socialization, academics, and individual development.
But it was a flawed review for several reasons, and not just because it's so fricken LONG. I'm not going to reread it now, but Brandon Hendrickson seems to claim that True Education is about taking a child through five developmental stages, and then he gives examples of what to teach at each stage.
Which, ok, sure, but what is the evidence for the five stages and their correspondence to the child's age? I don't recall real evidence, mostly lots of anecdotes along the lines of "seven-year-olds like puns and jokes, which correspond to the XYZ stage." Also, I don't recall Mr. Hendrickson attempting to reconcile these five developmental stages with the socialization/academics/development (SAD) model - which sides of the SAD triangle are we aiming for with the five-stage model, and why? Next, the five-stage model seems to ask a LOT of teachers. They're supposed to not just know a very broad range of material, but also be gifted comedians, storytellers, entertainers, etc. Sure, this works if you're paying lots of money to recruit the best teachers possible to a small school attended by the children of the PMC, but how well will it scale across the whole society?
Finally, as other commenters on that review already pointed out, the review, for all its length, has some very important gaps. Mr. Hendrickson spends a lot of time on the proposed curriculum, and very little on things like testing, classroom discipline/behavior standards, tracking (do we group by age or ability?), etc. What to do if, despite our best efforts, some children just can't/don't want to progress through the "five stages"? And if the reviewer did write about these things, then I apologize for not remembering, but again, if you wanted me to remember more, you would have made your review shorter.
Again, I don't mean to sound harsh, you obviously put a *ton* of work into your review and it has a lot of very good things about it. Congratulations on your win!
Could you help with a link to Scott`s-The Alexander review? I somehow made it to a google-doc with review-entries from A to G - but not "T". (I agree with the points you make about the winner, btw.. I was very surprised it won. But kinda pleasantly surprised. If ACX-readers are ready to go for such lengths, I feel in good company.)
Aaaah, Merci & Dankeschön! Could not take the wait; now I feel as Balthazar Picsou does, when he is diving into his pool of gold-coins. Alexander Romance: a very Scott'ish review that is, again. Njal was better poised to win the contest.
I’ve noticed that Jaynes’ Probability Theory has been reviewed each year for three years in a row, but never does well. Given that it’s a core text for the rationality community, I’m curious if many ACX readers have read the textbook? My sense is that it seems a bit like a book that people talk about without having actually read it, and maybe for good reason, since the author died before completing it and a lot of it is half done.
I haven't read the whole thing, but I've used parts of it as a technical reference for work. I....really wouldn't have expected it to be the kind of thing that one reviews for this kind of contest or, if it was reviewed, for it to do well. It's basically a stats text book, admittedly an important one.
>1. I really approve of both using ranked choice voting and limiting it to three choices.
Why? It doesn't get the Condorcet loser criterion the way true IRV does. Indeed, Educated Mind was quite possibly the Condorcet loser, so this is in contradiction with your third point.
I am fairly sure that the chance of tactical voting helping is much greater in this method than in full IRV, because you have to choose between honest preferences or candidates with a chance (like in plurality). IRV does have tactical voting, but it is nowhere near as bad as plurality, and as essentially an IRV-plurality hybrid this is partway in-between.
What mushroom said, with the added point that tactical voting in real IRV (which this wasn't) requires highly accurate information on not only which options will be get the most votes and what order they'll be, but also what the lower preferences of those options' voters will largely be. Infinitely more difficult than in most other systems.
And also: how can having more choice *possibly* be more mentally taxing? Just limit yourself to 3 voluntarily if you find it hard. Why limit other people's choices?
Bit of a meta-comment: In terms of sheer quantity of text, these 15 reviews must be the most I have read from a single source (/collator) in a long, long time. I used to read this quantity of text from The Economist when I cared more about the news; occasionally from the FT, NYT, or New Yorker.
What I mean to say is that Scott is now, for me, as good a source of interesting writing as *the best* journal publishing houses in the world. That's an extraordinary state of affairs, and I hope someone studies it one day. But not Scott, because I just want him to carry on with whatever he's doing.
Congratulations to all the reviewers, I thought they were all world-class pieces of writing.
I fully agree with you. I had the same realization during the course of this year. Trying to read other blogs and sources, but nothing comes close to consistent quality of reading on ACX. The rest is mostly commentary, as they say.
I'm surprised at the amount of negativity directed at the Educated Minds review in the comments here. I found the book's outlook on education both exciting and sensible, enjoyed reading the review, placed it as my first choice, and am happy to see it win!
In general, I liked the contest a lot, thank you to all the reviewers for your efforts!
I see a lot of comments that say that the review was written terribly, rambling, teasing and generally confused how it won.
As someone that voted for it as a second preference, (Njal's Saga as first), I figured I'll say why I liked it enough to vote for it.
For me, I enjoy a book review when it introduces me to new ideas, and does so in an enticing and interesting way. I felt that TEM book review took me on an interesting journey, introducing a new way to consider education and adding a new lense to view interacting with children and teenagers.
This was especially interesting to me as it fit in with what I have actually experienced myself, and can see with my children: my grade school child is fascinated with stories, and she can spend 40 minutes reenacting all the parts of the stories she heard at school, but can barely recall other lessons. Meanwhile, her older brother is really getting interested in extremes: highest mountain, lowest place on earth, etc. What I'm trying to say is that as I was reading, the ideas resonated, and seemed to fit my observations and place them in the context of a theory.
I also enjoyed his explanation of the problems with current education, with different types of schools satisfying two out of three "ideals" and leaving out a third, important one, and found it to make sense. Especially given how drastically schools can vary (from personal experience). The review left me with hope that an education institution that isn't completely broken is possible.
As for the rambling and teasing, etc., I didn't mind that so much; other overly-verbose reviews gave me more trouble, and a couple just lost my interest halfway. This one did not. A book review that delivers so much of the book's content makes me happy with it. I also felt that the review addressed many of the objections, concerns and questions that popped into my mind as I was reading it. I spoke about the review with friends and recommended it often enough that it felt to me that it makes the cut as a solid second choice.
As a last note, while I liked "Cities and the Wealth of Nations", I didn't want to vote for it, as I felt economics reviews were overly represented in this competition (likely due to last year's winner), at five out of sixteen finalists. But I do agree that CATWON was the best economics review.
I wonder if people with children / significantly younger siblings were more likely to vote for it. Did anyone also have this experience?
I agree totally. It was the review I took screenshots of to send to my partner with captions of my own reflections. If a review does that then there is clearly something special in your experience of it.
I mostly based my voting on the ones that I could clearly remember - this may well have biased me towards later ones, but I think moreso just being unusual. I felt like I had to vote for On The Marble Cliffs because it made me actually download the book, I plan to read that some time when I'm in the mood for Literature.
I'm surprised more people aren't declaring their votes. Well, here are mine:
1. The Rise and Fall of the Third Reich.
This was terrific and hugely underrated. Very much what I'd like to see more of here (in comments and even in some posts by Scott) and elsewhere: clear, logical grounding of relevant facts, and then detail-focused, logically justified conclusions from those facts. And especially, solid grounding of claims about the present and predictions about the future in the parallels and principles from actual history.
It's amazing to me how little of this there is, and how much people rely on Sprawling Theories of Everything, attempted grand explanations of the past and the future and the nature of society and life itself, justified by little more than "this just seems like something that might be true". Even in purportedly "empirical" communities like this. JJ's review instead does the radical thing of answering the question of what might happen, or what we should fear happening, by looking at what has, you know, *actually* happened, in the past. And getting a grounding in that, picking out the important details, and drawing out general principles that we can apply to other contexts.
I'm not surprised it didn't do well in the contest, though. The fact that so many people love their Sprawling Theories of Everything, and their gut reactions that "*clearly* candidate X is the new Hitler, that's just obvious, I don't need any logical argument or actual knowedge of Hitler to make that claim!" means they won't be very amenable to tightly-focused and clearly-argued pieces that pay attention to details and distinctions. Please write more pieces like this, JJ. Even though the people you describe in your first paragraph are most of the internet and a large portion of the voters in this contest, there are definitely those of us who see its compelling explanatory insight.
2. Man's Search for Meaning.
A very courageous, sweeping essay, complete with narrative twists, integration of multiple historical contexts with one another and with the present, and of course a focus on the *truly* important questions. As admitted in the essay, it's a bit of a mess in terms of focus and structure. But polish it up a bit, and it could be a total masterpiece.
3. Njal's Saga.
Incredibly entertaining and well-written. I bumped it down slightly for being gimmicky: the multimedia youtube aspect and trying to copy Scott's style. But I then bumped it up for *succeeding* at the latter. In light of the truth...the last two points are invalidated I guess, leaving "entertaining but a bit gimmicky". Please, no one copy the second part next year, thinking that's what made it "win". It's not.
If I'd been allowed a fourth vote, it would have gone to ZuoZhuan. This one had, in fact, the good aspects of each of the above three respectively: clear and solid, sweeping scope, and entertaining style. It just had each of them to a lesser extent, making this a hard choice for me. In the end, I decided it was probably written by an academic or other expert professional, and that I wanted to give my votes to less credentialed contestants. In light of the biograohical information, it seems I was wrong about this, and maybe this deserved an even higher place.
"The Rise and Fall of the Third Reich" was the one I liked least. As a German with a minor in history , I cringed at a reviewer who read just ONE single book about the period - and one that is really not up-to-date. The author retells a lot of tales about Hitler that later turned out to be just fake-news (e.g. Hitler was NOT the 7th member of the party), and is ignorant of a lot of analytical work that was simply not yet done at his time. Knowing no other reputable book about the topic before writing a review about this one is ... misled and misleading. (I do not say the book is worthless, far from it. Nor was the review badly written. Far from it! But was it competent enough to be published? Sadly, far from it. - Sadly, I am no expert to judge the accuracy of other reviews, I all enjoyed and felt enlightened by. Maybe wrongly, too.)
The complaints against the winner need some perspective - even with ranked choice voting most people will have voted against the winner. This would be true of most winners given the number of candidates.
If there is an official list of haters of The Educated Mind, add me there.
At first, it makes some good criticism of the existing educational system. But, as is the usual problem with populism, the actual challenge is not telling people how the existing system sucks (most people know already), but designing a system that sucks *less*.
The proposal, in a nutshell, is to tell stories at the elementary school, interesting trivia at high school, and leave the boring stuff (such as addition and subtraction, I suppose) for university. Yeah, that would definitely improve education a lot!
Seriously, if you think this is a strawman, then I challenge you to show me a mathematical curriculum designed to comply with The Educated Mind, and put it side by side with the existing one.
The nearest thing that exists in reality is probably Waldorf schools. If you look at the lists of their famous alumni, you will find approximately zero scientists, which is not a coincidence.
Having looked through the comments section for that book review, I’m wondering whether the described system (and the layout of the review) appeals especially to intelligent, academically oriented people with ADHD or similar executive functioning difficulties. I was always good enough at math that I didn’t struggle in the subject like a lot of kids with ADHD did, but I found the way it was taught dull enough that it put me off pursuing a career in mathematics or physics research. In hindsight, high-level mathematical research seems to involve more of a creative problem-solving approach that seems like a good fit for people like me. The same is likely true for many academic research fields and engineering careers that involve a design element.
I get the general impression that disciplines always seem rote/mechanistic at the training-wheels stage, while the fun/creative parts manifest later. Math however, seems deep enough that even undergrad math is still the training-wheels stage. I suspect this prematurely colors people's experience, and is partly responsible for the widespread "not a math person" vibe.
Absolutely wonderful as always, these book reviews always manage to concentrate so much good stuff on so many topics. I wonder, are there links to the Google Docs containing the rest of them? I see one in this post, but only for the A-G section.
In anticipation of maybe taking part in next year's contest, does a review have to be of a very recently published book, or would one published in say 2016 be OK, or 1916 for that matter?
Thanks, that was interesting.
So it WAS you!
I feel like this should come with a "Phoenix Wright pointing" picture...
Does this work? https://www.youtube.com/watch?v=fo210ISA8aM
"My client can't be guilty ... because he's also the judge!"
I knew it too. Interesting that Scott thinks it's good that we are quote "cattle domesticated by the state". Doesn't he also claim to be a libertarian?
I mean isn’t “the goddess of everything else” essentially an anti-libertarian parable?
Only to a socialist.
For example, the central motif is summed in this line:
> So I do not ask you to swerve from your monomaniacal focus on breeding and conquest. But what if I show you a way that my words are aligned with the words of your Maker in spirit? For I say unto you even multiplication itself when pursued with devotion will lead to my service.
That must be why
I'd say the opposite -- in TGoEE, mutually beneficial cooperation arises naturally from red-in-tooth-and-claw competition (as a consequence of the fact that cooperative groups tend to outcompete the others), with no need for centralized planning. (Apart from that of the Goddess Herself, but She's just a metaphor for a natural process.) I think it'd make more sense to interpret Moloch as anti-libertarian.
Right, order out of chaos; but that order has to prevail and stabilize. Laissez faire is basically the motto of cancer; the emergence of stable state structures negates that. Centralized planning is no different from other kinds of cooperation, it’s neither more nor less natural; what matters is whether it succeeds or not.
All true; my point was that stable structures do not *require* centralized planning -- if anything centralized planning requires some ordered structure to be possible in the first place -- and that they can very well emerge on their own without control, biological evolution being a central example.
Sure. I can’t imagine a hunter-gatherer tribe doing “centralized” planning, even though the certainly did some planning. It is only natural that more sophisticated forms of organization can appear only later in history.
Damn Scott writing anonymously won his own Book Review Contest. That deserves respect.
It was very clearly in his voice, I certainly wasn't alone in guessing ... but I did genuinely really enjoy it and voted for it anyway!
Last year I like many people read the "Arabian Nights" review he published during Book Review season (thinking it was a submission) and thought "wow that's a clear winner," then in comments saw it was him.
I think it's a combination of the readers self-selecting for enjoying his style, and our preferences having been shaped over time to appreciate that style.
Plus SAS being an* outstanding essayist. Especially when it comes to book reviews. - Maybe the other reviewers were trying/striving to emulate him?
* make it "the"
I'm betwixt and between here, because I really liked that review, it was my number one choice in the vote, and now here am I with egg on my face and a mouthful of feathers from eating crow.
How will I ever trust again (sob)????
lol, I feel cheated, too. At least, you can take pride in spotting the 'best review'. While I voted for Mind of a bee.
37th/156 is also commendable
It's impressive, but on the other hand we're all self-selected as people who like Scott's writing in the first place, so it's not /that/ surprising.
Yeah, but like given that all subscribers have already selected to read Scott, I don't find it so strange that his review would get most votes.
It was clear it was him and the blog is a place where people who like his style congregate.
I'd argue that Scott shouldn't enter his own contest because, due to the readership's inevitable liking for Scott's style, it gives other entrants too much incentive to mimic his style.
But, it's his contest.
Thanks, that was really fun!
Yours was my favorite. Please keep writing!
Thanks! I really appreciate it!
Your review was my favorite. Was really beautiful, I still think about it
Thank you!
Also really enjoyed your review - and voted for you to win.
Thanks!
> Njal’s Saga, reviewed by Scott Alexander. This one got the most votes, but I’m disqualifying it because it seems in poor taste for me to win my own contest.
There was a prediction market about which book review would win "resolv[ing] to contest winner". An excellent lesson in how precise wording matters!
Meanwhile the prediction market on whether a review of Scott's would "be announced as 3rd place or better" finds itself in a complicated place...
Yuuuup. As literally written it ought to resolve NO, but there's no way to definitively answer it without making a lot of people severely annoyed at best.
At the end of the day, every prediction market is really predicting the decision of the resolving authority; when the nominal criteria *also* depend on the decision of one individual, things get spicy.
Yeah, I think this is always going to be an issue with prediction markets - there's always going to be some sort of Lizardman's Constant for cases where the 'spirit of the prediction' and the 'letter of the prediction' disagree - it seems like a lot of predictions require lawyer-like precision to define well.
A sound prediction market probably meets the legal definition of a contract, so it's not surprising that the fine print is worth fighting over!
IMO literally written it should be YES, because having the most votes is surely "better" than 3rd place.
But getting $0 in prize money is surely "worse" than 3rd.
Well if Scott won his own book review then he would be paying himself with his own money... which is still net $0 for Scott
Not quite; he's paying himself instead of a different author, so he'd still actually be up by that amount of money.
Ah, but it's "better" along a metric that makes no explicit appearance in the resolution criteria! Surely Scott reserves the right to remove participants regardless of their vote total, yes? Would a market for determining "nth place or above" be bound to recognize a contestant who received a number of votes but was disqualified at the 11th hour for, say, plagiarism? (It's happened before!)
But ultimately I'm not sure I care that much about the object level. Point is, *that* analysis is now ambiguous due to a factor outside the consideration of the criteria. So one is left pitching arguments to Benjamin Cosman, who has more or less carte blanche to give out ten bucks of not-so-play money whichever way he chooses. Probably not the proposition people thought they were betting on!
https://manifold.markets/ShakedKoplewitz/did-scott-write-the-njals-saga-book#tOeGaVJ5pIuIPI0BcFVU
It was my first foray into prediction markets after just getting into acx around that time. Am sort of thrilled with the outcome. Will be a fun thing to look back on. Twas a fun intro! Am loving the blog and have loved any of these book reviews I've read.
So the book with the most votes didn't win. I now have no idea what that means for possible future election systems, save that it seems to indicate the Electoral College is a good idea and giving presidential elections according to the winner of the popular vote is not?
It would have been amusing if, in the 2016 American presidential elections, a 3rd party candidate beat Trump and Clinton, but after the results were in, he pulled off his mask, Scooby-Doo style, and revealed that he was just Barack Obama, and couldn't accept a 3rd term.
Or Bill Clinton. A lot of people would have voted for Third Time Bill instead of Hillary.
I thought that market was overconfident that Njal's Saga was Scott's, settling as it did at 80% yes.
I didn't disagree that it read a lot like his writing, but I thought the probabilities of "Scott just mentioned entering his own contest to confuse us, didn't actually do it" and "of course some ACX readers try to write like Scott, it shouldn't be that surprising if somebody does a good job" ought to account for more than a 20% chance.
I guess I should update in the direction of Scott's writing genuinely being that recognizable, and people in general being able to accurately recognize writing styles.
I don't know I'd call it "recognition" per se, and would hide behind the near tautology of 'people who vote in ACX contests are strongly selected to appreciate Scott-like writing'. Honestly I'm surprised more people don't seem to explicitly try to ape his style, after the past few years' results.
I'm not referring to the fact that it got a lot of votes, I'm referring to the fact that the prediction market about it was at a stable 80% chance of it being Scott's before the results were revealed. People were confident that it was Scott's writing when it was still anonymous. I thought they were being too confident, but they were right.
And also update away from "some people being good at imitating a writing style". I was pretty sure this was someone putting in real *effort* to copying Scott and succeeding. Actually figuring out the turns of phrase and precise attributes that make his writing so enjoyable, unlike all the ones who just "vaguely copy the gist of it". But no, it really was Scott so...I'm just overwestimating the ability of writers to do that I guess.
To be honest I didn't at all notice that it seemed similar and thought the Alexander Saga was a truer imitation of his style (though still an inferior copy).
I also thought it was overconfident at first, and I bought NO at 80% in June. But over time, no further finalists came along that were anything like his style and I kept updating upwards.
Also, I believe the top YES holder was somebody who knows Scott personally? That raised my probability estimate a bit too.
So pleased to see The Educated Mind review win! Certainly the most mind-opening of any review I’ve read on the last two years of ACX for me.
I guess this means we won't be seeing shorter book reviews next year, not that I mind at all
Yeah. :-/
Clearly the winning strategy for these reviews is to copy the entire book word for word, and then add another book's length of commentary on top. Really get into every last possible detail.
Skeptics annotated bible-style!
If you do this then the year after I will review your review, also copying it in full with added commentary.
It was by a wide margin the worst review, and the only one I couldn't even read all the way through. I guess being polarizing is good in an open field, but still, bleh. It was bad.
(I usually don't like critizing something this openly, but the author was annoyingly self-congratulatory in addition to being a bad writer and having bad ideas about education)
What are you upset about here, the book being reviewed, the author of that book, or the reviewer of the book?
I strongly disliked the reviewer's tactic of putting the thesis at the end of a very long review rather than at the beginning, and of teasing it throughout in order to get people to keep reading. Presumably, the thought was that some readers, if they read the thesis statement without first seeing what kind of evidence for it would be brought to bear, would immediately dismiss Egan as a crackpot and not read the rest of it. But I think that readers ought to be extended a little more faith than that.
This feels like the difference between expecting to read an essay versus expecting to read a work of fiction. Thesis-at-end makes for great fiction but is bad for information transfer.
I really enjoyed the early parts of the review, but was completely underwhelmed by the conclusion. I then went on to completely forget the conclusion even existed and in fact voted for it to win.
I have no regrets. The review convinced me that it is possible to somehow do education better than we are doing it, which honestly is something that I think I gave up on a decade ago.
I'm confused by this. The actual applied suggestions at the end were all slight variations on how we do things already (usually for the worse IMO, but not even outside the overton window).
Instead of taking an idea and presenting it in as clear, simple a way as reasonably possible, it had basically no good points to make (just a simple one at the start about how we want three different things from education that are hard to balance, which should have been a short summary), and used excessive wordiness to obfuscate this.
This is the worst sort of thing to write. It creates a sense of false knowledge, because the people who read it read a lot of complicated smart-sounding words so they get an impression they understood something, except there wasn't any real idea behind them to be understood. So people just leave confidently clueless. And I really hate the idea that some people will now think this is good writing and try to imitate the style.
(Also, it doesn't help that all the specific educational suggestions it has are both banal and bad ideas on the merits).
There are some categories where the marketing is the product. Books on education are perhaps in the same category as self help books. In both cases, the big difficulty a good text is helping its readers overcome is not ignorance, but rather hopelessness.
If someone walks away from such a text feeling that change is possible, he has absolutely gotten his money's worth.
I think the review was good, in that it really made me hate the book, hate the guy proposing the theory, and hate the theory.
So, um, congratulations Brandon?
Why? I can understand having a mountain-sized prior against any Brilliant Theory Of Education Which Will Help Everyone being at all effective; but with that caveat, the theory does at least seem as interesting as any, and I don't doubt its creator's good intentions… Hardly seems worthy of active hatred (unless you reserve the same scorn for all brilliant-new-theory-of-education-crafters on principle?).
Well, not having read the book but only the review, I may be being unfair. But the impression it made on me of yet another guy going breezily "Okay, we all know the education system stinks, but here's my One Weird Trick to fix it" and then the One Weird Trick is the old "take a handful of bright, motivated kids; give them one-to-one attention as much as possible; pump a ton of resources into the project; then after a few months come away saying 'theory proven' and let the model lapse" - yeah, it makes me grind my teeth.
That was very much not the impression I came away with from that review.
It's certainly not a description of what the One Weird Trick purports to be, but I'm not sure it's that inaccurate of a representation of what the Weird Trick has been in practice. Sharply limited (or zero) empirical findings is the typical hole in the heart of education research, and I'm not sure Egan does better than average there.
Here here!
It just reinforced for me that theorizing about education is still in the wilderness and in desperate need of some sort (any sort!) of empiricism.
Or rather: education is in need of the people theorizing about it to actually read the empirical literature that does exist. The book review, and the book it reviewed, just ignored all the research that exists. Frustrated it won.
Agreed. It was by far my least favorite review. An order of magnitude longer than it needed to be, smug and self-satisfied in all the most annoying ways, obnoxiously burying the lede in order to force the reader to slog through a piece that the author acknowledges is too long and poorly edited. I'm legitimately shocked it won.
Yeah; to me it rapidly stopped looking like a good-faith review and started looking like a pitch for a pyramid/network sales scheme. I quit with the scrollbar still in the top half of the window.
I slogged through it, and was prepared to overlook the smarmy used auto sales pitch, but when at the end it turned out the reviewer hadn't bothered to mention any of the voluminous subsequent empirical work, not even in a casual way, then my inner Freddie de Boer said "enough of this arrogant nonsense". I'm sad it ended up high, let alone winning, and I wonder if an online homeschooling community was brigading votes. My three choices all did well and I spotted Njal's Scott fingerprints, so mostly happy.
After looking at it myself, I agree wholeheartedly. Nearly every entry this year was really good, which makes this outcome even more shocking.
I don't think it was the worst review, but it was the one I had the hardest time reading. In addition to editing for length, it could have used more formal structure, including stating the thesis up front, and outlining the argument that it was going to make. As it was, it felt like one of those hikes where every time I think I'm close to the end, it turns out there's another huge stretch to do.
I did read it all the way through and I was really hoping it wasn't going to to win, mainly for it's ridiculous length. It was so extraordinarily long (and ultimately boring, unlike the land tax one) that I thought it might break that curse of the longest review winning, but that was surprisingly not the case.
Hah, yes, that was the other one that was most hard for me to read. But in that case, it was entertaining enough to keep me going, but I was stopping every two paragraphs to wonder "is that really true" and work through a few thought experiments in my head.
Hilariously I was sitting in a tent camping unable to sleep and still couldn't finish it. It was just so disrepectufl of my time, and not very interetsing on top of it.
The good points of it could have been summarize don one graph. "Education is sis doing different things and they conflict".
Full IRV rather than the three-option vote-exhaustion one doesn't like polarising very much (it fails Condorcet but can't elect the Condorcet loser).
Yeah I try to write myself a couple words of notes on each review after reading them so I'm not completely going on months-old memory when it comes time to vote. My thought on that review was: "too long and not worth it."
This year I stuck to the simple schema: like the reviews you would vote for, with final choice depending on voting scheme. I quickly checked each review at the end to verify I hadn't changed my mind (Zuozhuan and Cliffs were close calls).
I thought the top 3 were all very weak.
Yeah I really didn't like that one at all.
Yeah, single transferrable vote, in which the vote can only transfer twice (Once if you were a Njal's Saga voter), is going to make for a lot of vote splitting.
I, for instance, liked them all except Educated Mind and Search for Meaning, which means I checked the Manifold predictions so my secondary votes would be for alternate frontrunners... like Njal's Saga
Agreed. My pet theory is that parents loved it and everyone else hated it.
I am not a parent, but I felt it was far and away the best review such that none of the others were even close
Parent and teacher - and pro-Caplan. And still, I did not love that review, voted for another (not true-Scott-style Njal, not into Vikings too much). Drown your pet; new pet:" ex-school-students who suffered" are prone to like it. And suffer we did. Very surprised it won, being loooong and redundant at times. As a good teacher would write. But then, ACX readers really like to learn, I guess. It was quite convincing, too. Not at all written badly.
>Parent and teacher - and pro-Caplan. (at least on higher ed)
Ditto.
Speaking as a parent, I don't have time to read a 27,000 word book review. I got nappies to change, lunches to make, I gotta listen to little voices blathering on about something and politely pretend to pay attention. I'm too damn busy for this.
I'm a parent and I disliked it, for the same reasons as others have given: long, self-indulgent, wilfully unclear, etc.
1. I also found it far, far too long: as others have said, if you want that much detail read the actual book, especially when most of the review is just summary not analysis. I thought it started off great, with the triangular explanation of the main educational purposes, and had real clarity and insight. And then it just went into endless rambling (when that Q and A format started), and unlike other reviews it felt like *deliberate* rambling, padding out the review and teasing the conclusion. I'd *really* like to know what so many people saw in this, and also *why* people consistently think longer means better. Succinctness and brevity are generally regarded as important virtues in writing--but not around here apparently.
2. The actual theory from the book itself has (perhaps fittingly) the same problem for me: too much, too overly complicated and dense to be a remotely feasible educational approach. Good luck getting even a single school to properly implement those five kinds of knowledge, divided into three stages of learning, but where actually each of the five kinds needs to be present in every stage, so the stages are separated but really they're not, they're integrated, and every stage and every kind must include examples from every time and place on earth, that's essential, and must treat each student as an individual while also teaching the same material to everyone, make sure you do both at the same time, but make sure you also separate those approaches and don't confuse them, along with the following dozen other requirements...tell me this is not meant seriously.
I guess it has one advantage: since no one will ever be able to fully implement the author's insanely complex theory, he can declare victory and say it's never once been falsified. Well done.
3. I don't object to Scott using only-3-preferences-allowed "IRV" (given the difficulty in counting), but I think this is the reason this review won. 3 preferences is really much closer to FPTP than real IRV, and so polarising reviews that stood out, that maybe the majority of people loathed but a focused minority loved (much like Donald Trump) can end up winning.
It is almost as if he wants to sell educational services!
I think you just put your finger on it.
I got the same feeling I did when reading various diet or exercise books, where they go into exhaustively loving detail about how every step of their plan is based on what's natural and normal for humans, and they carefully list all the evidence in favor. It's an embodiment of "rationality as a tool of persuasion", where the core thesis could be summed up in a few pages, and the entire rest of the book is support for that core thesis. But not brutally honest debate with a steelman; instead, lots of "reasons" that may not ever have had contact with an opposing point of view.
It's the sort of book I want reviewed by someone who disagrees, but who is also rigorous and honest. Anything that survives that process is almost guaranteed to be useful. As with the review for "Why Machines Will Never Rule the World".
Yeah, next time books should be excluded from the book review contest.
I just want to say that I really enjoyed the Book Review Contest—I didn't finish (or even start) every entry, but I found the book selections wide ranging and in some cases utterly captivating. (On Marble Cliffs and Njal's Saga being the most unexpected). Kudos to all the reviewers and to you for running such an interesting contest.
Seconded. I like the publishing schedule of Friday afternoons, when I'm already looking for an excuse not to do any more work, too. A definite plus.
Same, I really liked the book review contest, I read & finished all the entries (except the last one which I've only started so far) and enjoyed them all
Thanks for hosting the contest once again Scott! I'm pretty happy I upgraded from finalist to podium this year. Congrats to Brandon and everyone else too! And thanks for those who believed in my review enough to place it first on the prediction market :D
For those who might be interested in Jane Jacobs further, I'll be writing about some of her other books in the near future.
(Also for the record I really liked The Alexander Romance and it was my only 10/10 on the preliminary round and I'm glad I publicly predicted on Twitter that Scott had written it)
I had the impression that the response to "Cities And The Wealth Of Nations" was pretty negative from what I remember of the comments but it did really well
For people who voted for that book review, do you agree with the core idea there?
It was a well written review about interesting ideas, I had it on my list even though I didn't agree with everything there.
Otoh the educated mind was neither of those, I have no idea what people were thinking with that one.
I'm pretty sure I would have noticed if most of the feedback was mostly negative. I think many did disagree with the core ideas from Jane Jacobs (I myself disagree with some of them), but liked the review as a piece of writing. It also got the most Substack likes, for what it's worth.
I really liked it as a piece of writing, but a lack of prediction in macroeconomics sets of warning bells; at the very least, Montreal seems to be doing more or less fine. I'm not sure much of the core thesis holds these days anyway, either under efficiently globalized supply chains or even nation-scale derisking.
Yeah that's reasonable. I think if I wrote it today I'd be somewhat less enthusiastic about her ideas, especially in the conclusion. I still think her books are great and... philosophically accurate, in a way? But I understand and sympathize with the skepticism on economic matters.
Also, I would have loved to analyze how well her predictions held up in the past 40 years, but unfortunately that would have been 1) way outside my expertise 2) significantly more work 3) significantly more words and 4) a risk of outing myself if I gave my impressions of modern-day Montreal.
Yes, that's pretty much how I felt about it. A very good review, and I disagreed with some of the ideas in the book, but it was tasty food for thought.
(For instance: transport has gotten faster and cheaper over time, so shouldn't that affect the size of the thing she calls a "city"?)
Yeah that's a great question, which as far as I know she doesn't explicitly get into. However she does allow for pretty large definitions of cities: in her last book, which I'm rereading at the moment, she calls Czechia, Slovakia, and Taiwan "quasi-city-states".
I got some push-back in the comments for hypothesizing such a thing on the level of the USA, so I'm glad to know that it wasn't an entirely absurd idea. :-)
I missed any intellectual context for Jacobs's ideas. A brief discussion of Von Thünen's "ring" model, the core-periphery model and Venables's and Krugmans' New Economic Geography would have helped place Jacobs in some sort of intellectual tradition of inquiry.
I was one of the people coming down hard on the ideas in the book. I think it being the first review to be posted was a big advantage in terms of recall value.
I thought it was a really well written review of a book with a lot of ideas that are wrong in an interesting way.
Hello! I'm Michael Zhang, the author of the "Science Fictions" book review (https://michael-zhang.medium.com/trust-scientists-less-trust-humanity-more-9eb1f5af98d4). I'm an astronomer studying exoplanet atmospheres. I was honored and pleasantly surprised to get an honorable mention, so thank you!
For everyone who read the book review, I'm curious what you thought about it (either about the book or about the review). Do you think the replication crisis reveals that something is deeply wrong about science, or is fraud, hype, bias, and negligence just part of the human condition and an inescapable part of the scientific process? Has the replication crisis changed your worldview? Do you agree with Stuart Ritchie's ideas for reforming science? And finally, what did you like or not like about my review (as distinct from the book)?
Since you ask, here goes. I have no particular strength as a writer or editor, so please take everything I say with a huge grain of salt. :-)
I think "fraud, hype, bias, and negligence" are all part of human nature, and the process we call "science" is an attempt to distill truth out of these impure inputs. And this will of course be imperfect: "out of the crooked timber of humanity, no straight thing was ever made". The replication crisis is a sign that not only is the process flawed more than we thought, but that the internal checks are also flawed in ways which encourage other flaws. I can't remember what my worldview was like beforehand, but I do seem to recall being more likely to be credulous of pop-sci claims than I am now, but there were other reasons for me personally to shift away from that, so I can't attribute it all to the replication crisis. FWIW, I'm not a scientist in my day job, but I have friends who are. Given that, his ideas for reform all sound reasonable to me, but I find his omission of ideological bias to be a worrisome blind spot. How do scientists react to the potential of a politically inconvenient truth? Do they seek it out, or do they avoid being associated with it?
As for your review: It's easy to read, and a good size, maybe even a little on the short end (not that that's bad if you said everything you want to say), although again, perhaps this is partly because it's so easy to read. I think it could stand to have slightly more formal structure. In your last paragraph, you start talking about the nature of some of the results, and what that might say for human nature as a whole, and I think this is an interesting thread that you might have touched on throughout the review. I suppose my main criticism is that it's not telling me anything new? I went into it broadly agreeing with the thesis, and I didn't hear anything particularly new (except for that last paragraph). All his ideas for reform sound good, but afterward I still don't know: How practical are they? Which are being implemented, and why those, and why not others? Are there better ideas for reform out there? Who disagrees with what he's saying, and why? If there's difficulty in making reforms, where does it come from and how can it be neutralized?
Frankly, I think you might have run into one of the biases that you write about in the review, "we want to discover new and exciting phenomena".
(Also, there's a 7th sin, envy.)
I liked it! Even though I do agree the replication crisis has been talked to death by now. I remember when everyone was excited about priming, and then a summer camp had us do power poses... it was worth it because I got over my fear of being non-conformist in public, and they gave us a Kindle. But when I heard power poses failed to replicate, I felt vindicated for not having liked them.
I think the replication crisis revealed some major flaws in how journals chose what to publish, which have been partially to mostly fixed since. Humans aren't going to stop being flawed, but with the right process we can get pretty accurate results despite that.
The replication crisis hasn't had much effect on my worldview, or if it did it got lost among all the major changes that were happening in my teenage years.
I'd tell you what exactly I liked about the review but I have no idea, unfortunately. Anyway, congrats! You wrote a thing that's good :)
The throwaway line with the sins annoys me a good bit. Not only is it an unnecessary word sink to list all seven Deadly Sins, not only do you call them the four deadly sins before listing the seven, not only do you say they don't apply when Pride, Greed and Sloth clearly do, but you also only wrote six of them and missed Envy, one of the most scientific-study relevant ones.
With regard to hype, the book is using it itself; the trachea story is horrible, but were people trying to replicate it based on his claims? If not, it's one guy who got caught and the system is working as intended.
My basic thought is that it's really easy to never be wrong, you just stay silent. The method is not "deeply wrong", it's a necessary evil that grows into a larger, less necessary version of itself without constant vigilance. In both directions. Journals are incentivized to publish false positives, but a null journal will be incentivized to publish false negatives; it's all mud forever.
Surely the point of null result journals is providing balance to positive result journals? Right now, if a researcher gets a null result, they're incentivized to tweak it to get a positive result, since otherwise they can't publish. If both positive and null journals exist though, they can just do their research honestly and publish in whichever journal ends up applying. This works even though each journal will only publish either positive or null results.
Researchers will always be incentivized to tweak their results to make themselves look smarter, and no one would want to read a journal filled with "I thought this but was wrong". It would exist primarily as a counter-culture takedown of previous positives.
You're saying researchers will find an effect, but will tweak their results to have no effect just so they can be contrarian?
If they're trying to contradict an existing study, yes they'll do that. Otherwise they'll continue to tweak them to make it look like they got a positive result because the positive result will still be more prestigious.
I think the grammar of the sins sentence is OK? He previously mentioned "four cardinal sins of science", and what they are. He then says the "The four deadly sins of science are not pride, greed, wrath, lust, gluttony, and sloth, but fraud, bias, negligence, and hype." To me, this is saying, "the 4 members of category X are not A, B, C, D, E, or F, but instead H, I, J, and K". This seems fine? There's lots of potential members of category X, and he's ruling out 6 of the 7 candidates that probably come first to mind.
It's not "or Sloth" though, it's "and Sloth": all six are placed in the same grouping of potential Four Deadly Sins. And outside that, there's no reason to specify Four here; the original is Seven, so to preemptively declare there's four hollows the comparison before it's made.
WOW! Thanks all around for the reads (especially everyone who read through to the long, bitter end), thanks for the feedback, and thanks for the votes! I’ll be posting a “Highlights from the Comments on the Review” on my substack. (Never fear, it’ll be shorter. Slightly.)
I didn't think you'd win based on the length, but it *is* an excellent review and your first place is well deserved. I guess this shows how verbosity is very much not penalized on ACX, despite the many comments complaining about it.
Agreed, just I dislike "verbosity" - slightly pejorative: remember the NYT calling Scott "verbose" - though the NYT is nothing but - Oxford dic. says: the fact or quality of using more words than needed; wordiness. "a critic with a reputation for verbosity". ACX-readers like reading well written stuff and to learn along. That is why many of us like your excellent substack! Your place on the podium is well deserved! Bien joué!
Please send me an email when you're done so I don't miss it and can remember to put it on a links or open thread.
Thank you so much for writing that review. It came out right as I was planning how to homeschool my child and I had never heard of Egan before. I'm now working my way through his books for more ideas on how to teach him. I voted for your review as my first choice, since it is the one I am confident will change both my life and my child's life.
This essay might you: a concise critique of the underlying premises of the educational system.
https://sudburyvalley.org/essays/ten-wrongs-cant-make-right
I was *outraged* that Njal's Saga didn't win until I read on.
Where can we read all entries?
See point 1 at https://www.astralcodexten.com/p/open-thread-274
Thanks so much!
I feel like I should have an evening dress or something for this grand occasion, but oh well.
First, I must apologise and not alone eat crow but humble pie and my hat. To everyone who said "Scott has a book review in this contest and it's Njal's Saga", I was wrong and you were right. The deceit, the duplicity, the deception! I am shocked, shocked I tell you!
Now excuse me while I choke down this mouthful of feathers.
I am disappointed Zuozhan review didn't do better, so good luck to T! It was an excellent review and educated me on the subject.
Kill two birds with one stone and wear a hat like Tonto's from the Lone Ranger.
"I’m planning another contest next year. I haven’t decided if it will be book review or generic essay"
I would really love to see a reader submitted generic essay contest, although I suspect there would be 100s of entries since the barrier to entry is so small.
I’d like to see one of each.
> Seth is a chemistry
Typo
Well, aren't we all a chemistry, when you get down to it? The reactions that power life?
Missing word, looks like - Substack bio says "chemistry Ph.D.".
New meme: "I am a chemistry that lives in a society".
So do we have to wait until next year for an AI-written book review to place among the finalists?
Or am I still too optimistic there?
“Njal’s Saga, reviewed by Scott Alexander”
Yeh, it was the best out of a great list. You should think about making a living out of writing.
Gave me a good laugh, thanks
The LRB and other publications that have book reviews as long form essays, allow the writers to review more than one book on the same topic (for instance if there are two books on Stalin, or whatever) - is that possible in this competition, next year?
Yes.
It seems like it was possible this year too. "Man's Search for Meaning" reviewed two books.
I was surprised to see Njal's Saga not in the top 3, based on the overwhelmingly positive response to it elsewhere. Then I scrolled down, and everything made sense again.
Congrats to the winners, and thankyou for the honorable mention!
I finally got around to posting my review here if anyone would like to give it a read: https://open.substack.com/pub/ninedimensions/p/book-review-the-design-of-everyday
If anyone else has published their non-finalist reviews, please share!
I posted my non-finalist review: https://michael-zhang.medium.com/trust-scientists-less-trust-humanity-more-9eb1f5af98d4
I'll give yours a read. Thanks for posting! It's a lot easier than reading the Google Doc (which takes forever to load).
Thanks Michael! I have bookmarked yours to read later too
Thanks for posting all the scores, I was very curious how well I did.
Getting 54th out of 145 is just good enough to encourage me to try again next year. I probably won't be a finalist, but I think I can beat my high score! Though I wish there was a way to get feedback from those low scoring reviewers so I can figure out how to improve.
Could you link us to your review? I could go back and read it & then share some thoughts.
Hopefully this link works.
https://docs.google.com/document/d/1AtGIIv371v0Yu35eNsIxJr67dw4SHOiGdKrqmoKt2hg/edit#heading=h.f78y9pqjkokp
If it doesn’t, I reviewed “The Discarded Image” by CS Lewis
Here are some thoughts, I hope you don't mind that they jump back and forth a bit:
There's a couple grammar mistakes (using principle instead of principal; the second sentence starts with "How could..." but doesn't end in a question mark; TV isn't capitalized; "LIterature" is weirdly written with a capital 'i'). They're small details but when I've submitted scores in the past for this contest I'd take a point off if I remember noticeable grammar/spelling issues. Which doesn't sound like much, but +1 point to your average would put you right up on the verge of being a finalist.
To be honest I felt disappointed by the review. I assumed from the title that I would learn about the mind of medieval people. But what I actually learned felt underwhelming. My main takeaways were (1) they had a certain model of the universe and (2) that model... led them to really enjoy detail?
But I didn't even really understand #2 — one of the examples of "vividness" was "‘So stant Custance and looketh hire about’". That's vivid? (The other examples aren't much better at being vivid...) So overall the thesis didn't really "click" for me. One note I wrote down as I was reading through it: "You spend a long time discussing the ancients' model of the universe when I thought I was getting a review about medieval literature.".
I also felt confused b/c you started by saying they "were nerds"...And then I thought the meat of the review would dig in and show me medieval writers nerding out on King Arthur stories (or whatever) the same way modern nerds nerd out about Superman v. Goku. Who did they write fan fiction about? Whose canon were they debating? I felt like this was an interesting 'hook' but then it never actually got discussed in detail.
I know in Scott's old article on nonfiction writing advice he suggests having pictures, and I tend to try to follow that advice. A couple well-chosen images/diagrams may have given you a tiny boost (and the model of the Universe would be a natural thing to diagram). My general impression of your writing was that it could be "tighter", if that makes sense. Aside from Scott's old article, for writing advice I also like Steven Pinker's The Sense of Style and Julian's guide: https://www.julian.com/guide/write/intro
"references that no longer hold" --> Not sure I've ever heard of "holding" as a verb to use with "references"
One thing I noticed about your writing style, you seem to really like making lists:
How could they make sense of our in-jokes, backwards science, pop culture references, and unspoken assumptions
The fictional settings these characters are placed in have different rules, histories, and laws of physics
The impulse to systemize, to catalog, and to understand is strong among nerds.
"They dedicate hours of their lives to filling fan-wikis, assembling timelines, and collecting reference works."
"multiple authors and ages, Pagan and Christian, Greek and Roman, fiction and non-fiction, "
" where all the muck sinks to and collects."
Not saying it's good or bad, but something to be aware of just in case you aren't. Use it with intention!
Hope that helps! And hope you don't mind the criticism, but you did ask for it :)
PS -- if you want feedback *before* submitting a review next time around, feel free to contact me & I'd be happy to be a "beta reader"
Thank you so much! This is exactly the kind of feedback I wanted.
I did end up leaving out a lot I wanted to include because I ran out of time. Reading your comments, I realize the whole thing was a bit half-baked. I should probably start next year's book review right now, so I have plenty of time to edit it. I may take you up on your beta reader offer!
I had one final thought that goes very nicely with what you've said in this comment. I didn't like the comment you made about being out of time before your deadline in the Conclusion section. On the one hand, I understand what you're going for (e.g., a general "this book is really interesting and I wish I could say more")
But it also feels disrespectful to your readers, in a way. As in, you didn't respect them enough to write your review early so you'd have a reasonable amount of time to make it the best that you could. And if you want to be closer to being a finalist, which it obviously seems like you do, then you should show your readers that respect!
So, yes, I 100% agree — start earlier :)
I was very salty Njals saga lost til I saw the explanation. Bravo
I’m thrilled to hear The Educated Mind won. I enjoyed most of the book reviews in the contest but that one genuinely changed how I think about education and motivated me to dig deeper into idifferent learning and teaching theories.
Welp! I was surprised not to see Njal's Saga in top 3, and low on the prediction market, but this explains a lot. I had no idea it would have been Scott's, I just really enjoyed it.
I traded down on you having written the Njal's Saga review, because you linked to your own channel and I thought that's too obvious
Even if it was a different channel, throwing videos into your book review feels like something that most entrants wouldn't do because they'd assume it was against the rules. Only the guy who makes the rules would be so bold.
The bad news is that next year everyone is going to include videos in their damn reviews.
Shocked that “The Educated Mind” won, it was a chore to get through and it seemed aware of that and felt a need to constantly reassure the reader of the reward at the end (at long last reaching the thesis). Pleased with the other two though, “Cities” was my favorite and “Njal’s Saga” and “On the Marble Cliffs” were near contenders.
I read most of the entries, but I wasn't able to get through The Educated Mind one. It came off as really generic education studies stuff, as I recall.
I'm happy to see that On the Marble Cliffs placed so high, it was my favorite.
I somehow missed the opportunity to vote, despite being a regular reader of ACX. It looks like it was only posted on Monday, and then voting closed on Wednesday. There really needs to be more than two days of voting, especially after the contest itself took place over many weeks.
It was open from Friday the 8th?
https://www.astralcodexten.com/p/vote-in-the-2023-book-review-contest
Oh, I see it was. I think that must have been just after I last checked ACX then. Surely I'm not the only one who sometimes goes a whole *five days* without checking ACX.
Whenever I go 5 days without checking, suddenly I realize I'm 2 weeks behind. :-/
I felt a bit rushed as well - I didn't want to be biased by reading them over a period of literal months, so I read them all in the voting period and only barely made it.
I'm not sure if this is a symptom or our societies polarization, or if it's the fact that this kind of thing being true that is causing it, but what I'm taking away from the comments here is that the most polarizing review was also the most popular.
What's slightly interesting to me is that, given that it was a ranked choice voting system, that _mostly_ shouldn't have been true (or at least, in general on average that shouldn't be true).
I also find that interesting! Educated Mind got second/first place (depending on how you count disqualifications) in all 15 voting rounds.
Vote exhaustion greatly reduces the effect of preference flow (since, y'know, preferences after the third *don't* flow), so that's not especially surprising.
Also a ton of the votes were thrown out.
I think it mostly shows that the comments don't reflect the voters
Yes, it shows what has been a consistent feature of this blog for at least as long as its been ACX: The readership and the commentariat are two different beasts.
Isn't that always true across all internet communities? The lurkers outnumber the posters pretty consistently.
My question is if it also holds in real-life communities. I feel like when I'm in small gatherings in real life there are fewer permanent lurkers, but it might also be a function of size. It's harder to hide in a group of ten than in a group of one hundred.
My point isn't that propensity to comment isn't uniformly distributed among readers of ACX. It's much more specific: The population of ACXers who post are different across a number of characteristics from the population of ACXers who read but don't participate in the comments. The example that springs to mind from one ACX survey or another I remember is political leanings. Commenters tended to be more right wing than readers.
Reminds me of a game I played and worked on as an open source dev where the entire forum and dev community was focused 90% on MP play, but if you looked at stats 97% of the actual player base was SP only.
Was that by any chance freeciv?
No but very similar project (Wesnoth).
As someone who reads regularly but rarely comments I can confirm this.
Ranked choice is actually really good at finding polarizing choices. Every step asks which options got the fewest first place choices, then kicks them out. The end result is that a option which is everyone's second choice gets removed instantly; the only thing that matters is the number of diehard fans.
This is one of the reasons I dislike ranked choice. Approval is, honestly, better.
Approval voting gives you the contestant who annoys the least amount of voters. The problem here is the least known contestants can often end up winning; people seldom disapprove of what they don't yet know of.
Plurality voting (FPP) encourages the contestants to define themselves as what the other is not and for most contestants to withdraw from competition so that only two realistic choices remain.
Ranked choice is somewhere in the middle: allowing some polarisation (you must be well-known to win), but not so much that you discourage too many people from giving you a high secondary ranking.
The middle way is always best.
Aristotle was very wise.
All hail Thanos.
You assertion about approval voting suprises me; in the few votes I have seen using it, if I don't know who someone is and what their capabilities are, I don't vote for them.
Do you have studies I can read?
Sorry, I don't. Just experience from watching a lot of elections; a relative unknown can get some lucky favourable media coverage (for example at a debate), that's most of what people know about them, then win and have quite disappointing performance in office.
The archetypal example here in NZ was Peter Dunne.
IRV with vote exhaustion and most-votes-wins (which is what was done here) is far worse about this than full majoritarian IRV (in which you only need a fairly-small primary vote to snowball to victory).
Do not conflate ranked choice with instant runoff; there are other (Condorcet) methods that still use ranked ballots but do not have the unique flaws of IRV.
ETA: everyone's second choice, if first is fragmented, will absolutely prevail under Condorcet.
No, no, no, no, no! Sorry, but this is really bugging me: this was NOT a ranked-choice voting system. And it wasn't proper instant-runoff voting either. A real ranked-choice system lets you rank ALL the options (at least if you want to) and counts all of your preferences. If the majority (as I strongly suspect) would have preferred "On The Marble Cliffs" to "The Educated Mind", a real IRV count would have redistributed the preferences from all other entries (once they had been eliminated) between those two, and the former would have won.
How it was actually counted was MUCH closer to first-past-the-post than to actual ranked-choice. I don't begrudge this at all, as I understand a ranked-choice vote would have been difficult and time-consuming to count. And also, I would have had no problem with just using FPTP in the first place.
But for the love of God, please let's not pretend this was a real ranked-choice vote. It should be described as either "modified FPTP" or "very limited RCV/IRV" or something similar. And being in effect a form of FPTP, it's not at all surprising that it has that system's same problems of electing highly polarising options.
I'm not sure I understand. What is your understanding of how this voting worked, and what specifically did it do or not do that makes it not count as ranked choice?
Okay, imagine this was the first-preference vote (totally made up, and using only five contestants):
The Educated Mind: 88
Cities and the Wealth of Nations: 50
On the Marble Cliffs: 32
Man's Search for Meaning: 19
Zuozhuan: 11
Case 1, under first-past-the-post, there are no further rounds or preferences, and Mind wins.
Case 2, how it was done with only three preferences allowed: Zuozhuan is eliminated, its voters all had Meaning as their second choice (they like epic historical stuff), which now has 30 votes, still the lowest, and is eliminated next. Those 30 votes flow to Cliffs.
Mind: 88
Cliffs: 62
Cities: 50
Now imagine all the Cities voters had Zuozhuan as their second choice, and Meaning as their third. Since those have both been eliminated, and they were only allowed three preferences, those votes all exhaust and have no further effect. Cities is eliminated, and Mind has 88 votes to Cliffs' 62, and Mind again wins. None of the 112 people who didn't vote for Mind get to express a preference between Mind and Cliffs (just like they didn't in case 1).
Case 3, proper instant-runoff where people can express a full set of preferences (from 1 to 5). Same as above, up to when Cities is eliminated. This time, those Cities voters have their fourth preference read: 45 of them had Cliffs ranked fourth, and Mind last. Those 45 votes flow to Cliffs, and in the final round...
Cliffs: 107
Mind: 93
...Cliffs wins this time! Why? Because of the 112 people (the majority) who *didn't* vote for Mind at the start, almost all of them (107) preferred Cliffs to Mind. Or, more roughly, of those who didn't *love* Mind, almost all of them *hated* it, and they were a majority.
Both FPTP and only-3-preferences "IRV" gave no weight to that majority's preference for (let's assume) *all* other entries over Mind. Thus, I'm claiming both case 1 and case 2 (where case 2 is how the actual contest was counted) did not allow voters to fully rank their choices, and this (in my hypothetical, and very plausibly in reality) disenfranchised the actual desires of the majority. Thus I don't think it can fairly be called a ranked-choice vote. It functioned similarly to first-past-the-post, favouring a polarising candidate that got a lot of love and a lot of hate, over one with much broader appeal. Only a vote with full ranked-preferences would have favoured the latter.
Let me know if there's any part of this you didn't understand, or that you don't agree with, because I *really* want people to see this point. And in particular to see how much the details of the voting system used can change the outcome.
It's even more muddled than that because The Educated Mind didn't get the most votes, Scotts review did, and was then discarded. I'm not sure where the Saga votes ended up.
I don't think it's a big deal either way, but yeah, this shouldn't affect your view on ranked choice or whatever.
A better way to collect votes (with google forms) might be to have a list of all candidates and ask the voter to establish a preference ordering by assigning arbitrary points to them, with the understanding that the sorting of the candidates by these points (with ties randomized) will establish their preferences.
Just allow them to put a floating point number for every candidate, and treat any empty fields as zero (and discard the votes of anyone who assigns not-a-number). So if you like two candidates and really hate another one, you could just assign 2, 1 and -1 to them (leaving the others at zero) and be done with it.
He even posted the Video on his own YouTube Channel he has used for SSC videos in the past. He wasn't trying his hardest to hide it. In general, it was the funniest review by far, though I only ranked it 3rd because two others were more informative.
Scott: Now that it's been established that it was you, can you explain your thought process behind *why* you didn't comment even in passing on almost any of the parts of Njál's Saga for which it has been famous and beloved for centuries? I get that the point for the review was essentially to crack a joke about legal systems very different from ours, but surely the concept of a review would normally oblige someone to engage at least cursorily with the substance of a book? I guess most of your readers haven't read Brennu-Njál and couldn't tell, but reading the review as someone who *is* familiar with the book, seeing not one word about e.g. the attack on Hlídarendi in favor of the ridiculous legal quibble parts nobody cares about is almost surreal.
I guess I didn't find the attack on Hlidarendi as interesting as you did.
This sounds to me like a more polite version of "fuck off and shut up", and I respect that I guess, but regardless of personal tastes, can you, having read the book, nevertheless apprehend that a review of Njál's Saga which mentions the name "Gunnar" one time in a parenthesis is in some objective sense... not really focused on the same things that the actual book is? I think if I hadn't read the book, and then read it based on your review, I wouldn't have expected or even recognized the book I got at all.
After looking at The Educated Mind, I agree with the other commenters here that it was terrible, quite possibly the worst review of the contest. I'm amazed it got any votes, let alone first place, and that's saying something since I thought that nearly every entry deserved to win this year.
Huh, I didn't realize Scott had reviews in the contest. That sorta distorts the betting market.
> Njal’s Saga, reviewed by Scott Alexander. This one got the most votes, but I’m disqualifying it because it seems in poor taste for me to win my own contest.
What a chad move.
I am SO happy Njal's Saga won. Had no idea it was you, and I loved it so much.
Now I can't decide if intentionally not voting for Njal's Saga cause I was pretty confident that was the True Scottsman actually made sense. I guess it at least ensured none of my votes got discarded? But it was the most entertaining, well-written (even by Scott standards - felt like "classic" SSC-era), and breaking the nonfiction monopoly is a big plus in and of itself. Not expressing those preferences feels...inconsistent.
The Educated Mind...would have expected it to top5, given the demographics, but not actually win. The length and imprecision were real bummers, even though I overall enjoyed and "liked" it. I think I'd have given up if not for priors being adjusted by FdB and Caplan - that is, having the expectation that ponderous ed theory critiques can actually be enlightening and moderately entertaining, so it's probably worth slogging through. The juice was worth the squeeze, at least in strict marginal idea-profit terms, if not best-uses-of-my-time utility.
Main bucked expectation for me was Why Machines Will Never Rule The World finalizing at all. I guess there's a lot more community (or voter anyway) split on AI fundamentals than I thought.
First of all, I loved the Njal's Saga review and I agree with other commenters who said that it was by Scott Alexander. I also really liked the Romance of Alexander review, and I said to myself, that one also looks like a Scott Alexander review! I couldn't tell which, if either, was by our illustrious host, or whether there were contestants who were just really good at imitating Scott's voice. (What is the kabbalistic significance of Scott Alexander reviewing the Romance of Alexander?)
Second, The Educated Mind winning first place was... well, it was a choice.
I can't say I *disliked* the review, because it had some excellent parts and made very good and important points, especially about how success in public education is difficult/impossible, because we're asking the education system to do different/conflicting/incompatible things: socialization, academics, and individual development.
But it was a flawed review for several reasons, and not just because it's so fricken LONG. I'm not going to reread it now, but Brandon Hendrickson seems to claim that True Education is about taking a child through five developmental stages, and then he gives examples of what to teach at each stage.
Which, ok, sure, but what is the evidence for the five stages and their correspondence to the child's age? I don't recall real evidence, mostly lots of anecdotes along the lines of "seven-year-olds like puns and jokes, which correspond to the XYZ stage." Also, I don't recall Mr. Hendrickson attempting to reconcile these five developmental stages with the socialization/academics/development (SAD) model - which sides of the SAD triangle are we aiming for with the five-stage model, and why? Next, the five-stage model seems to ask a LOT of teachers. They're supposed to not just know a very broad range of material, but also be gifted comedians, storytellers, entertainers, etc. Sure, this works if you're paying lots of money to recruit the best teachers possible to a small school attended by the children of the PMC, but how well will it scale across the whole society?
Finally, as other commenters on that review already pointed out, the review, for all its length, has some very important gaps. Mr. Hendrickson spends a lot of time on the proposed curriculum, and very little on things like testing, classroom discipline/behavior standards, tracking (do we group by age or ability?), etc. What to do if, despite our best efforts, some children just can't/don't want to progress through the "five stages"? And if the reviewer did write about these things, then I apologize for not remembering, but again, if you wanted me to remember more, you would have made your review shorter.
Again, I don't mean to sound harsh, you obviously put a *ton* of work into your review and it has a lot of very good things about it. Congratulations on your win!
Could you help with a link to Scott`s-The Alexander review? I somehow made it to a google-doc with review-entries from A to G - but not "T". (I agree with the points you make about the winner, btw.. I was very surprised it won. But kinda pleasantly surprised. If ACX-readers are ready to go for such lengths, I feel in good company.)
I assume/hope Scott will post it on the blog at some point, but it's also available here: https://docs.google.com/document/d/1AtGIIv371v0Yu35eNsIxJr67dw4SHOiGdKrqmoKt2hg/edit#heading=h.qku6a39xims7
A good way to get a link to any review from the three years of the contest is: https://codexcc.neocities.org/
Aaaah, Merci & Dankeschön! Could not take the wait; now I feel as Balthazar Picsou does, when he is diving into his pool of gold-coins. Alexander Romance: a very Scott'ish review that is, again. Njal was better poised to win the contest.
I’ve noticed that Jaynes’ Probability Theory has been reviewed each year for three years in a row, but never does well. Given that it’s a core text for the rationality community, I’m curious if many ACX readers have read the textbook? My sense is that it seems a bit like a book that people talk about without having actually read it, and maybe for good reason, since the author died before completing it and a lot of it is half done.
I haven't read the whole thing, but I've used parts of it as a technical reference for work. I....really wouldn't have expected it to be the kind of thing that one reviews for this kind of contest or, if it was reviewed, for it to do well. It's basically a stats text book, admittedly an important one.
Congrats to the winners! I'd also like to commend T. for her Zuozhuan review, which was my favorite.
Thanks to Scott for putting this on, and to everyone who read.
1. I really approve of both using ranked choice voting and limiting it to three choices.
2. I gave my first choice to Cities and the Wealth of Nations and I'm glad to see it place.
3. Of all the book review finalists the only one that would annoy me by winning was the one that did just that.
>1. I really approve of both using ranked choice voting and limiting it to three choices.
Why? It doesn't get the Condorcet loser criterion the way true IRV does. Indeed, Educated Mind was quite possibly the Condorcet loser, so this is in contradiction with your third point.
More potential for tactical voting when you can be exhaustive, but mainly it's more mentally taxing to rank as many options as you want.
I am fairly sure that the chance of tactical voting helping is much greater in this method than in full IRV, because you have to choose between honest preferences or candidates with a chance (like in plurality). IRV does have tactical voting, but it is nowhere near as bad as plurality, and as essentially an IRV-plurality hybrid this is partway in-between.
What mushroom said, with the added point that tactical voting in real IRV (which this wasn't) requires highly accurate information on not only which options will be get the most votes and what order they'll be, but also what the lower preferences of those options' voters will largely be. Infinitely more difficult than in most other systems.
And also: how can having more choice *possibly* be more mentally taxing? Just limit yourself to 3 voluntarily if you find it hard. Why limit other people's choices?
I posted my review here:
https://tapwatersommelier.substack.com/p/book-review-mans-search-for-meaning
In a week or so, I will write a post answering to the comments from the original posting in June.
Bit of a meta-comment: In terms of sheer quantity of text, these 15 reviews must be the most I have read from a single source (/collator) in a long, long time. I used to read this quantity of text from The Economist when I cared more about the news; occasionally from the FT, NYT, or New Yorker.
What I mean to say is that Scott is now, for me, as good a source of interesting writing as *the best* journal publishing houses in the world. That's an extraordinary state of affairs, and I hope someone studies it one day. But not Scott, because I just want him to carry on with whatever he's doing.
Congratulations to all the reviewers, I thought they were all world-class pieces of writing.
I fully agree with you. I had the same realization during the course of this year. Trying to read other blogs and sources, but nothing comes close to consistent quality of reading on ACX. The rest is mostly commentary, as they say.
Thank you!
I'm surprised at the amount of negativity directed at the Educated Minds review in the comments here. I found the book's outlook on education both exciting and sensible, enjoyed reading the review, placed it as my first choice, and am happy to see it win!
In general, I liked the contest a lot, thank you to all the reviewers for your efforts!
What an excellent way to leverage skill at edutaining children.
In defense of "The Educated Mind" book review:
I see a lot of comments that say that the review was written terribly, rambling, teasing and generally confused how it won.
As someone that voted for it as a second preference, (Njal's Saga as first), I figured I'll say why I liked it enough to vote for it.
For me, I enjoy a book review when it introduces me to new ideas, and does so in an enticing and interesting way. I felt that TEM book review took me on an interesting journey, introducing a new way to consider education and adding a new lense to view interacting with children and teenagers.
This was especially interesting to me as it fit in with what I have actually experienced myself, and can see with my children: my grade school child is fascinated with stories, and she can spend 40 minutes reenacting all the parts of the stories she heard at school, but can barely recall other lessons. Meanwhile, her older brother is really getting interested in extremes: highest mountain, lowest place on earth, etc. What I'm trying to say is that as I was reading, the ideas resonated, and seemed to fit my observations and place them in the context of a theory.
I also enjoyed his explanation of the problems with current education, with different types of schools satisfying two out of three "ideals" and leaving out a third, important one, and found it to make sense. Especially given how drastically schools can vary (from personal experience). The review left me with hope that an education institution that isn't completely broken is possible.
As for the rambling and teasing, etc., I didn't mind that so much; other overly-verbose reviews gave me more trouble, and a couple just lost my interest halfway. This one did not. A book review that delivers so much of the book's content makes me happy with it. I also felt that the review addressed many of the objections, concerns and questions that popped into my mind as I was reading it. I spoke about the review with friends and recommended it often enough that it felt to me that it makes the cut as a solid second choice.
As a last note, while I liked "Cities and the Wealth of Nations", I didn't want to vote for it, as I felt economics reviews were overly represented in this competition (likely due to last year's winner), at five out of sixteen finalists. But I do agree that CATWON was the best economics review.
I wonder if people with children / significantly younger siblings were more likely to vote for it. Did anyone also have this experience?
I agree totally. It was the review I took screenshots of to send to my partner with captions of my own reflections. If a review does that then there is clearly something special in your experience of it.
I mostly based my voting on the ones that I could clearly remember - this may well have biased me towards later ones, but I think moreso just being unusual. I felt like I had to vote for On The Marble Cliffs because it made me actually download the book, I plan to read that some time when I'm in the mood for Literature.
I'm surprised more people aren't declaring their votes. Well, here are mine:
1. The Rise and Fall of the Third Reich.
This was terrific and hugely underrated. Very much what I'd like to see more of here (in comments and even in some posts by Scott) and elsewhere: clear, logical grounding of relevant facts, and then detail-focused, logically justified conclusions from those facts. And especially, solid grounding of claims about the present and predictions about the future in the parallels and principles from actual history.
It's amazing to me how little of this there is, and how much people rely on Sprawling Theories of Everything, attempted grand explanations of the past and the future and the nature of society and life itself, justified by little more than "this just seems like something that might be true". Even in purportedly "empirical" communities like this. JJ's review instead does the radical thing of answering the question of what might happen, or what we should fear happening, by looking at what has, you know, *actually* happened, in the past. And getting a grounding in that, picking out the important details, and drawing out general principles that we can apply to other contexts.
I'm not surprised it didn't do well in the contest, though. The fact that so many people love their Sprawling Theories of Everything, and their gut reactions that "*clearly* candidate X is the new Hitler, that's just obvious, I don't need any logical argument or actual knowedge of Hitler to make that claim!" means they won't be very amenable to tightly-focused and clearly-argued pieces that pay attention to details and distinctions. Please write more pieces like this, JJ. Even though the people you describe in your first paragraph are most of the internet and a large portion of the voters in this contest, there are definitely those of us who see its compelling explanatory insight.
2. Man's Search for Meaning.
A very courageous, sweeping essay, complete with narrative twists, integration of multiple historical contexts with one another and with the present, and of course a focus on the *truly* important questions. As admitted in the essay, it's a bit of a mess in terms of focus and structure. But polish it up a bit, and it could be a total masterpiece.
3. Njal's Saga.
Incredibly entertaining and well-written. I bumped it down slightly for being gimmicky: the multimedia youtube aspect and trying to copy Scott's style. But I then bumped it up for *succeeding* at the latter. In light of the truth...the last two points are invalidated I guess, leaving "entertaining but a bit gimmicky". Please, no one copy the second part next year, thinking that's what made it "win". It's not.
If I'd been allowed a fourth vote, it would have gone to ZuoZhuan. This one had, in fact, the good aspects of each of the above three respectively: clear and solid, sweeping scope, and entertaining style. It just had each of them to a lesser extent, making this a hard choice for me. In the end, I decided it was probably written by an academic or other expert professional, and that I wanted to give my votes to less credentialed contestants. In light of the biograohical information, it seems I was wrong about this, and maybe this deserved an even higher place.
Thank you! I also really enjoyed "The Rise and Fall of the Third Reich" review, and it was one of my votes.
"The Rise and Fall of the Third Reich" was the one I liked least. As a German with a minor in history , I cringed at a reviewer who read just ONE single book about the period - and one that is really not up-to-date. The author retells a lot of tales about Hitler that later turned out to be just fake-news (e.g. Hitler was NOT the 7th member of the party), and is ignorant of a lot of analytical work that was simply not yet done at his time. Knowing no other reputable book about the topic before writing a review about this one is ... misled and misleading. (I do not say the book is worthless, far from it. Nor was the review badly written. Far from it! But was it competent enough to be published? Sadly, far from it. - Sadly, I am no expert to judge the accuracy of other reviews, I all enjoyed and felt enlightened by. Maybe wrongly, too.)
The complaints against the winner need some perspective - even with ranked choice voting most people will have voted against the winner. This would be true of most winners given the number of candidates.
For sure, but I think if you had some sort of scoring system it would have had one of the lowest averages by far.
I am delighted by two of the honorable mentions: Prince of Persia and the Design of everyday things.
I am bit sorry for Lying for Money, which was very good, and Zuozhuan, which was excellent.
I wanted to mention that I thought the Zuozhuan review was excellent - my personal choice for the best among the offerings.
If there is an official list of haters of The Educated Mind, add me there.
At first, it makes some good criticism of the existing educational system. But, as is the usual problem with populism, the actual challenge is not telling people how the existing system sucks (most people know already), but designing a system that sucks *less*.
The proposal, in a nutshell, is to tell stories at the elementary school, interesting trivia at high school, and leave the boring stuff (such as addition and subtraction, I suppose) for university. Yeah, that would definitely improve education a lot!
Seriously, if you think this is a strawman, then I challenge you to show me a mathematical curriculum designed to comply with The Educated Mind, and put it side by side with the existing one.
The nearest thing that exists in reality is probably Waldorf schools. If you look at the lists of their famous alumni, you will find approximately zero scientists, which is not a coincidence.
Having looked through the comments section for that book review, I’m wondering whether the described system (and the layout of the review) appeals especially to intelligent, academically oriented people with ADHD or similar executive functioning difficulties. I was always good enough at math that I didn’t struggle in the subject like a lot of kids with ADHD did, but I found the way it was taught dull enough that it put me off pursuing a career in mathematics or physics research. In hindsight, high-level mathematical research seems to involve more of a creative problem-solving approach that seems like a good fit for people like me. The same is likely true for many academic research fields and engineering careers that involve a design element.
I get the general impression that disciplines always seem rote/mechanistic at the training-wheels stage, while the fun/creative parts manifest later. Math however, seems deep enough that even undergrad math is still the training-wheels stage. I suspect this prematurely colors people's experience, and is partly responsible for the widespread "not a math person" vibe.
You are criticising the book not the review. That no such curriculum exists, doesn’t mean it couldn’t.
Absolutely wonderful as always, these book reviews always manage to concentrate so much good stuff on so many topics. I wonder, are there links to the Google Docs containing the rest of them? I see one in this post, but only for the A-G section.
In anticipation of maybe taking part in next year's contest, does a review have to be of a very recently published book, or would one published in say 2016 be OK, or 1916 for that matter?
One of the winners was published in 1939, so...