the sabermetric party line
the chorus is growing. soon the sabermetric party line will be accepted by the masses.

for an example of the sabermetric party line, click here. i found the link on baseball musings.

did you reed the article? that used to be me. i accepted and espoused all the principles of sabermetrics. i thought i was smarter than general managers. i thought i was open-minded.

hopefully the writer of the article really is open-minded. it doesn't much sound like it though:

I refuse to pick on such low hanging fruit but if you subscribe to and understand sabermetric principals to any degree whatsoever, go check out the debate I had yesterday over there to gain a new appreciation of just how far ahead of the average fan you are in your understanding of the game.

if you reed the article you realize that its main flaw is that the phrase "sabermetric principles" is used as a synonym for "truth":

And if you have a deeper grasp of Sabermetric principles than I do (countless baseball fans do) and I stray somewhere, please point that out to me. I don’t ever want to stop learning.

while it's a nice sentiment, the implicit assumption in this line is that the only way to better understand baseball is to learn more sabermetrics. i was there once. i got to the point where i fully understood sabermetrics. then i realized it was horribly flawed.

- Extensive research has shown that a team’s win-loss record correlates quite closely to its differential between runs scored and runs given up.
- Extensive research has also shown that certain offensive events can accurately predict the amount of runs a team has scored and conversely, the prevention of such offensive events can accurately predict how effectively a team prevents runs.
- Given this, you can apply these same formulas to individuals to predict how well they, individually, create runs.
- Because the run prevention side is truly a team effort, a combination if pitching and defense, boiling it down to individuals is a bit more complex.

point 3 does not follow from points 1 and 2, but i'm being nitpicky. i basically agree with all these statements. most sabermetricians then go on to say that since you can measure a player's runs created, and you can measure a team's run/win correlation, you can thereby measure a player's wins created. it's an attractive idea, but it's not accurate. we'd like to think it's logical, but it isn't.

the problem is all runs are not created equal. runs scored in close games produce more wins than runs scored in blowouts. sabermetric party line says it all evens out. it doesn't all even out. there are certain skills that correlate to success in close games. the most important are contact hitting, defense, and baserunning.

power hitting tends to create the most runs. but power hitting is trumped by power pitching. this is because most power hitters are susceptible to the strikeout.

contact hitters are not as susceptible to the strikeout, and therefore suffer less of a drop against power pitchers.

by power pitchers we mean strikeout pitchers. now there is a correlation between strikeout pitchers and close games. the best pitchers have high strikeout rates. the best pitchers keep the runs down. therefore, it follows that contact hitters are more successful in close games. more success in close games means more wins.

check it out sometime. look at a team's record in close games. you'll find that teams that do well in this area tend to have contact hitting, pitching, and defense. sabermetricians view this stat as "random".

defense and baserunning are valuable in close games because they maintain the same run value while being in a lower-run environment. so they are more likely to produce wins. defense never takes a day off. power hitting does.

yes a home run is more valuable in a close game than an out. i'm talking about a home run by the offense versus an out by the defense. but home runs are way less common in close games. outs are not.

sabermetricians have tried to simplify things by making assumptions. some of those assumptions are wrong. i hope this article has helped to make that clear. to restate: you cannot look just at runs created when you compare players. you must take into account the *type* of player.

Just equipping yourself with that much knowledge gets you a long way to understanding the game. It’s just so damn logical that I cannot believe how venomous some become when confronted with it. Now I am no mathematician and consider myself more of a Sabermetrics salesman than any sort of pioneer but this all makes sense to me and I can grasp it. You won’t find me blazing any new trails in the field but you can bet I will be analyzing the game itself through the application of what I have learned in the last couple of years.

it's funny, i *am* a mathematician.
- posted by julien @ 5:39 AM