How Good of an Offensive Metric is OPSBI?
Posted: June 16, 2011 Filed under: Research 5 Comments »Update: Just so everyone is clear: I do not agree with the concept of adding RBIs to OPS. If you are looking for a “quick stat,” I recommend looking at OPS all by itself; it does a fine job. The purpose of this was to simply look at OPSBI and see how well it correlates with a more advanced statistic.
Last night, Jim Bowden put up a post on ESPN.com, wherein he created a new stat called OPSBI. Basically, you take the OPS of a player, multiply it by 1000 and add RBI to it. I can honestly say I’ve never thought of doing this, but let’s see how it stacks up to an advanced hitting metric like Fangraphs batting runs. I took only players with greater than 1500 plate appearances in the past three years. This provides a 133 player sample. Here ya go:
(Click picture to make it larger)
R^2 of 0.84 with batting runs? Not as terrible as you may think.
Yeah but if you just use .OPS it’s like .98. Why muddy the waters?
Because the more letters in a metrics name, the better it is? I am in no way agreeing with the concept of adding RBI to OPS, I just thought it was interesting to show that it actually does correlate pretty well to offensive output.
Well, yeah, of course it does. It’s based 90% on OPS. It’s like mixing beer and water, with a 90/10 beer/water ratio, then saying “see, this gets you drunk 90% as effectively as beer!”
[...] Curt Schilling’s stat PDF with FIRE I recently did this with Jim Bowden’s OPSBI and it was too much fun to not do this [...]
[...] Anyway, it doesn’t matter. This was simply one vote in one season that will be forgotten easily. So instead of continuing a discussion about this particular vote, I have decided to create a framework by which all future MVP voting should follow. I call it the BARBI system and I would be remiss if I did not thank Jim Bowden for his inspiration. [...]