Looking Back at Moneyball - 20 years later

[MFS62]

It has been 20 years since the season we read about (and later saw) in Moneyball.

How do you feel about the increased use of Moneyball concepts (analytics) in the evaluation process since then?



BA has an article about this, but I can't get beyond the paywall to find it and post a link.



Later

Edited July 9, 2022 by Guest

[Ceetar]

Moneyball actually came out in 2004, it was _About_ the 2002 season.



I think my favorite moneyball concept is that players are bad sources about what is "good" baseball, and what happened in a game/season and why.

[MFS62]

Corrected. Thanks.

It comes across to me as the way the new generation of business focused owners can understand what they have bought.

I'm more concerned about the way numbers are being used for player development. For example, how many good contact/ spray /OBP hitters have been messed up by coaches concerned with launch angle and power? We may never know.

Later

[Ceetar]

the answer is pretty close to zero, because now we recognize high OBP as talent. If you mean bad contact spray hitters, those players weren't good and we're not messing them up so much as cutting bait for better players.

[Johnny Lunchbucket]

I think, Moneyball's legacy is bifurcated.



On the one hand it opened eyes that there was some value in applying a kind of economic analysis to the consideration of ballpayers, or sports overall and what made the book so effective and important, was that it captured a moment when that kind of thinking met with entrenched resistance, and Beane's own cognitive dissonance about his own career, and a kind of uneasy hybrid future was born.



One the other hand, its notoriety made it way too easy for people who never fully appreciated all of the above, to think of "Moneyball" as a monolithic thing that meant nerds with slide-rulers were running baseball.



Such a great book, (and a fun but flawed movie) and so complicated!

[Frayed Knot]

The number of the book's detractors [Joe Morgan, Mike + Mad Dog, others] who never actually read it would have been amusing if it weren't so annoying.

[roger_that]

What Moneyball (sabrmetrics, actually) did was quantify small things that no one considered important before because measuring them was difficult and it was unknowable how much each one affected winning, if it did at all. So we let managers, in particular, play hunches and do all sorts of wacky shit because we couldn't say for sure that it was wacky or that it was shit. But now we can--so SB attempts are way down (because we now KNOW that if someone can't steal bases at close to 75% success, you don't send him), small ball strategies are much reduced, the complete game has died, etc., which were all (plus fifty more) resisted at first because we've always played hunches and done weird shit. Now, not so much. A few holdouts remain, but very few. Moneyball popularized sabrmetrics, dumbed it down for the masses, but exploiting market inequities is winning baseball, and ignoring it is for losers.



Where I think the lightbulb went on over people's heads was that each of these market inequities is tiny, a percentage point or two of winning percentage at most, but there were so many of them that they did add up to the difference between making the playoffs and not.

[MFS62]

Considering all the publicity, do you think the overall impact on the game has been greater than we think it is, or less?



Later

[Edgy MD]

I don't think we know as much about stolen bases as we think we do. And I don't think 75% is the real break-even point. Indeed, I think that point shifts a lot over the course of a game.



I think the accidental fallacy of Moneyball was convincing people that such understandings as the break-even point in stealing, the existence of clutch hitting, the value of defense were closed matters — something I don't think Michael Lewis or Billy Beane or Bill James would have endorsed. I think we opened up new and important questions, but some things we stopped refining, and so an old set of prejudices was largely displaced by a new set.



The point was — or should have been — the ongoing rexamination of received wisdom, rather than the creation of a newer, equally closed category of received wisdom.

[Ceetar]

nobodies closed off to continued understanding except like, people that read moneyball and think "oh, well we missed OBP, but we understand now" And stopped thinking.



We're learning new things all the time, better quantifying defense, etc.

[Johnny Lunchbucket]

Right. The idea wasn't that stolen based were worthless necessarily but there was an economic value assessed to them that-- in that moment-- was out of line with their dollar value with respect to their contribution to winning.

[Edgy MD]

nobodies closed off to continued understanding except like, people that read moneyball and think "oh, well we missed OBP, but we understand now" And stopped thinking.

Those are people. And there are a lot of them.

[roger_that]

Edgy MD wrote: I don't think we know as much about stolen bases as we think we do. And I don't think 75% is the real break-even point. Indeed, I think that point shifts a lot over the course of a game.

No stat measures absolute constants through all games, but when you steal less than 75%, your gains are minuscule (or non-existent)and below 65% you are hurting your team. I wrote 75% as the point at which you're definitely helping your team. Stolen base attempts, per team, are down significantly over the past few decades, and SB percentages are up, marginally, over that time. Teams are looking for small edges, and here they've identified one thanks to sabrmetrics. Before 1980 or so, you'd get people who couldn't steal bases trying 10 or 20 times a year, and getting thrown out over half the time. Now, not so much. Probably amounts to a gain of less than a win per team per year but every edge counts.

[Ceetar]

Edgy MD wrote: nobodies closed off to continued understanding except like, people that read moneyball and think "oh, well we missed OBP, but we understand now" And stopped thinking. Those are people. And there are a lot of them.

I don't think there's any real subset of people that treat the 2004 knowledge of baseball as a pinnacle. Maybe a handful, but mostly those are people that just happened to decide they'd understood everything at the same time as moneyball. Reluctantly accepting that OBP is good. Very few of them are arguing in good faith though.

[Edgy MD]

I agree that there's going to be bad faith. But it's more that there's only so much processing power some of us have as the key data gets pushed to margins of the math.



The arguments of data can be elusive, and how it's presented helps. Others can be too easily swayed by data presented well but misrepresenting the facts.



I think we all will get stuck somewhere, but as long as players are going up there swinging at the first pitch with the tying run behind them in the ninth, I'm going to wonder how well we're learning.

[roger_that]

Edgy MD wrote: wonder how well we're learning.

That's the point, I think. Some will think it's worth it to make decisions that give a teeny tiny edge consistently, and some won't. But those teeny tiny edges add up over the course of a season.

[Frayed Knot]

Edgy MD wrote: I think the accidental fallacy of Moneyball was convincing people that such understandings as the break-even point in stealing, the existence of clutch hitting, the value of defense were closed matters — something I don't think Michael Lewis or Billy Beane or Bill James would have endorsed. I think we opened up new and important questions, but some things we stopped refining, and so an old set of prejudices was largely displaced by a new set.

And Bill James himself has publicly said so.

He and his followers upset the status quo by questioning what had always been seen as "true". In a piece (ten years ago?) titled 'The Fog of Numbers' (IIRC),

James cautioned the current wave of stats hounds not to assume that their generation has any more of a monopoly on truth than the old one did and that the

new status quo deserves just as much questioning as did the old. Maybe, he opined, clutch hitting as a skill really does exist but it's merely that the evidence

hasn't been found yet.



To take a similar argument on a different subject, ethanol was once considered the holy grail of environmentalists while nuclear power was the ultimate evil.

Since then, Al Gore admitted that the ethanol push was both a failure and a total waste of (your and my) money, while some former Greenpeace leaders have

since made a 180 and now embrace nuclear as at least a piece in the fight against global warming. These flip-flops (if you want to even call them that) don't

mean that you don't agree on the problem only that no one view necessarily has a monopoly on the answers. In this context it often bothers me when I hear

the term "settled science" being thrown around because it tends to imply that everything there is to know on the subject is already known.

[Ceetar]

Bill James is a good example of someone that went a certain distance and then just kind of stopped. The clutch hitting thing is akin to saying "well I dunno, maybe the vaccine will have crazy side effects 2 decades from now!"

[Edgy MD]

Does anybody else think that analogy works?

[roger_that]

Edgy MD wrote: Does anybody else think that analogy works?

I would need to understand it to answer that question.





James has certainly backed off "Clutch hitting doesn't exist" but it's mostly a jiu-jitsu move. Now it's more like "Certain hitters MAY be marginally more skilled in the clutch situations than others, but far more significant is the small size of samples we have available to measure it, and as that size grows larger, everyone reverts to the norm." In other words, if one guy bats .280 in clutch situations and a very similar batter hits .260, you might be able to claim the first guy is a better clutch hitter but if we had a few more decades (somehow) of their at-bats, they would meet at .270 and the streams might even cross. But since we don't have 15-decade careers to look at, no one can say you're definitely wrong to scream "CLUTCH HITTER!" at the first guy and "BOOOOO!" at the second one, so knock yourself out. But it still adds up to "Clutch hitting doesn't exist." We just can't prove it doesn't.

[Ceetar]

Edgy MD wrote: Does anybody else think that analogy works?

"there's no evidence for something but I'm still gonna insist it must be out there somewhere"

[roger_that]

There's plenty of evidence, just no definitive proof. Which is probably impossible to find--it's proving a negative which can't be done.



There are of course plenty of instances of clutch hitting, as there are instances of just-plain hitting. But show me the proof that there are certain hitters who have the ability to hit in the clutch, year after year, and certain ones who lack that ability.

[smg58]

I think it improved the knowledge base of the average baseball exec and the average fan considerably, and that's a good thing. But the games are longer and less interesting as a result. Most hitters are going to wait for a pitch they can knock out of the park, and will happily take a walk if they don't get it. They are also not afraid of striking out if their approach leads to more runs in the long haul. These things can all be justified by the stats, but I would argue that the little things and the diversity of players' styles and the strategy involved are what made baseball fun.

[Ceetar]

I think it improved the knowledge base of the average baseball exec and the average fan considerably, and that's a good thing. But the games are longer and less interesting as a result. Most hitters are going to wait for a pitch they can knock out of the park, and will happily take a walk if they don't get it. They are also not afraid of striking out if their approach leads to more runs in the long haul. These things can all be justified by the stats, but I would argue that the little things and the diversity of players' styles and the strategy involved are what made baseball fun.

But all these things are easily fixable if MLB really cared.



I find the game MORE interesting now than 18 years ago personally.



A lot of the "less interesting" is from the broadcast/marketing/advertising side, specifically the lack of diversity thing, because TV Networks want more uniformity and predictability. They don't want to have to understand nuance, etc.



But if you want more contact, shrink the strike zone and deaden the ball a little to reduce home runs (something MLB maybe is trying to do but they're remarkably bad at it). It'd be remarkably easy to adjust the strike zone once it's robotic. The same with strike outs/homers. Currently the strike zone shrinks on 0-2, but if you want to shift the Home Run/Strike Out match, you can not shrink it on 0-2, which means you're more likely to strike out, less likely to walk, etc, so you're gonna make more outs per home run, and it might not be worth it as much. It'll incentivize contact.



But I'll reiterate ,I much prefer the home runs and the strikeouts to soft groundouts to second. And I'd caution Rob "I hate baseball" Manfred on thinking that that's what's keeping people away from the game that would otherwise watch.

[Frayed Knot]

Edgy MD wrote: Does anybody else think that analogy works? "there's no evidence for something but I'm still gonna insist it must be out there somewhere"

No, he said that it Might (very different from 'must') exist to some degree or another even though current thinking says that it doesn't.

But mainly he was simply using the concept of clutch as an example of a topic where we shouldn't act as if we already know all there is to know. Just because the current status quo replaced the previous one doesn't mean that the current one shouldn't also continue to be challenged.

[Ceetar]

yes, but go ahead and challenge it with data, not lack of data. It's not like there's a nugget of data that suggests clutch hitting exists and we just haven't isolated it.



This _IS_ the same nonsense being wielded during the pandemic. "Cloth masks might not be that effective therefore I'm not gonna bother" or "people that are vaccinated still get Covid so it must be ineffective"



We have MORE data, MORE understanding. Yes, obviously continue to question, but this is NOT the same status quo, it's a more informed one. We know for example, in almost every single case, that a runner on second with 1 out is a lower scoring situation than a runner on first with no outs. In millions of situations, this holds true. we have the results of what happens. If you're going to question that, you better have some sort of subset of data or situation that you think data's not capturing, to explain why you're questioning it.

[roger_that]

The burden of proof is on the one who asserts the existence of something, not on the one who questions its existence.

[Edgy MD]

=Ceetar post_id=99108 time=1657470039 user_id=102]
yes, but go ahead and challenge it with data, not lack of data. It's not like there's a nugget of data that suggests clutch hitting exists and we just haven't isolated it.

[Frayed Knot]

=Ceetar post_id=99108 time=1657470039 user_id=102]
yes, but go ahead and challenge it with data, not lack of data.

[batmagadanleadoff]

I think clutch hitting, as a skill, is like those old tales passed down from hundreds of years ago from people who lived by candlelight because electricity didn't exist and when they wanted to have chicken dinner, had to go out and hunt down and catch their own chickens and then kill and clean them themselves by hand because there were no Kentucky Fried Chickens around. And we live our lives according to what these primitive bastards who knew practically nothing abut the world they lived in believed, because we're barbaric primitive bastards ourselves, but fool ourselves into thinking otherwise because we have air conditioning.



Put aside for a second,the total lack of objective numerical evidence for clutch hitting. If clutch hitting were a separate and distinct skill, totally independent of regular hitting, we'd have a list of hundreds, if not thousands of clutch hitters, backed up by the math, because they've been playing baseball and keeping meticulous detailed records of the sport since practically the Civil War.



But the real reason clutch hitting doesn't exist is because clutch hitting itself makes no sense. It's a ridiculous absurdity. Me, personally, I dismissed the idea of clutch hitting by the time I was in Junior High School. Because if clutch hitting existed --- if there really were batters that could truly elevate their performances in clutch situations, however clutch is defined --- then those batters would "clutch up" all of the time, not just in "clutch" situations. The 25 HR's a year hitter who could supposedly hit in the clutch at will would want to hit 40 HRs instead if 25. Or 70. The .330 clutch hitter would want to hit .400. Or .600. And so this extraordinary but mythical clutch hitting skill would be used in every at bat, not just in clutch situations.



Besides, every hit is a clutch hit. So, for example, when three straight hitters come up with two outs and nobody on base and each of those batters single, generating a run for their team, each of those singles was clutch. The first batter in that sequence also gets credit for a run scored, the last batter -- an RBI. The middle batter gets neither a run nor an RBI. But his single was just as crucial to the run producing sequence. Every single in that chain was the proverbial weak link in the chain because take away any of those singles, and the whole chain collapses -- the run never happens. The point being that it's just as important to get a hit with nobody on base as it is with a runner in scoring position. So the whole concept of clutch hitting is flawed as well.