Moneyball Worked

Moneyball worked, even though the A’s never won the World Series. And no, not just in the sense that it created the best wins for value of players that we had ever seen. It wasn’t just that it allowed a small market Oakland A’s team to compete with he big market teams. In 2002, it literally made the Oakland A’s the best team in baseball, in the way that the best team in baseball reveals itself. No analytics for making the best baseball team can account for the brevity of playoff baseball.

As noted in the book by Michael Lewis and the 2011 movie, the concept revolved around getting the best value for money spent on players. Doing so involved figuring out how to “buy runs.” That is, buying players who get on base. Because baseball has so many aspects to what makes a player valuable, attributes that are not necessarily as important to actually winning games sometimes become overvalued. A player may be lauded for being able to steal bases and for being a great fielder, but might not be as instrumental in actually winning games as another player who simply gets on base more often.

Using sabermetrics, the Oakland A’s starting in 2002 were able to acquire lesser valued players who had qualities that were more likely to win games. Traditional baseball conventional wisdom valued runs batted in (RBI) and batting average. The A’s were then able to buy players with lower RBI and batting average, but who had higher on base percentage (OBP), for cheaper prices. They were able to capitalize on the misdiagnosis of what made a baseball team successful.

This worked incredibly for the A’s in the 2002 Major League Baseball season. They tied the New York Yankees for the most wins in the regular season. The difference was that the A’s had $44 million in salary, while the Yankees had $125 million. But the A’s fell in the American League Division Series (ALDS), three games to two. Their strategy was lauded as an innovative next step that improved their prospects. It was soon emulated across Major League Baseball (which of course nullified the A’s advantage, and thus put them at the same financial disadvantage). But it was characterized as “falling short,” because it did not result in the A’s wining the 2002 World Series.

But the A’s not winning the World Series does not in any way represent a failure in sabermetrics making them the best baseball team in the league. The accurate statistical analyses of sabermetrics are meant for the course of a large amount of games. And any attribution of a good baseball team or player is describing them over a large amount of games. The building blocks for effective offense in baseball is getting hits, getting on base. But even the best batters of all time more often than not do not get on base. If a player gets a hit 30% of the time for his career, he is in the Hall of Fame. That means the best offensive player can easily not get on base at any given at bat.

The same goes for the best pitchers. An incredible pitcher who has an earned run average (ERA) of 2 and an ERA of 1.5 in the playoffs for his career could easily give up 4 runs in 6 innings on a given day. When you have a team of players who can get on base, hit the ball, have great pitchers, and make defensive plays, all the aspects of what makes a team great, these qualities become revealed over the course of a long season, rather than in one game or in a few games. This is simply the nature of baseball as a sport. It makes it unique from every other sport there is. The best teams in baseball can go through certain amounts of games with mediocre or poor results without it being an indication of them not actually being that good.

Because baseball is a combination of individual statistics coming together, there is an element of randomness to it. If nine players from two teams all have a 25% chance to get on base, that can result in quite different outcomes. For one team, it could mean that their hits are spread out in a way that results in zero runs. For the other, it could result in ten runs. But over a period of many games, the averages manifest themselves correctly.

For that reason, the A’s losing the ALDS three games to two is not an indication of any improvement needed in moneyball. There is nothing about sabermetrics that can be perfected to account for short amount of games. Because what makes a good player or team in baseball intrinsically manifests itself over a long period of time. The regular season is honestly the best determinate of the best team in baseball. Playoff baseball does not offer enough games to accurately show if a team is better. But of course having no playoffs would make the MLB season much less interesting. Many more teams would be long out of the running and interest would decrease. Playoffs are fine.

The A’s use of sabermetrics, or moneyball, during the 2002 season made them the best team in baseball. There was no last hurdle that it failed to clear. Playoff baseball is exciting and I would never get rid of it, but it simply is not indicative of who is the best team. As far as baseball can be analyzed, moneyball wasn’t simply a great idea that came up just short. It was a great idea that succeeded as much as is possible for baseball analytics to succeed. It is widely used in MLB now. But it hasn’t been “perfected.” It succeeded exactly as intended from its first season in use, for the A’s in 2002.