Revisiting Moneyball
djpardis.medium.com
djpardis.medium.com
He thought that analytics was changing the probabilities of discrete events by single digits. Essentially, nobody was doing anything wrong, there were just optimizations that were/are available.
Remember that the book/movie is about the A’s, who were eliminated in the first round of the baseball playoffs.
The one that pops into mind offhand was putting the "cleanup hitter" as leadoff, even though at the time "leadoff hitter" was a very specific physical archetype as was the "cleanup hitter". Yet the latter was often the best hitter on the team, and by putting them top of the batting order they'd get more at bats over the course of a season.
That is actually huge in baseball. As an example, the player who was least likely to get a hit last year did it in 19.6% of their at bats (this typically isn't represented as a percentage and would instead be listed as .196 batting average), while the player who was most likely to get a hit did it 33.2% of the time, meanwhile the league average was 24.3%. That means "changing the probabilities of discrete events by single digits" is what separates the average player from the outliers at both the top and bottom of the talent pool.
On the face of it, NFL’s playoff of a single game deciding forward progress seems more high variance, rather than a best of x series.
Football also seems more high variance just due to the explosive, physical nature of the game.
I wonder what the stats say about lower seeds winning the tournament for MLB vs NFL playoffs.
Meanwhile the best regular season team of my lifetime, the '01 Seattle Mariners, still only win maybe 75% to 80% of the time vs. the historically terrible 2024 White Sox.
Now put the '85 Bears against a historically inept team instead of one of the other best teams in the league that year. I'm not sure the '76 Buccaneers win 1 game out of 100.
It's just not possible to physically dominate a game in baseball the way it is in football.
Why it’s so much harder to predict winners in hockey than basketball
If you can have a competitive team filling 80% of the seats as your competitors at 50% the payroll..
Back then, it was relatively easy to just look around and prod relatively available data and find new angles if you went in eyes open. If you were trying to justify spending all that money on that kid who isn't working out with you, the approach lets you dig the hole you want.
One of the interesting 'post-Moneyball' stories is when old-school scouting methods came back onto the scene. People started overvaluing the new popularized statistics, and the market advantage was to combine the analytics and traditional approach in a cost-efficient manner.
Some of the smarter teams in the NFL seem to be figuring out that maybe running backs aren't completely fungible, as has been the mantra for a while.
There is no durable thing you can simply identify as your edge in metrics that you can stick to for years.
There are things that work with a very high percentage ...and there are things that are enormously satisfying and exciting to do.
If you're interested in winning you will methodically do the "correct" things.
The problem is it's just so much fun to do a firemans carry ...
In the Moneyball narrative, the non-analytical scouts are branded as "stupid" or "thick" or even "bigoted". I see them as more human and less robotic. I bet they have more fun than Billy Beane (and the book suggests they do).
But, as in business and personal relationships, finding what works best and putting it into practice tends to be more likely to succeed than simply hoping it will work--look at me, it has never been tried before!
In both, it's way more interesting to watch a bunch of less talented people with personality and style duke it out, than it is to watch perfect, best of the best execution go toe to toe.
That's why we look at a short video of an unbelievable Forward 3½ Somersault combined with a Pike dive, and we say to ourselves, "that's great", before quickly switching to look again, for the 100th time, at a less technically impressive but more emotionally charged goal scored by our favorite team 25 years ago.
I rarely watch sports nowadays, also because it is difficult to get really into a team or an individual when they are younger than you are and make millions.
I’ll agree. Was kind of a big sports fan in the 80s (nba, nfl, mlb). The characters were pretty fun. It was a little crazy but fun. Some recent documentaries/books on the Celtics I checked out reveals more things going on behind the scenes. The press had more access. The Celtics broadcaster “Johnny Most” was really not very partial to other teams, but came up with fun nicknames. One game on the radio they stopped announcing and started laughing because he lit his pants on fire with a cigarette https://en.m.wikipedia.org/wiki/Johnny_Most
I tried watching sports now, the players are so good but it lacks the fun / heart. It’s professional and big business but somehow when you don’t care it’s a lot less fun. Though I did enjoy the women’s world cut in France quite a bit.
what about grabbing a folding chair and hitting your opponent with it?
the problem is when 'entertaining' optimizes away reality.
As if there is some kind of no-arbitrage condition arbitrage, to be made between humans playing baseball.
This is nothing new though. Baseball is America's pastime in a very deep way. I collected baseball cards in the 80s like I was trading stocks in another bullshit, market fundamentalist delusion.
Do baseball fans ever discuss potentially changing the rules or game setup to mitigate this?
The actual change was lowering the height of the strike zone (to help hitters.)
> Nearly a quarter of the way into the 2023 regular season, the rule changes MLB implemented this year are continuing to have their intended effects -- namely, a quicker pace resulting in a significant reduction in overall time of games, more hits and more steal attempts.
I'm not old enough to remember a pre-moneyball world, but MLB has been mitigating the effects of the “three true outcomes” problem quite well over the last few years. The pitch clock is the best thing to happen to baseball since integration.
Occasionally an old head will grumble about it, but even the old guys tend to be pretty happy about the pitch clock.
That makes sense statistically, but I think most fans would intuitively disagree. When your team plays great all season and chokes in the playoffs, you rarely come away feeling that the regular season was their “true quality.” Typically, such events are seen as revealing that the team wasn’t as good as it seemed, or at least not when it mattered most. There’s probably truth to both perspectives.
That experience led me to have a couple thoughts:
- Many people think that Moneyball was about On Base Percentage. Really, it was about how to use analytics to figure out what the undervalued metric is to focus on.
- Moneyball is great on its own. When you try to do your own version of Moneyball in a sport you care deeply about, it's as if you find another "level" of the book. For example, this quote from the book (paraphrased): "You can't see the difference between a .275 and a .300 hitter. It's an extra hit every two weeks." You realize HOW MANY parts of sports are essentially invisible unless you take the stats and then analyze them.
- Another example of the above is the quote "it's a system only a rich team could afford but only a poor team looking for an edge would buy". Having worked at a lot of startups (4), I've often told this story when people are deciding whether to pay for some 3rd party tool that will have a big impact. Sometimes it's not worth it but sometimes it can really change the outcome.
- Going in to the whole process, I wasn't sure how well I could track stats in something as complicated as college level tournament paintball. I started with just a clipboard with some paper and a stopwatch. At the end of the day, that was really all I needed to get DEEP insights on how the game worked, game theory in different scenarios and which players were adding or taking away from our win percentage. A reminder of "you can just do things"
- One unexpected outcome: I got to the point where I could watch a game and know exactly what the win probabilities were as the game went on. It was like watching the World Series of Poker where you see the odds of each hand. This was both great b/c I new the next game theory optimal call or play to do. It was terrible when I knew the odds of winning a big game went below 5% (although we did occasionally win some of those too)
0 - https://www.nytimes.com/2009/02/15/magazine/15Battier-t.html
There's a market distortion from the rest/majority of teams incorrectly evaluating the players. Knowing that and taking advantage of it is underpaying if you believe in paying people their real worth instead of just what you can get away with paying them.