In the final section of "OPS: Be Gone! - Part 2" Tom Tango demonstrates how players with the same OBA, same SLG, and widely differing batting averages have varying run values according to linear weights. Disturbingly, the players with the lower batting averages seem to have the higher offensive value, according to the set of linear weights that Tango is currently using (or was using, when that piece was written). As part of the back-and-forth in the comments on Tango's blog, he challenges me to explain this, presumably in such a way as to defend my contention that absolute average (ABSO: the Average of Batting, Slugging, and On-base averages) is a better metric than OPS. This post is an attempt to do so.

Linear weights are calculated for a particular year, or for a particular range of years. They measure how much each offensive event contributes to (or detracts from) the scoring of one or more runs, over that particular sample. For samples in which home runs are common, the value of the walk will be closer to the value of a single than for samples in which home runs are less common. Walks do as good a job as hits in getting runners on base, but do not do as good a job in moving over existing runners. The fact that the home run clears the bases, and drives in runners regardless of which base they happened to occupy, emphasizes getting on base and de-emphasizes moving runners over. In a homer-friendly environment, "get 'em on, get 'em over, get 'em in" becomes simply "get 'em on, get 'em in."

In such an environment, not only does a walk become almost as good a hit, but the sacrifice and the stolen base become less valuable, and consequently more rare. Why risk making an out trying to get a runner to second or third, when the guy up next is almost as likely to drive him in from first? Why hit behind the runner when you can hit it out of the park? The fundamentals become less fundamental in a homer-friendly environment.

It was not always so. When Crosley Field (then called Redland Field) was built in 1912, it was 360' down each line, and 420' to dead center. Because of the dead ball being used at the time, it was almost a decade before a player hit a home run over the fence there. (Pat Duncan was the first major-leaguer to do it, in 1921.) Edd Roush, the Reds' best player during several of those years, accumulated 186 sacrifice bunts as a Red. Roush considered hitting to be an extension of bunting, and never took more than a half-swing at the ball, letting the weight of his four-pound bat do all the work. He stole 199 bases as a Red, and although we don't know exactly how many times he was caught stealing, it is doubtful that he was successful much more than 50% of the time.

In the modern environment, such numbers would be sneered at as highly counterproductive, and perhaps rightly so. But it made sense for the dead-ballers to play for one run at a time, because "small ball" was all they had. In the modern environment, conversely, it makes sense to simply avoid making outs and wait for the three-run homer. But both these environments are arguably "outliers" when considered against the entirety of baseball history. For many years in between, there was a place for both small ball and the long ball. I wager that if you calculate the linear weights for those years in between, you will find that among players with the same OBA and SLG, but widely differing batting averages, the players with the higher batting averages will have more value.

If this turns out to be so, I would argue that the linear-weight results for the current environment indicate that the current, homer-friendly game is seriously out of balance and in need of correction. College basketball has recognized that the old three-point shot was too easy, and is moving the line back a foot this year. (Incidentally, I predict a negative impact on the linear-weight value of the 20-foot jump shot.) Maybe it's time for MLB to recognize that home runs are too common in the current environment, and make a similar adjustment.