Monday, December 09, 2013
Post-game WPA
WPA is a wonderful and perfect metric whose implementation is exactly as designed: to determine the change in win expectancy following an event. It is attached at the EVENT level.
As an additional layer, we can associate an event to one or more participants, but the actual "division" of that change in win expectancy is not so clearly determined. That we have been able to identify the player(s) involved isn't enough for us to determine the causative agents of that event. In a nutshell, we may know who is involved, but we don't necessarily know who is responsible and to the extent they are responsible. And from one perspective, we might not even CARE about responsibility, much like we don't care about the responsibility of picking lottery numbers, simply who is involved, who is actually holding that winning ticket.
All of this is determined in real-time, on the idea that the future will hold "average" results. So, our decisions (and this is not limited to baseball, but to ANYTHING in life) has as the context that we should only expect an "average" future.
But, what about after-the-fact? It's all well and good to say that the guy who opened the game with the leadoff homer really set his team on a path to winning. But what if the game ended 10-0? Do we necessarily want to associate a greater "number" to one run over another? After all, we may not know how the game would unfold, but we do know that there are 27 outs. Once all that has been logged in, can't we then retroactively establish the actual impact at the GAME level?
Remember, WPA is an EVENT-level metric. That's why it's perfect, because it does what it sets out to do. But, someone may come along and is looking at a GAME-level perspective. Clearly, WPA won't help them, just as much as a Philips screwdriver is useless to you if you are holding a handful of nails. That doesn't invalidate the existence of the screwdriver. It does invalidate you and your logic if you persist in using that screwdriver.
Studes came up with something fascinating several years back, and that is a Leverage Index that is determined after the game, and applied equally to all events in that game. It's not clear that this is a great metric, or even a useful one. But, it is rather fascinating, and it's worth further research.
Based on the data, I have estimated a standard multiplier for calculating the impact of a hit or out in a game with a particular victory margin. It's still under development, but I don't think the final version will look much different than this:
Margin Impact 1 1.38 2 1.13 3 0.97 4 0.86 5 0.76 6 0.66 7 0.63 8 0.57 9 0.51 10 0.47
Recent comments
Older comments
Page 1 of 151 pages 1 2 3 > Last ›Complete Archive – By Category
Complete Archive – By Date
FORUM TOPICS
Jul 12 15:22 MarcelsApr 16 14:31 Pitch Count Estimators
Mar 12 16:30 Appendix to THE BOOK - THE GORY DETAILS
Jan 29 09:41 NFL Overtime Idea
Jan 22 14:48 Weighting Years for NFL Player Projections
Jan 21 09:18 positional runs in pythagenpat
Oct 20 15:57 DRS: FG vs. BB-Ref
Apr 12 09:43 What if baseball was like survivor? You are eliminated ...
Nov 24 09:57 Win Attribution to offense, pitching, and fielding at the game level (prototype method)
Jul 13 10:20 How to watch great past games without spoilers