it’s called WPA (Win Probability Added).

It’s a fun little tool that every year people try to trot in MVP discussions as being meaningful even though it’s obviously a junk stat. And by junk stat I mean junk stat – it’s the only stat in baseball that actually punishes a player for better performance. An extreme example, but the most obvious, is to look at a player that hits a solo home run in the first inning and then another solo home run in the ninth inning with a 1-0 lead. Even though that hitter is single handedly responsible for all the scoring of the game, his WPA is only around +13.4% (depending on the outs and home team).

If that player had the foresight to only strike out in that first at bat and then hit a home run with two outs in the bottom of the ninth in a tie game, he would have had a WPA of +43.7. WPA also over-values pitchers that play behind bad offenses and offenses that have to overcome poor pitching performances.

Although, I will note that there’s a variation on the win probability that looks at particular in-inning situations and creates a run expectancy matrix. Baseball Prospectus uses this data to create a stat called Adjusted Runs Prevented that looks at the game stat when a reliever enters and compares the number of runs that would score on average to how many runs actually scored. So if a pitcher comes in with no outs and a runner on third the run expectancy of that game situation is 1.452 (chart here). If said pitcher only allows the one inherited run to score, they’ll be credited for +.452 runs prevented. If instead a pitcher came into the same situation, but with two outs, the run expectancy would be .334 and allowing the runner to score would would give the pitcher a -.666 runs prevented. Since this looks only at in-game state and not the score, I think it’s the ideal measurement for reliever effectiveness, but I should note that if there’s any aspect of the game that makes the most sense to be measured by WPA, it’s reliever performance. WPA can be a rather blunt tool at times, but it does a better job than ERA when you’re dealing with inherited runners.

I’m a Diamondbacks fan, but I was watching the Colorado feed of the game and I heard your announcers talk about how Hawpe was getting most of his RBI in those more “important” late inning situations, but personally, I find that line of reason to be extremely faulty. Sure, the ninth inning game-winning home run is really flashy, but if I had to choose between a hitter that does his best hitting in the first five innings or the last four, I’d choose the former every time.

It’s absolutely huge for a team to score early. It takes the pressure off your starting pitching, increases the chances of a short outing from the opposing starter and makes it easier to score in the late game since modern baseball managers use their lesser relievers when their trailing than they would if it was a save situation. Of course, these were the same announcers that later said the main failing in Conor Jackson’s offensive game is his inability to hit with two outs and RISP. I guess the concept of small sample sizes must elude them.