Oswalt and FIP

When we talk about pitchers, we often refer to a pitcher’s FIP, which builds an expected ERA based on a pitchers walk rate, strikeout rate, and home run rate. In general, FIP works a lot better than a stat like ERA, because it removes a lot of the noise from results that pitcher’s don’t have a lot of control over.

However, one of those components is a bit more fickle than the other two, and it can have a huge effect on a pitcher’s performance – that component, of course, is the home run rate. Let’s take Roy Oswalt, for instance.

In March through May, he made 12 starts, pitched 76 innings, and had a respectable 20/55 BB/K rate. However, he gave up 16 longballs during that stretch, so his FIP was a downright poor 5.35, which nearly matched his 5.45 ERA. Those home runs were killing him, even though he was commanding the strike zone fairly well.

Since June kicked in, though, no one’s been able to take Oswalt out of the park. In his last 17 starts, he’s thrown 114 2/3 IP, posted a better 23/95 BB/K rate, and cut his home runs allowed down to 5. He’s certainly pitching better, as the increase in strikeouts and decrease in walks show, but five home runs in almost 115 innings? That’s nutty. As such, his FIP for that stretch is 2.93.

Just so you can see how much the home run rate affects FIP, if we change Oswalt’s HR rate in his first 12 starts to match what he’s done in his last 17 starts, his FIP drops from 5.35 to 3.31 – in other words, that difference in the home run rate is worth two runs a game. As much as we love to evaluate pitchers by walks and strikeouts, home runs have a massive effect on their performances, and a pitcher can succeed in MLB simply by keeping the ball in the park.

However, because home run rate is far less of a skill than walk or strikeout rates, most pitcher’s can’t succeed for a long team just through HR suppression. That’s why BB/K rates are a better predictor of future success than HR rates, even though HR rates have more of an impact on run scoring.

Just a little reminder that while FIP is a nice tool, if you see a guy running strong based on a remarkable HR suppression streak, it’s not as likely to be real as if he’s doing it with no walks and lots of strikeouts.

Dave is the Managing Editor of FanGraphs.

Newest Most Voted
Inline Feedbacks
View all comments
15 years ago

Do we consider HR/FB to controllable or not? Assuming it’s essentially random, and given that the vast majority of HR are on balls in play classified as FB (as opposed to LD), wouldn’t we be better off using FB% in place of HR/9. It’s not quite as elegant as the /9 consistency, but it would be more true to the idea of controllable skill without adding complexity.