Pick Your Pop Culture

So, I've like written about music for 25 years, and like I've got a lot to say and not enough people to pay me for it, and like I like to write about TV, and books, and movies, and stuff like that.

Wednesday, June 01, 2005

The Numbers Game

The St. Louis Post-Dispatch headline in the sports section read, “The Most Indispensable Player on the Cardinals.” The story made the claim that Jason Isringhausen is the one guy the team can’t live without. Move over Albert Pujols, the second greatest player in all of baseball over the last three years by any measure you’d care to name. Nope, the daily paper thinks the most important player is a guy who pitches in less than 70 innings per season. That’s 70 out of a minimum of 1438 innings for the whole team.

How did this happen? How does conventional wisdom overtake reality again and again? It’s not just baseball, you know. Check out how many people still believe that Iraq had weapons of mass destruction. Evidence is not as important as one would think it might be. I mean, I’ve lived my whole life ready to change my opinions of something if proof comes along that I was wrong before. But, apparently, this is not common.

Baseball leaves behind the clearest trail of proof that has ever been assembled in human endeavor. You can tally up the numbers, and see just what each player has contributed to his team’s success. Everything is a zero sum game. For every out recorded by a pitcher, there is an unsuccessful at bat for a hitter. And vice versa. And, outs are the most precious commodity in any baseball game. Each team has 27 and only 27 of them to work with (unless there are extra innings, and then each team gets an equal number again and again until one team wins). So, anything that prevents outs should help one team while hurting the other.

And yet, for some reason, the vast majority of baseball fans – led by the astonishingly simple-minded sportswriters and broadcasters of the world – cowtow to statistics which don’t reflect this simple rule. There is no less useful statistic in all of creation than the number of saves tallied by a pitcher during a season. It tells us absolutely nothing other than the fact that this pitcher makes more money than all his peers in the bullpen, and he makes that money because his manager puts him there.

I wanted to read “The Numbers Game,” written by Alan Schwarz, when it came out last year, but I wound up waiting until it recently hit the paperback market. Schwarz has done an excellent job of researching the history of baseball statistics and he’s uncovered some fascinating stories. I was impressed to find there were at least a few people in the 19th century who understood that batting average, and runs batted in, were misleading numbers. And, yet, ask a modern-day baseball fan who are the best hitters in the game, and these will be the two stats they use to tell you.

Saves weren’t invented until 1969, and prior to that, pitchers were never employed exclusively to pitch the 9th inning. In fact, before the mid-1980s, when Tony LaRussa hit upon the bright idea of putting Dennis Eckersley into the role of what we now call a closer, the best relief pitchers were brought in when the game was on the line, which may or may not have been the 9th, and which may or may not have been for only one inning.

Schwarz looks at statistics from a wide variety of angles, but it’s clear he’s on the side of the mathematicians who have expanded the science of baseball study over the years. Not that there’s anything wrong with people who have come up in the game without knowing the exact percentage of runs which can be expected to score with runners on 1st and 2rd and one out. It’s just a little more fun to have that tidbit in your back pocket while you yell at the manager for deciding to bunt in a situation which is already likely to lead to good results. Trading an out for a miniscule increase in the likelihood of scoring a run is just silly.

So, yeah, I was talking about saves. The rule states that a pitcher who gets the last three outs of a winning game, as long as he enters the game with his team ahead by three or fewer runs, or if he gets fewer than three outs yet faces a batter with the tying run in the on-deck circle, is credited with a save. It doesn’t matter if the last three hitters are the bottom three in the order for the worst hitting team in baseball, or if the three previous hitters, retired by another pitcher, were the three most dangerous hitters in the game. Saves only measure who gets the last out, whether or not that out was the most difficult out.

Watch Jason Isringhausen pitch sometime, and prepare to chew on your knuckles. I mean, the guy is a decent relief pitcher, but he has a nasty tendency to put people on base in tight situations, unless he’s facing the real bottom of the order. In that case, he’ll blow people away. He’s had a lot of saves in his career with the Cardinals because the Cardinals have won a lot of games while he’s been here, not the other way around. I’m glad he’s on the team, though, frankly, I wouldn’t pay him the huge salary he commands if I were in charge. But, the most indispensable guy on the Cardinals, a team that has three of the ten or so best hitters in the game today – that’s Pujols, Edmonds, and Rolen, if you don’t already know? That’s just crazy talk. It’s the equivalent of believing false evidence without bothering to examine the correct evidence right under your nose.

“The Numbers Game” is a blast to read. It’s like watching the race to discover the atomic bomb or something, as formulas come into existence which destroy conventional wisdom. I find baseball to be infinitely fascinating, of course, but it’s even more so when you realize just how much there is to still be learned about it. I do worry about the concluding chapter of the book, though. Schwarz seems convinced that a new computer software and hardware package called Tendu will revolutionize baseball statistics. By next year, if all goes well, we should have access to every data point from every major league game, which, I think, will confuse the issue.

Baseball becomes clearest when the most data is crunched. Until you have hundreds or even thousands of examples of situations, randomness will infect what you see. Any twenty or thirty at-bats can result in any sort of situation. There’s little that annoys me so much as hearing announcers prattle on about how Roger Cedeno is 6 for 9 against a certain pitcher, as if that makes it very likely he’ll get another hit. The fact is, Roger Cedeno isn’t a very good hitter at all, but he’s still a major league ball player, and thus is capable of getting six hits out of any random nine at bats in his career.

Tendu will oversaturate us with data that is essentially random noise. Want to know how many times Albert Pujols pulls the ball with an 0-2 count and the temperature 75 degrees? Well, we’ll know, but so what? What will that mean to our understanding of the game? I fear micromanaging by insane statistics in ways that will make Tony LaRussa look like Homer Simpson.

Of course, ultimately, we’ll be able to use this data in ways much more interesting. The more information we have, the more likely we’ll be able to know what’s going on over a larger number of events. And, the beauty of baseball has always been that for all the joy of each individual game, it’s the lifetime of experience with the sport which makes it something to love.


Post a Comment

<< Home