Sox Watch

Well, Mirabelli finally homered. That's about all I really want to say about this game.

Actually, I'll say a bit more, because there are a bunch of interesting WPA lessons here. Focusing on the stats helps distract me from what a painful game this was.

First of all, you might wonder, in such an abjectly miserable game, how Lowell, Youkilis, Crisp, Nixon, and Gonzalez managed to turn in fairly hefty positive WPA totals. The answer is simply that Beckett was so bad that he absorbed most of the negative WPA himself. By time the second inning was (finally) over, the game was essentially lost, and Beckett left the game with a -0.540 WPA total. It's pretty tough to accumulate more than 0.500 in negative WPA for a game singlehandedly. In fact, this is the first time it has happened all year. Going over 0.500 means that you essentially lost the entire game, and more, yourself.

Each team enters the game with a 0.500 (50%) WP (Win Probability), and by the end of the game, the Sox' WP was 0.000. So they lost 0.500 along the way. So how is it possible for Beckett to lose more than 0.500, with a total of -0.540 for the game? The answer is that the WP goes up and down over the course of the game, which means that in some cases, the pitcher is giving back WP that the offense has previously earned. That is what happened here. At the end of the top of the first inning, the Sox' WP was 0.452. In the bottom of the inning, the Yankees scored a run, and by the end of the first, the Sox' WP was down to 0.397. Varitek took some of the debit for his unusual defensive manuevers during the Cabrera run, and Beckett absorbed a total of -0.038 WPA for the inning. Then, in the top of the second, the Sox got two runs back, making the score 2-1 by the end of the top of the second. At this point, they have a WP of 0.557. Then, over the course of the bottom of the second, Beckett proceeds to squander most of that total. By the time he leaves the game, the WP is down to 0.055, and he has lost 0.502 in the inning. This, combined with the 0.038 he lost in the first, produces his total of -0.540 for the night.

One odd side effect of Beckett's huge WPA number is that it means that the rest of the team actually had a net positive WPA of 0.040 for the night. This is where the numbers for Lowell, Youk, Crisp, and Nixon come from. Crisp doubled in the first for +0.058 WPA, while the three straight hits for Nixon, Youk, and Lowell in the second were worth a total of +0.224. Gonzo's positive number came from his defense. I gave him 75% of the credit for hauling in Beckett's errant throw to second to force Damon on the FC play where Cabrera reached.

Another anomaly here is Mussina, who pitched a pretty strong game, striking out eight and allowing four ER. He probably would have gone more than six innings in a tighter game. Yet he finishes the night with a -0.072 WPA, the worst on the Yankees. At first glance, this seems strange. But you have to remember what WPA is telling you. It's not telling you how good the player is, or even how well he played. It's telling you how much he contributed to the victory (or loss), and so it depends tremendously on the game situations each player is performing in. In Mussina's case, his WPA simply reflects the fact that the NY offense won the game last night. Mussina gets a minor positive WPA for getting through the first inning unscathed, and then a larger negative WPA for giving up the lead in the second, making him net negative at that point. After that, the Yankees bats take over, and by the time Mussina walks onto the mound in the third, the game is over - it doesn't really matter how well, or how badly, he pitches from this point forward. His WPA total from the third inning onward is only 0.040. This is also why the totals for Riske, Tavarez, Delcarmen, Proctor, Villone, and Rivera are all very close to zero. Combined, those six pitchers threw 8.1 innings of shutout ball (Riske allowed a couple inherited runners to score), but their WPA total is a combined 0.009, because none of them were pitching in situations where their performance had a big impact on the outcome of the game.

Finally, I'd like to do a WPA analysis of the bizarre play that brought Melky Cabrera around to score in the bottom of the first. I break baserunning plays up into smaller segments to properly record WPA debit/credit to the players involved. See, for example, the play on May 28 where Manny threw Joey Gathright out at the plate. Last night, we had Cabrera on first in the bottom of the first with one out in a scoreless game. The Sox WP at that point was 0.450. As Cabrera advanced around the bases, the WP diminished at each point:


With a man on first,  WP = 0.450
With a man on second, WP = 0.434 (a loss of 0.016)
With a man on third,  WP = 0.407 (a loss of 0.027)
With the run in,      WP = 0.374 (a loss of 0.033)
Total net loss           = -0.076

So we have a total of -0.076 of WPA to divide up among the players involved. I broke it down like this:

Advancing from first to second - This was due to the wild pitch. I blame Beckett for 75% of this and Varitek for 25%, splitting the -0.016 accordingly, so Beckett gets -0.012 and Varitek gets -0.004.
Advancing from second to third - This was due entirely to Varitek's throw to second. I assigned all of the -0.027 WPA for this portion to Varitek. I could have chosen to assign some of this to Lowell for missing the throw to second, but I felt that this was an uncatchable throw.
Advancing from third to home to score - This was tougher to account for. As Remy said during the game, nobody practices for this type of situation, so it's not clear what the assigned roles are. But the fact is that everyone was out of position and nobody was guarding the plate. I figure that either Beckett or Varitek should have been there, so I divide the -0.033 for this play evenly between them, which is about -0.0165 for each.

So at the end of the play, we have Beckett with -0.028 and Varitek to -0.048. This gives Beckett about 37% of the debit for the play, and Varitek about 63%. This fits pretty well with my intuition about the play. Beckett kicked things off with a bad (but not horrible pitch), and finished up with a bad throw back to the plate, while Varitek takes a larger share for missing the WP, the crazy throw to second, and getting out of position at the end of the play.

Individual Player WPA Contributions
Monday, 6/5/06
Red Sox 5, Yankees 13