The Wild Card Games and World Series Win Probability Added

A year ago (wow, I did not remember it had been that long) I introduced the concept of World Series Win Expectancy and Win Probability Added. The idea is simple: you assume that every post-season game is a coin flip and use that assumption to determine the probability of winning the world series from any position in the playoffs, and then within each game you treat the Win Expectancy scale from 0 to 1 as corresponding to the interval from your WSWE if you win to your WSWE if you lose. Individual events that cause a change in your game Win Expectancy, and which therefore accrue Win Probability Added (positive or negative), generate World Series Win Probability Added, and we can therefore be pretty precise about which players and events have had the most impact.

This October, just for fun (and because I have lots of stuff to be procrastinating from) I'll keep a leaderboard of World Series Win Probability Added. I'll keep track of which players have done the most so far in the post-season to help their teams win the World Series, and of which players have done the most to help them fail to win it. I'll also keep track of the most momentous handful of plays. Right now this isn't a super exciting leaderboard. I'm only keeping track of one game, namely the (insane!) Royals/Athletics Wild Card game from last night, because Baseball-Reference (which, unlike FanGraphs, gives me WPA numbers to the tenth of a percent) doesn't put up its box scores until the next day. So right now this is just the WPA numbers from that amazing ballgame divided by eight (because if you win you have a 1/8 chance of winning, whereas if you lose you have, well, a 0/8 chance of winning, so there's one-eighth of WSWE on the line). Tomorrow's update will include the WPA numbers from both Wild Card games, divided by the same constant! But then after that we'll start having players with multiple games played and games of different leverages, so things will start getting interesting. Anyway, leaderboard is below the fold.



Top 10 players by WSWPA:

Eric Hosmer, Kansas City Royals, +7.38%
Brandon Moss, Oakland Athletics, +4.29%
Christian Colon, Kansas City Royals, +3.18%
Josh Reddick, Oakland Athletics, +2.81%
Billy Butler, Kansas City Royals, +2.69%
Brandon Finnegan, Kansas City Royals, +2.64%
Alberto Callaspo, Oakland Athletics, +2.10%
Josh Willingham, Kansas City Royals, +1.68%
Jarrod Dyson, Kansas City Royals, +1.49%
Josh Donaldson, Oakland Athletics, +1.29%

Note that 100% of those players are batters (the pitchers combined to pitch well enough to lose three baseball games between the two teams), and also that four of them played for the losing team. Josh Willingham did precisely one thing in the game, namely hit a single. Eric Hosmer really earned that spot at the top, with five different positive events, one of which was the third-most-important play of the game. Brandon Moss, obviously, spent a long time looking like the story of the night. Ah well. Brandon Finnegan played in the college World Series this year, apparently. (I wonder what his cWSWPA was!) More than 100% of Jarrod Dyson's value came on a steal of third base (he gave a touch back with a sacrifice bunt, which isn't really fair to him but whatever).

Bottom 10 players by WSWPA:

Jason Hammel, Oakland Athletics, -5.46%
Yordano Ventura, Kansas City Royals, -4.26%
Dan Otero, Oakland Athletics, -3.78%
Jason Frasor, Kansas City Royals, -3.06%
Sean Doolittle, Oakland Athletics, -2.11%
Alex Gordon, Kansas City Royals, -1.65%
Jon Lester, Oakland Athletics, -1.61%
Jayson Nix, Kansas City Royals, -1.58%
Jed Lowrie, Oakland Athletics, -1.54%
Omar Infante, Kansas City Royals, -1.38%

Five players from each team. The winning pitcher is the fourth guy on that list. Six of the ten are pitchers, despite the fact that just 13 of the 41 players in the game were. Interestingly Jon Lester was already in negative territory entering that eighth inning, by which I mean he was sitting at -0.42%. Surprisingly little of the damage done was actually done with Lester on the mound; the very first thing that happened after Gregerson came in did more damage than Lester's entire four-batter stretch had. Sort of the bad version of Willingham, Jayson Nix appears on this list because he pinch-hit and struck out, going 0-1 on the day. That left the winning run on third base, though. Not a good entirety of your interaction with the game.

One interesting omission from these list is a guy who in a sense should be on both of them. Prior to the last event of the game, Salvador Perez had cost his team 35.8% WPA for the game, and therefore 4.48% WSWPA. Prior to that point, he had started the day by getting out right after Alex Gordon had also gotten out to open an inning three consecutive times. Then he struck out with the tying run at third and one out (on three pitches, I think none of them remotely in the zone. It was awful.) Then he passed a ball, I don't even know if that's included in these calculations but it might be. Then he grounded out to second with the winning run at third and two outs. His big whiff in the eighth inning cost nearly 19% of game WPA, and the failure in the 10th cost an additional 12%. And that was after starting the day a pretty routine 0-3.

Then he won the game.

His walk-off hit down the left-field line, scoring Christian Colon from second with two outs, sent the Royals from a 60.2% chance of winning the game, and hence a 7.525% chance of winning it all, to 100% and 12.5% odds, respectively. It therefore earned 39.8% gmWPA and 4.975% WSWPA, giving him a final figure for the night of +4.0% gmWPA and +0.5% wsWPA. Those numbers feel trivial. Salvador Perez had no net effect on that game. Except that's not right, because he practically lost the game for his team, twice, and then also very literally won it for them. Adding things up can lie to you sometimes.

Obviously the walk-off hit was the biggest event of the night. Here's a list of the top 5 plays so far this post-season by WSWPA:

AL WC Game, bottom of the 12th, 2 outs, runner on 2nd, OAK 8, KCR 8: Salvador Perez singles to left field off of Jason Hammel, Christian Colon scores. +4.975% WSWPA for Kansas City.
AL WC Game, top of the 6th, 0 outs, runners on 1st and 2nd, KCR 3, OAK 2: Brandon Moss homers to center field off of Yordano Ventura, Sam Fuld scores, Josh Donaldson scores. +3.975% WSWPA for Oakland.
AL WC Game, bottom of the 12th, 1 out, bases empty, OAK 8, KCR 7: Eric Hosmer triples to left field off of Dan Otero. +3.738% WSWPA for Kansas City.
AL WC Game, bottom of the 12th, 1 out, runner on 3rd, OAK 8, KCR 7: Christian Colon singles to third base off of Dan Otero, Eric Hosmer scores. +3.463% WSWPA for Kansas City.
AL WC Game, top of the 1st, 2 outs, runner on 1st, OAK 0, KCR 0: Brandon Moss homers to left field off of James Shields, Coco Crisp scores. +2.488% WSWPA for Oakland.

Basically that game was Brandon Moss hitting two home runs and then the bottom of the 12th. Except here's the thing. Here's a little spoiler: that list is not going to change in tomorrow's update! Not a single play from tonight's NL Wild Card game cracks the top-5! The biggest deal was obviously the Brandon Crawford grand slam, and it fell a tenth of a percentage point of gmWPA short of Moss's first home run. There were a full thirty-two additional plays from the AL game which ranked fully ahead of the 5th-most important play from the NL game. Thirty-two! 32! That's insane! That's like half an entire baseball game of plays that were more important than all but five plays from today's game! And that's in addition to the five listed above. It's probably not that uncommon for one team to have no more than 37 plays in its own half-innings for a whole baseball game. So when I say that game was Moss and the bottom of the 12th, it was that and then it was a whole hell of a lot more. Doing a "top-5" list from that game is just not doing it justice because my god what baseball game.

And by November, none of those plays will still be on my leaderboard.

For what it's worth, the Wild Card games were the highest-leverage games we'll see for a little bit at least. The next chance of a game with 12.5% of WSWE on the line would be a Game 3 of a Division Series tied 1-1. Any Game 4 would also have as 12.5% Leverage Index, and a Game 5 would obviously have a full 25% on the line. But it's entirely possible for a Division Series never to have a game as important as these: if one team sweeps, the first two games had Leverage Indices of 9.4%, and the last one of 6.3%. So the events of that game are going to stick around at least until we get into the League Championship Series, probably, unless we see any Game 5's.

I will try to do one of these updates daily, or at least for every day when there's baseball, but I am also a law student with, y'know, stuff to do, so that may not happen.