Baseball and process control

We are well into the baseball playoffs and are thus running out of time for baseball posts. To the rescue is a story from Marketplace that relates to baseball and, in a way, statistical process control (The Yankees’ monopoly in baseball, Oct 14).

The gist of the piece is that the Yankees of have dominated professional baseball essentially from its inception. The easy answer for this is they have spent a lot of money. But there are also things in the way that baseball runs that make spending on tons of talent count:

But there’s more to it. Just about everything in baseball’s structure militates against parity. Start with the long baseball season. In the same way an opinion poll sampling 100 subjects will be a more precise than a poll sampling 10 subjects, baseball’s loooooooong season lends itself to an accurate reflection of talent. If two teams play one game, anything can happen. But if they play a good many games, the better team will win the majority of the time.

Then consider the playoffs. Only the eight best teams in baseball make it to the postseason — fewest of any major sports league. And, unlike the Super Bowl, the World Series is a best-of seven game, again, a large enough sample size so that best team should win. The Yankees may have the most baseball titles, because they buy the best players. But the imbalance is allowed the flourish because of baseball itself.

It is worth pointing out that divisional play and multiple rounds of playoffs didn’t come to baseball until 1969. All of those Yankee teams with Ruth or DiMaggio or Mantle went from the regular season straight to the World Series. This only re-emphasizes the point of the article. A system that stresses a big sample and exposes the teams that survive that to few rounds of short series will favor the deepest and hence richest team.

So what is the link to process control? Well if you want to properly evaluate a process, you need a sample. The outcome of individual items is random and subject to variation. The Yankees can drop a game in July to the pre-Showalter Orioles. That doesn’t mean that the Orioles are better team. It just means that the Orioles had a better day. Process control is similar. Just because one item is out of specification, that doesn’t mean that something is wrong with the process. Similarly, just because one item is on the money doesn’t mean that everything is OK.

Samples allow for that variation to average out and that let’s you see whether the process is in control or not. And they also let you see who should go to the post season and who should get yet another high draft pick.