Random Baseball Statistical Analysis

A post of mine a while ago laid out my philosophical problem with the statistic OPS. I did a little fiddling around with some data using the preferred stats that I invented, and I think the results are interesting. I took the 400 players with the highest OPS+ numbers through history and computed their Not-Out Percentages, Specialness Ratios, and Total Summary Averages. Then I ranked these 400 players by all the stats I had, and did some correlations of their rankings. The correlation between Total Summary Average rank and OPS+ rank is around 0.64: a genuine correlation, but a fairly rough one. However, OPS+ is adjusted and all; the correlation between rank in OPS, unvarnished, and rank in TSA is 0.89, quite good. If I exclude stolen bases and caught stealings from TSA, since they are not included in OPS, the correlation goes to 0.92; I'm inclined to think that including basestealing makes the stat more valuable, though obviously since it's somewhat broader it conflates more different things. In any event, it doesn't make a whole lot of difference. Rank in Specialness Ratio and rank in TSA have a correlation of 0.62, while rank in Not Out Percentage correlates at an 0.55 level with TSA rank. Specialness Ratio and Not Out Percentage have a correlation of -0.25.

Conclusions: TSA and OPS match up fairly well, though not perfectly. This makes sense: they are both intuitively related to broad-based measures of baseball offensive skills. Both of the components of TSA have only a rough correlation to TSA as a whole, which is explained by the fact that they are inversely correlated with each other. That, too, makes sense: there intuitively ought to be a conflict between not getting out very much, stuff like drawing walks, having a high average, and not getting caught on base, and doing a lot each time you don't get out: hitting home runs and stealing bases. If you don't exclude baserunning from TSA, the people who like TSA more than OPS are, let's say, Rickey Henderson, the speedsters; guys like Wade Boggs, who practically never stole a base in his life, are penalized under TSA relative to OPS. If you do exclude baserunning, the players who like TSA more than OPS are power hitters: guys like Reggie Jackson, Matt Stairs, Nick Swisher. Guys who like OPS more than TSA are more pure hitters: the player with the greatest negative differential between the two in this direction is Tony Gwynn; others like Clemente and Jeter are on that list. Babe Ruth has the highest TSA overall, at around .750. Most of the guys with a Specialness Ratio above, say, Willie Mays are really, really recent power hitters, the steroid guys plus Ryan Howard. Ted Williams has the highest Not Out Percentage.

Note that, since I'm not a total baseball statistical wonk, I don't know how to adjust for league environment for my TSA stat.