Where sports, statistics, and literature meet: Comparing David Fleming to Bill James
Statistical Modeling, Causal Inference, and Social Science 2025-03-31
Jonathan Falk writes:
In honor of the start of the baseball season, I am reading David J. Fleming’s Baseball Obscura 2025, a book in the form of the sorts of essays that Bill James wrote in those original Baseball Prospectus’ back in the late 70’s-early 80’s. Fleming worked for James’ website for years and has a firm grasp of the aura of a Jamesian essay on, say, the Padres, and the style is consciously Jamesian, but I am reading these essays dutifully, and not with delight. I am not asking you to read it (though you are of course free to do so — who the hell am I to tell you what to read?) but I see at least three possible reasons that I’m not transported, which in turn correspond to some of your posts on literary topics.
(a) Imitating James isn’t the same as being James. No matter how much you admire James and no matter how much you copy his style, he’s Bill James and you’re not. In literary works, influences that are too strong come off as imitations (or even worse, parodies).
(b) Bill James was teaching while entertaining. And the insights were genuinely new. We can now apply the tools (and their offshoots) but the delight of discovery is gone. In literary fiction, the first unreliable narrators (Don Quixote, Tristram Shandy, on up to Holden Caulfield and Humbert Humbert) provide some shock of the new which is sadly absent when the method is applied again. (I realize that I’m contradicting myself slightly here by citing works with hundreds of years of difference, but I hope you get what I mean.) Novel means surprise.
(c) Maybe the zeitgeist has moved on and Jamesian essays need to be replaced by something else. What was great in an era of statistics laboriously compiled on legal pads by hand has been supplanted by arguably more informative (and arguably not) but definitely more easily generated data summaries generated from better data sources by computers. Sue Grafton refused to age her detective Kinsey Milhone into the age of cellphones because the existence of cellphines and computers would corrupt what she saw in the detective process. Of course, novels are free to be embedded in any era, but the zeitgeist moves on. One wonders what psychoanalysis or a modest dose of Prozac might have done for Raskolnikov, and you can’t help but wonder while you read Crime and Punishment, and someone who writes today and sets his works in Czarist Russia is tainted by that anachronistic knowledge.
Baseball season cannot start fast enough…. I’m going stir-crazy.
I too have fond memories of reading the Bill James Baseball Abstract every spring . . . I think it was 1986 that I first discovered it, then I went back and found the old ones and read them too. It was tragic when he stopped coming out with them every year.
I have a few things to add to Jonathan’s explanations;
1. There’s a limited number of things that can be said in that classic Baseball Abstract format. Even James stopped doing it after a few years! So when you say it’s too bad that his follower can’t do it . . . well, maybe nobody can.
2. As Jonathan says, we’ve all moved on, and as readers we want something new. Even in a robust genre such as Agatha-Christie-style mysteries, we demand some novelty in form or subject matter. We wouldn’t want too faithful a reproduction.
3. In his later baseball writing, James started to b.s. a bit. I think he became a bit too believing in his own pronouncements, and he also became too much of an insider for my taste–I talk about this a bit in my essay, A Statistician Rereads Bill James. But this may have made his writing more readable–he had the flexibility to say whatever came to mind. In contrast, a follower of Bill James could be more constrained by the evidence.
4. The other constraint is . . . well, I went online and looked up David J. Fleming’s Baseball Obscura 2025. It has some sample pages, and, yeah, they’re pretty much exactly in the format of those Bill James Abstracts from the ’80s. Maybe that’s too much of a constraint too. You could say that James was himself constrained in that way, but not really: it was the format he chose, and he had the flexibility to change it if he wanted.
Beyond that, it’s hard for me to judge. I was much more of a baseball fan in the ’80s than I am now, so most of the names in this new book mean nothing to me. That said, one reason I remained such a fan during that past decade was because the Bill James books sustained my interest.
Overall, I find those pages of Fleming’s book to be pretty compelling. I think I’ll buy it. Maybe it will get me more interested in baseball this summer. Also I’m sad that books aren’t what they used to be. It’s not like in the old days, when I could stop in the bookstall in the train station and pick up the new Baseball Abstract every spring.
P.S. Since we’re on the topic, here are some earlier posts:
– How did Bill James get this one wrong on regression to the mean? Here are 6 reasons:
– Bill James on secondary average
– A collection of quotes from William James that all could’ve come from . . . Bill James!
– Bill James does model checking
– Bill James and the base-rate fallacy
– Why can’t I be more like Bill James, or, The use of default and default-like models
– Confusion over British elections and the relevance to political punditry of a dictum of Bill James
– Maintaining competitive balance in basketball: I disagree with Bill James