One of my pet peeves, when examining offensive performance metrics or debating player comparisons, is the lack of accounting for when a hitter makes an out that moves runners versus an out that does not. The so called productive versus unproductive out assessment. Recently, after making a comment about Espinosa's very poor offensive performance this season, which no one challenges, I also mentioned his performance in previous years which included a high strikeout rate. Further, I threw in a comment about Adam Dunn in which I stated that the Nationals were better off without him, for a similar reason that, in my opinion, even a career average season by Espy is not good. That reason is a very high strikeout rate which a relatively high home run rate and/or walk rate may not compensate for. Well, mention of Dunn in an unattractive manner brought on the wrath of RobBob, a strong defender of Dunn previously. That got me motivated again to think about developing an offensive metric that takes into account both the value of hits, and weighs extra base hits more, but also factors in productive outs, which move runners. Here is a rather half-baked formula that I present for comment and revision, with the further hope that any new metric can be calculated with easily available statistics.
Offensive Production Statistic = A + B + C where
A = (total bases) / (# of at bats) = traditional slugging pct
B = (total bases advanced by runners due to at bats of individual) / (total # runners on base summed over all ABs)
C = (runs batted in) / (total # runners on base summed over all ABs)Explanation and examples
Term A: the slugging pct., is independent of runners on base, so it gives ample credit to a player on a weak offensive team
Term B : a measure of how productive the batter is with runners on base, including via walks, HPB, hits, errors, and outs. The denominator normalizes for the number of opportunities to advance runners.
Term C: extra credit for advancing runners to score, and also normalizes for the RBI opportunities.
- A player walks with bases loaded, so for this AB the contribution to each term in the formula is:
A = 1 / 1 = 1
B = 3 / 3 = 1
C = 1 /1 = 1 so OPS contrib. for this AB is 1 + 1 + 1 = 3
- A grand slam home run, gives
A = 4 / 1 = 4
B = 6 / 3 = 2
C = 4 / 3 = 1.33 so OPS is 7.33
- A strikeout with the bases loaded (along with a DP, the ultimate rally killer)
A = 0 / 1 = 0
B = 0 / 3 = 0
C = 0 / 1 = 0 so OPS for this AB is 0. Note that the denominator of term B (3) will contribute to raising the denominator when the formula is used for a sample of ABs, thereby lowering the term B contribution to OPS and penalizing the player for freezing 3 base runners via his strikeout.
I look forward to comments and proposed revisions to the OPS formulation. It has at least minor flaws, which may not be correctable with easily available player statistics. The hope is that the final version of OPS will improve our knowledge of player offensive ability beyond what is available now, which does not add in productive outs or hitting with runners on base to the slugging percentage.