Step Aside Bill James (Just Kidding)
July 23, 2004
This topic began as a blog entry gone awry which got too bloated for the home page, so I moved it. Here it is in all its awkwardness:
While the Angels and A's struggle to keep up with the Rangers, many of us this year have asked how this happened. I thought I'd try to find an answer.
Bill James, stat guru extraordinaire, calculates projected win percentage with a relatively simple formula incorporating runs scored and runs allowed. It doesn't do much for predicting future events, but it does a good job explaining past history. Basically, he says, "hey dummies, if you score more runs than you allow, you give yourself a better chance of winning!"
While we all know that, we always want to weight pitching much greater than hitting. If you don't score, though, you can't win. Perhaps that explains why Texas is successful this year without any aces on the staff. Using an even simpler method than James' formula, let's look at the 2004 Rangers:
In 2004, the Cubs (51-44) are struggling more than expected even though they are giving up less than 4 runs/game (3.95). The Rangers (54-39), on the other hand, give up almost 5 runs/game (4.99), yet remain 15 games above .500. Common sense tells you why - the Cubs score 4.66 runs/game and the Rangers score 5.65 runs/game.
The Cubs actually have a larger spread between runs scored (RS) and runs allowed (RA), but are 3 games behind the Rangers. I suspect that the Rangers are either more consistently above the average with their bats (better to score 6 runs 3 days in a row than 18, 0, 0) or with their pitching. Let's see:
50.5% of the time, Texas scores more than their average RA (4.99).
64.5% of the time, Texas gives up less than their average RS (5.65).
58.9% of the time, Chicago scores more than their average RA (3.95)
68.4% of the time, Chicago gives up less than their average RS (4.66)
My suspicions were wrong, as the Cubs have actually surpassed their averages more than the Rangers. On Chicago's side, it appears that both sides are doing their job, with the offense faring better than the pitching. Of course since the percentage is based on the opposing statistic, a slight slump in pitching (bumping the RA to over 4 runs/game) would actually hurt the scoring percentage. I did it this way because comparing RS to RS accomplishes nothing. We need to know how consistent run scoring helps, given a constant (RA), and vice versa.
What can we do with these percentages now?
I don't know, I'm no statistician. I'm just a guy with a website.
Assuming that offense and defense (including pitching) are equally important, let's average the percentages to determine the projected winning percentage for each team.
Rangers: 57.5%, or .575
Cubs: 63.7%, or .637
Does this make mathematical sense? I'm guessing no, but it sure looks good, doesn't it? It's safe to say that Bill James' methods are better tested and, well, correct compared to mine. Nevertheless, using these projected win percentages, the Rangers' and Cubs' records should be:
Rangers: 53 - 38
Cubs: 60 - 35
Hmmm. Seems a bit high for the Cubs, and I suspect that my calculations, if used on every team, might yield more wins than losses across the league, thereby making my work faulty on several levels. Not wanting to test that, I'll just say that, based on my figures, warts and all, the Rangers are for real and are right about where they should be, and it's time to put some money on the Cubs because they are due to catch up to the percentages.