Bonus Podcast: Nate Silver Talks with Steve Kettmann
This conversation will also air in this week’s episode of “Hot Takedown” — FiveThirtyEight’s new sports podcast. Subscribe on iTunes here.
The New York Times on Wednesday published an op-ed with the headline “Don’t Let Statistics Ruin Baseball.” Some of the baseball nuts at FiveThirtyEight staff couldn’t contain themselves after reading it. Here’s an edited transcript of our Slack conversation.
cwick: This is the kind of thing that gets our attention around here. A sample of the op-ed: “Thanks to ‘Moneyball’ and stats-driven fantasy leagues, advanced statistics have changed how fans think about the game. On the whole that’s a positive trend — but not when the numbers begin to eclipse a more nuanced appreciation of baseball.” Nate, you created FiveThirtyEight for all sorts of reasons, but surely it has no more sacred duty than to respond to this piece. What’s a sabermetrician to think about this op-ed?
natesilver: One litmus test for any writer: Does he know his subject? Steve Kettmann might know a lot about baseball, but he doesn’t know very much about baseball statistics. For instance, he describes how wins above replacement compares a player against “some hypothetical median player, the ‘replacement.'” Actually, it compares him against a replacement-level player, who is way worse than the median major leaguer.
He doesn’t describe fielding independent pitching (FIP) correctly. Its goal is not to provide a “broader measure of a pitcher’s performance than the traditional E.R.A.” Rather, it’s to measure how effective a pitcher is independent of his defense (fielding) — as its name implies!
On the other hand, he favorably cites “a previously obscure statistic: batting average against relievers.” That statistic hasn’t become much less obscure — it’s not something that we stat geeks are talking about very much. And that’s because we know it’s probably just random noise. Batting average is noisy enough, let alone when you cut the sample size by two-thirds.
The point being, the better advanced statistics are all about deepening our understanding of how the game is played. If a pitcher has a 2.85 ERA, is he really good — or is it the defense behind him? That’s what FIP can tell us (although there are better ways now to account for the impact of positional defense).
By contrast, when statistics were cited in, say, the 1980s or 1990s, they were mostly just trivial crap. “Claudell Washington is 6-for-25 on cloudy days against left-handed relief pitchers with runners in scoring position.” Shit like that. Totally useless and distracting. And there were dozens of them on every broadcast.
carl: Nate’s response shows how these debates are always about which numbers people arbitrarily approve and which ones they don’t. You can’t have baseball without numbers. How would you count strikes, outs, runs and wins?
cwick: Right, there have always been stats in the game because the game is such a structured one. Baseball happens in discrete steps. There’s an action and there’s an outcome. Pitcher throws, it’s either a ball, a strike, a foul, a hit, or an out. (Or some other possibility I’ve overlooked.) That creates stats that are robust and complete. Stats can’t help but bubble up and out of those.
benm: As I understand it, he is all for statistical advancement of baseball to enhance understanding of the game and quality of competition and other more utility-based purposes, he just objects to the obsession with statistics from an aesthetic perspective. He compares it to listening to a symphony: There’s nothing wrong with understanding how the tuba works or how many clarinets you need, but when you’re there in the audience, you should just let the music wash over you and appreciate it.
I find watching baseball tedious in part because I feel like nothing I’m seeing in front of me matters — and I can trace this back to when I was young and learned the factoid that the difference between a good hitter and a bad hitter was “one hit a week.” So if I had a more robust appreciation of the non-statistical aesthetic quality of baseball, it might make it easier to enjoy a game.
But the other way it cuts is that, without statistics, I probably wouldn’t be interested in baseball at all. Like many people, my fandom started out with baseball cards. Fast-forward 30 years, and while I’m not as nutty about the game as some of my colleagues, I still take time to follow the fascinating statistical developments in the league, and can appreciate Mike Trout or Billy Beane’s greatness in a way that has something of an aesthetic aspect for me.
carl: Like with so many of these pieces, the headline and opening are more radical than the author seems to really be. For example, this is pretty hard to argue with, unless you’re a straw man: “There is a risk that numbers become an end in themselves, and arcane stats proliferate. A good rule of thumb is that the more a stat relies on abstraction, the less likely it’s going to be consistently useful to a wide audience.”
rarthur: I agree with @carl. These pieces always boil down to, “This is the exact, right, perfect amount of numeric detail that should be permitted in the game.” And it’s never clear why we should have stopped at batting average/RBIs/saves, as opposed to moving on to more accurate and more descriptive statistics.
carl: The related point is that the stats that seem “traditional” (to use his word) just have to do with familiarity, not how arcane they are. The definitions of saves — which itself seemed radical to some when it was introduced — RBIs and pitcher wins are arcane and arbitrary.
cwick: @carl, I hear you, but I think there are some passages in here that are pernicious in their misinterpretation of sabermetrics. “When it comes to watching a matchup of, say, the Mets pitcher Matt Harvey and Giancarlo Stanton of the Miami Marlins, statistical analysis is about as helpful in deepening an appreciation of the human drama unfolding before us as it would be for a Pavarotti aria.” PITCHf/x data and Stanton’s BABIP have a ton to tell us about that matchup
carl: @cwick, that’s only pernicious if we assume that statistical analysis can’t help us appreciate an aria. The jury is out on that. Where’s opera’s Bill James?
cwick: @carl that’s a story for you!
ollie: You’ve all nailed the statistical points. I’d like to mention something else. Kettmann seems very concerned with our attention spans. Let’s take his analogy of music, and going to the symphony to hear Mahler’s Ninth. I need to pay attention. To clear my mind. How would I enjoy it otherwise? But if I knew that disease plagued Mahler’s mother and 13 siblings, or that he had an abusive father, or that he was obsessed with death, and that the last line of that symphony’s score reads “ersterbend” — German for “dying away” — I’m going to appreciate the piece of music more, not less. This echoes how I feel watching Anthony Rizzo bat knowing that he had an OPS+ of 163 against lefties last year, or whatever. Statistics are a complement, not a substitute.
rarthur: Right, @ollie. Above and beyond that, I would never presume to tell a music theorist that it was incorrect to be thinking about notes and scales while listening to an aria. You can enjoy it the way you want to. If you prefer not to know the statistical underpinnings of baseball, no one is forcing you to learn them.
natesilver: @rarthur: Actually the music theorists are pretty damned interesting when you corner them at a party!
carl: @ollie and @rarthur — totally agree. His implication is that quantitative context is less valuable — less artistic, less charming — than qualitative context. That’s personal taste.
cwick: Well, Kettmann’s point is that it’s becoming increasingly difficult not to pay attention to the stats, @rarthur. Which I sympathize with — I’m the words guy at a numbers site! But his argument comes across as a Luddite’s. “The art of hitting a baseball starts with emptying the mind. As Jonathan Fader, a psychologist who works with Mets players, told me: ‘Essentially, what we’re trying to do in sports psychology is helping people to not think.’ Fans and writers need to adopt a similar attitude.”
natesilver: Some of this is just plain old anti-intellectualism. Us nerds want to understand how the world works. Statistics are one way to do that, but not the only one. @carl, you’ll remember when we were at the US Open last year and I was constantly peppering you with tennis questions. Some of them were stats-y questions and some weren’t. I just wanted to learn more about tennis. I’m curious about the world, both as a “fan” and as a journalist.
carl: Yeah, @natesilver, I think we nerds are fine with people who don’t want to understand the sport or opera or whatever they’re watching. They should be fine with those who do. We can all sit side by side and enjoy the game. Except for those terrible tweeting/texting/Googling press-box reporters.
ollie: Or the guy on his phone at the symphony, @carl.
carl: @ollie probably checking operareference.com.
carl: Which is available. We should buy it before Sean Forman does.
cwick: opera-reference.com, obvs. (God, that might be the nerdiest joke I’ve ever made.)
carl: Also available.
benm: I know that a musician may appreciate Bach in a way that I can’t. But I think the visceral experience is still very immediate for them, so I don’t think that would be inconsistent with Kettman’s argument. I presume that in his ideal world, we all understand these stats but don’t let them be the main thing we appreciate when watching. They should be more like butter or MSG than like steak or fish.
natesilver: @carl: I’m totally fine with the dude who just goes to the game with his friends and drinks about 5 beers and cheers when the home team hits a home run. Sometimes that’s me! I’m NOT fine with journalists who are incurious about the world, however.
david: I think Ben hit on something important. This argument — and I can understand both sides — reflects opposing approaches to life. Some people appreciate these events — whether Mahler, Springsteen, or baseball — at the purely sensual, aesthetic level, and don’t see any need for their cerebral cortex to interfere in that appreciation. Others need to find deeper patterns and hidden movement in everything they see, and those patterns are usually there to be found. Each approach is legitimate, but the two sides will never really understand each other.
rarthur: I would challenge the idea that the two sides can’t understand each other. At its best, baseball (or perhaps sports more generally) bridges the gap and allows you to experience both a statistics-driven and purely aesthetic enjoyment. In the World Series Game 7 last year, as Madison Bumgarner pitched inning after inning, I was simultaneously trying to calculate whether it was a good idea to bring him back out and also just enjoying how awesome, how calm, how controlled he was. And I was thinking about him pitching up in the zone, and about how Sal Perez was tired and injured, and everything all at once.
natesilver: On one level, I agree with Kettmann about the aesthetics of actually going to a sporting event. Some of the scoreboards — like the new Jumbotron at Wrigley — are overkill. They rarely abide by good principles of information design, which usually means clean, somewhat minimalist presentations that fit naturally into their environments.
carl: My ideal world would not be one in which I or Kettman or anyone else decides how we should appreciate baseball, music or anything else. I think in our current, suboptimal world, there are plenty of ways for him and for like-minded fans to avoid the statistical distractions he dislikes — apparently, according to @natesilver, by avoiding Wrigley.
natesilver: @carl: It’s really about distractions in general, and not statistical distractions. I like going to soccer games because it’s a very clean experience. At the World Cup at Maracanã in Rio, there weren’t a lot of statistics. But there also weren’t a lot of distracting PA announcements, or cheap gimmicks, or anything else.
So give me a clean experience at the ballpark — I really don’t need to see Bumgarner’s FIP on the center field scoreboard. But also give me a fast Wi-Fi connection in case I’m in “curious nerd” mode and have something I want to look up.
cwick: @natesilver’s prescription for baseball’s future: Wi-Fi in stadiums
We reached out to Steve Kettmann for a response to our conversation. It’s below:
Thanks for seeking me out and asking me to offer a response. I see far more common ground between the view I articulated in the piece and your arguments than I might have expected. For example, I loved the discussion of how knowing more about Mahler only adds to the experience of listening to the music — but I’d point out that usually you read the program (and the biography) before the lights go down and the music starts. My piece does not argue against having statistical analysis in mind during the watching of a game; it argues against not even bothering to watch the game, because the action on the field is considered irrelevant. Benm states directly he finds “watching baseball tedious in part because I feel like nothing I’m seeing in front of me matters.” Wow! We can all enjoy baseball, even with differing perspectives on how best to understand it, but anyone who says they find watching events on the field to be “tedious” to me does not really love baseball, they love the playground of numbers the game provides.
Are the players just numbers to you? That’s seriously what you’d like people to believe? I don’t buy it. My reference to Jonathan Fader and the mental side of baseball is characterized as anti-intellectual, because I talk about the importance of “not thinking.” It does not bother me if the esteemed crew responding to my op-ed has other areas of interest than the question of how baseball players do what they do, but in fact the mental side of performance is the big uncovered story in baseball and in sports. More and more teams, in all sports, are hiring mental-strength coaches and counselors. Ignore all of this, if you like, but others might find it thought-provoking.
Nate Silver accuses me of being “incurious about the world” for daring to question how much is too much. I was an Oakland A’s beat writer for the San Francisco Chronicle from 1994 to 1998 and had a front-row seat to watch the unfolding relationship between the young Billy Beane and his mentor, Sandy Alderson, who introduced Beane to Bill James and Eric Walker and advanced statistical analysis. I’ve spent four years researching that period for my book “Baseball Maverick,” published this week by Grove Atlantic.
My argument in the Times is in part based on a distillation of what Alderson himself would say. Alderson believes that the human side matters; that you learn from close study of advanced statistics, but you also tune into human elements. Talking today with Jonah Keri for his podcast, Alderson observed “most analysts believe the intangibles are subsumed in the numbers. But I’m not sure that’s the case.”
Look, we’re all media savvy here and understand writers write articles, editors write headlines. I think, as I observed, that the application of advanced statistical analysis to baseball is a boon to the game, and is providing tremendous energy that is shaping the way the game is played — and the way the game is understood. Where I part ways with some of you is simply in whether it’s worth watching the games themselves: I was lucky enough to spend a summer at the New Yorker working with the great Roger Angell, and I will always be a fan of his close observation of the details of a baseball game. I’m saying: Can’t we have it all? Love stats, but don’t forget also to love the quirky little details of a baseball game as it develops in real time.
Editor’s Note: For the record, all of us (except maybe benm) enjoy watching baseball.