Monday, May 21, 2007

No Two Yankees are Not on Fire

Four hamstring injuries (and a fired strength/conditioning coach), a broken leg, a broken finger, a struggling foreign import, and Carla Pavano. These are the days of our Yankee lives this season, and the result has been a frustratingly bad record of 19-23 heading into a three game series against the Red Sox in the Bronx. But I come here not to bury the Yankees, but to praise them, and offer hope. Hang on a second, while I put these on.

Okay, I'm back. We shall begin, as we normally do on this site, with Pythagoras. The three or four people who regularly read this site are rolling their eyes, sure. But it should make intuitive sense to anyone who thinks reasonably about sports that if a team averages a certain number of runs per game and allows a certain number of runs per game, that team will, over the course of a season, win a predictable number of games. This is so, and repeatedly proven by the folks at Baseball Prospectus, within a range of 3 wins or so (which I think is pretty damn good). With this, we can then see if a team, during a season, is over or underacheiving its expectations based on the number of runs it scores.
Indeed, the expected win percentages calculated by Baseball Prospectus show the Yankees as one of the most underacheiving teams of 2007. Only the Cubs and Reds have similar underperformance, while the Brewers seem to be the most overacheiving. I use the word acheive specifically, as opposed to performance; the latter word is exactly what is suggested by the Pythagorean Win Percentage. And the PWP suggest that the Yankees should be 23-19. They have scored 227 runs (5.4/game, tied for 3rd in the ML) and allowed 200 runs (4.8/game, 25th in ML). While the pitching isn't great, it makes sense that a team that scores more than it allows should be above .500, at least.
So why are the Yankees so underacheiving? The classic explanations, which I haven't seen convincingly proven anywhere, are that teams fall below their Pythag Win Percentage when they fail to win their share of close games. The Yankees are 2-8 in one-run games. They also suggest that having a poor closer would sway this one-run game success, and unfortunately the greatest closer of all time has blown three saves this year for the Yankees (and accounts for 3 of the 8 one-run losses). On the offensive side, some would suggest that the result is the ever-elusive Clutch Hitting. While we can argue till doomsday whether it exists for individual players, the sample for a whole team should be big enough to draw some conclusions. So, the Yankees offensive stats:

So there is some evidence that the Yankees are underperforming in "Late & Close" situations (seventh inning on with the game either tied, winning by one run, or with at least the tying run on deck). Part of this can be explained by the 9th inning: hitters are going to do worse against the best pitcher in the opposing teams' bullpen (the closer, theoretically).

My point isn't one of "everything is fine, nothing is ruined." To the contrary, I think that "luck," and specifically better performance in one run games, is only part of the story of a 2007 team turnaround. The other part is better pitching. Allowing 4.8 runs per game just isn't all that good, even if it does project to a winning record with the Yankees' offense. The pitching simply needs to get better. So it's time to do some fuzzy math. The Yankees have played 42 games. Forty-two, in addition to being the answer to the Ultimate Question, is also a little more than one-fourth of the season. The second quarter will feature a similar schedule, as both sections contain six games against the Red Sox, three against the Mets, and three against Oakland. Series against Pittsburgh and Arizona are comparable to series against Tampa Bay and Seattle, while Colorado and Texas are similar. So let's see what could potentially happen over the next 42 games.

To this point, Yankees starters not named Wang, Pettitte, or Mussina have pitched 104.6 innings, giving up 74 runs, or 6.36 runs/game. That is awful. Now, let's have some fun: Roger Clemens, in 2006, pitched 113 innings and gave up 34 runs. Baseball Prospectus has adjusted for his age and move to the American League and projects an ERA of 3.36 for Clemens in 2007. The Rocket could make his first start for the Yankees on May 28 against Toronto; assuming he pitches every five days that gives him seven starts in the Yankees next 42 games. In addition, some combination of Tyler Clippard, Matt DeSalvo, and Phil Hughes will also be making starts every fifth day instead of Kei Igawa and Friends. Those three have combined, so far, to allow 12 runs in 33.7 innings (3.2 runs25-16. /game). PECOTA has Clippard at 4.88 ERA and Hughes at 3.78 ERA for 2007 in the majors (no card for DeSalvo, which isn't surprising).

So let's assume, for a minute, that Wang, Pettitte, Mussina, the Yankees' bullpen and offense all perform similarly for the next 42 games. In this experiment all we're changing is the identity of the Yankees' fourth (Clemens) and fifth (Hughes/Clippard/DeSalvo) starters. They will replace pitchers who in roughly, 27% of all innings for the Yankees, and allowed 6.4 runs/game. By my calculations, if the new guys lower the 4th/5th starter runs/game to the level of the Big Three and the Bullpen (about 4.36 runs/game), the Yankees will have a Pythagorean record of 26-16. That's a three game improvement over the first 42 games. Best case- the Yankees exceed this by a standard error (let's say 2 games) and end up, after 84 games, at 47-37. Of course it is possible that the one-run games futility continues and the Yankees underacheive their Pythagorean expectations just as badly as in the first 42 games- it did happen to the Cleveland Indians last year, who ended up 12 games below their expected win total.

I think it has been shown, though, that it is reasonable to expect the Yankees to perform substantially better in the second quarter of the season- not only because of "bad luck" tossed their way via injury and failure in close games, but also because the pitching rotation will further increase the gap between runs scored and runs allowed. In the meantime, it is useless and only discouraging to look up at Boston in the standings; as long as the Yankees take care of business against non-Boston opponents, they are Doing Their Job. After next weekend the Bombers don't play the Red Sox until September, so leaderboard watching is useless. Besides, there is another way into the playoffs besides winning the division. Ask the Marlins.

No comments: