Quality Starts: Just One Part of the Story

I kept hitting Refresh yesterday on my Internet Explorer as I watched one of my favorite young sleepers – Tampa Bay’s Matt Garza – throw an absolute gem in his first start of the year against the Red Sox. In Boston, mind you. 7 IP, four hits, five strikeouts and one earned run in what turned out to be a win. But just barely.

My Refresh Rate increased significantly during innings 7-8-9 as I watched Brian Shouse (who?) and Joe Nelson (thought it was Jeff) tag team the eighth through a gut-wrenching pair of hits, a walk and a run that narrowed the Tampa Bay’s lead to 4-2. By mid-afternoon, I was hitting Refresh in between the 30-second auto-refresh on ESPN (until I remembered Game Cast). Troy Percival popped in for the ninth and promptly let up a blast to Jason Varitek that forced me to release an expletive into the sedate, intra-cubicle common area as I slowly became convinced that Garza’s effort would be sacrificed in the name of old hacks who can barely hit 90 anymore including one 40-year old (this Brian Shouse character).

And thus I began to ponder: How often has this happened, and who are the most frequent victims? Garza ended up winning the game as Percival managed to fumble his way through the middle of the Boston lineup, but the question stands.

Earlier this year, the MLB Owner’s Edge blog on FanBall took up a discussion on “Quality starts: Are they really quality?” that brought into question the QS statistic tracked since about 1985 that categorizes a “quality start” as one with six innings pitcher and three runs allowed. It’s worth noting, as Ray Flowers did in the post, that a start with 6 IP and 3 ER still results in an ERA of 4.50 for the outing, arguably a bit high for a “quality” start.

My thoughts took on another tangent, as stated earlier, that focused on the conversation of Quality Starts (I’ll keep them, since they are pretty much standard) to wins by a pitcher. You can make the argument that QS shouldn’t be the measure, but I personally think it’s at least a decent proxy and worthy of using as a standard. Below you’ll find a breakdown of all pitchers in 2008 with at least 21 QS.

Some familiar names, as expected. But the range really isn’t remarkable, don’t you think? From 8 wins to 22. 3 losses to 15. ERAs of 2.53 to 4.09. This obviously results from poor run support or a sucky bullpen, but let’s put some data behind it.

Below is a ranking of W/QS for pitchers with at least 21 QS. You can see that

This stat simply gives us an idea of how often the pitcher won compared to every time he handed over to his bullpen after at least six innings, with no more than three runs allowed. A few usual suspects here, too. Cliff Lee, Mike Mussina, each guys who probably are converting

A quick point that guys like Dice-K, though outside the bounds of our study here, also fall into this category, with 18 wins for 14 quality starts. That’s over a 100% conversion percentage. Either he was very lucky, good at getting out of jams, was supported by a good offense, or held up by a solid bullpen, or potentially some combination of the three.

One more extension of this, and I’ll leave it to the good reader to think over for his or herself. What if we include Run Support (RS) for each individual pitcher? Run Support is essentially the unconstrained (both earned and unearned) ERA of the opponent’s pitchers for every game our pitcher appears in. Ranked from greatest RS to least, we see the statistical result of good efforts from good pitchers, sacrificed by any one or many of the factors I listed above.

Finally, a rare graph, but one that I can’t help including.

So what is (are) the point(s) here?

1) QS doesn’t mean much; a QS to TS (Total Starts) percentage of above 60% is probably a pretty good indicator of a “good” pitcher, regardless of team, but it doesn’t mean the guy will win the Cy Young. Heck, he might not even put together wins in the double digits.

2) The best pitchers in the league right now find a way to win by going deep into games without a whole lot of run support. Roy Halladay is a prime example. The guy pitched 246 innings last year, but managed to win 20 games and convert almost 90% of his quality starts into wins despite an RS of just 5.45.

3) Pitchers who receive high RS have a hard time converting QS into W because either they are A) entirely erratic or B) have a bad bullpen. I’d venture that Buehrle is an example of the former (he was credited with 12 losses), while Haren is an example of the former (in 2008 the Arizona bullpen posted a record of 17-28).

3) Build your fantasy staff around pitchers on good teams, with good bullpens. This is obvious, right? But so many of us fall victims to the Matt Cain’s of the world. He’s a good pitcher, yes; but his statistical output – simply because of the team he plays on – may well be sub par. Now, if you’re drafting for ERA or Ks, go right ahead, you’ll probably reap some rewards there. But simple win totals are a really poor way to judge a pitcher’s overall performance. And without wins, a pitcher’s overall fantasy output drops significantly.

I’ll admit this is a small sample size; I apologize as I just upgraded to Microsoft Excel 2007 and now only have several dozen MB of free memory which is causing my computer to freeze up every 4 minutes or so. But I’d venture that this trifecta of QS/TS, W/QS and RS are integral to identifying solid pitching performances, regardless of team.

Share/Save