Thursday, August 16, 2007

TGWNA Salutes the "Quality Start"

I just asked a good friend of mine, who also happens to be a professional sports reporter (we'll all pretend that we don't know who he is), what one stat he would look to, if he could only choose one, to judge the quality of a starting rotation/pitching staff (I realize the two aren't the same thing, but I don't want to make this such a narrow qualification.

His response: Starter's WHIP, or team ERA.

A week ago, I probably would have agreed with the team ERA part, but WHIP's a little too narrow as a metric to be something I'd look at to give me a "general quality." Unfortunately, team ERA lacks for context; it tells us what happened over the entire season, on average, as opposed to what happened in smaller trials, like the games the team in question actually needed to actually win to actually get to the actual playoffs.

If I may, an aside: I was reading something on some site from earlier this year that was talking about the Yankees and their massive differential between their actual record and pythagorean record. The person writing about this — unfortunately, I cannot remember who — commented that the reason for the divide was that the Yankees were "losing a lot of close games and winning a lot of blowouts," or something to that effect. My reflex to these kinds of statements is "Quantify that, cocksucker," but I thought about it for a second, and realized that this probably happens more than most stat-heads are willing to admit. Teams with awesome offenses tend to have huge variances in game-to-game performance (eight runs one night, two the next) whereas teams with shitty pitching staffs always seem to get shelled.

But here's where I'm going to sound almost like I'm admitting that Joe Morgan is right about something: The goal of any individual pitcher, and a pitching staff as a whole, is to create winning situations. That doesn't mean that wins, as a stat, are useful. It does mean, though, that the most important thing a starting pitcher can do is give his team a reasonable chance of winning, and the best available statistic for measuring that is the quality start.

Apparently, few agree with me, since finding quality start numbers on the web is surprisingly difficult. The venerable ESPN has the QS numbers from 2006 and this season to date, but nothing before. I can't find an older database anywhere, which makes it a little harder to prove my point by the numbers. But I'll try anyway.

While watching the D'Backs continue to mount one of the strangest pennant runs I can remember, it occurred to me that, for a pitching staff that was considered so good, the D'Backs were light on pitchers I would consider being worth much (excluding Brandon Webb, of course). If you asked me, at the beginning of the year, what word came to mind when I thought of Livan Hernandez and Doug Davis, I would have immediately said "walks," followed by "losses," and "justifiable homicide." It would not occurred to me that both would be among the most reliable starters in the NL when it comes to delivering eminently winnable games for their teams. In fact, Davis and Hernandez have been as reliable as Webb has been.

Brandon Webb: 25 starts, 16 quality starts
Doug Davis: 25 starts, 16 quality starts
Livan Hernandez: 24 starts, 15 quality starts

That's certainly not what you'd expect without looking at the numbers.

The beauty of the quality start is that, instead of attempting to eliminate noise like many advanced statistic, it embraces an Occam's Razor-like approach to distilling each individual start into something simple and meaningful: did the pitcher give his team an excellent chance of winning with only a pedestrian amount of offense? Any time, in today's game, a starter can hold a team to three or fewer runs while throwing at least six innings, it's a borderline crime for that starter to not get a win.

In the case of the D'Backs, people have trotted out a lot of reasons as to why they've been able to keep winning despite a run differential that compares (un)favorably with a team like the Marlins. The Snakes' bullpen is excellent, which is a big factor, and it appears they're acting as a converse to the Yankees, by losing big (22 losses by five or more runs) and winning small (26 one-run wins). And it's worth mentioning that not many people think Arizona can continue to defy the pythagorean gods; BP's Joe Sheehan said in Thursday's column about the Snakes, "The Diamondbacks are simply unlikely to keep playing at a 90-win pace while also being outscored, no matter how good their bullpen performs. They will likely see either their run differential improve, or their record sag, because run differential is a more powerful indicator of team quality and future performance than actual record is."

But I really think the biggest factor this year has been the steadiness of the staff's top three starters, who have enabled a team with a fledgling offense (read: inconsistent and low-scoring) to be in a ton of games where a modest offensive output could earn a win. Webb has done this by being the awesome pitcher he's been for the last three seasons, while Hernandez and Davis have done this despite mediocre "standard" statistics like ERA, WHIP and K/BB.

Here is the quality starts top-10 for 2007 so far:

LA Angels - 68
NY Mets - 68
Cleveland - 67
San Diego - 67
Oakland - 66
Arizona - 65
Chicago Sox - 64
Toronto - 62
San Francisco - 62
Chicago Cubs - 61

Not surprisingly, this list includes most of the best teams in the league. The Angels, Mets, and D'Backs lead their respective divisions, while the Indians, Padres and Cubs are all either the WC leaders or within a game or two of a playoff spot. The A's and Blue Jays are mediocre teams, and the Sox & Giants are downright horrible. But in the cases of the A's, Sox and Giants, you're talking about three of the five worst offenses in the major leagues; they're the very definition of teams that could drop a 20-game-loser on Walter Johnson himself.

But, in addition to the D'Backs, I want to point out the presence of the Mets at the top. The Angels, co-leaders in quality starts, are constantly lauded for having one of the best pitching staffs in the majors. The Mets' "pitching woes" are a regular subject of debate on Baseball Tonight. The reason is name recognition and expectations: Most everyone expected the Mets' starting rotation to be horrible this year, while seeing that the Angels' staff is good simply confirms what everyone expected.

It's taken long enough to get this up, and while I have more to say on the subject, I'll post this now and write more later.

No comments: