Using Advanced Metrics to Evaluate Pitching Quality: xFIP

by waiverwire Last Updated 2014-09-01 21:24:50

Using Advanced Metrics to Evaluate Pitching Quality: xFIP

In addition to a full time job as a software product manager, Alex Zelvin works part time for Fanduel.com (Zoobird on FanDuel) and co-owns Dailybaseballdata.com

A New ‘ERA’ of Measurement

The most important point in this article can be summed up in just one sentence: You should be using xFIP to evaluate how well a pitcher has pitched, rather than using ERA. Another way to say that is that xFIP is a better predictor of future ERA than past ERA is.

What is xFIP?

xFIP is a statistic that uses the parts of a pitcher’s performance that a pitcher is primarily responsible for to measure how well a pitcher pitched. For the most part, research has shown that pitchers have a great deal of control over strikeouts, walks, and home runs, while other aspects of their performance are primarily (although not entirely) outside of the pitchers control. Further, the primary means by which pitchers control home runs is through their ground ball rate. So by creating a statistic based entirely on strikeout rate, walk rate, and groundball rate, we get a much better predictor of future performance. xFIP is that statistic, and it largely eliminates the influence of defense, park, and random chance on our evaluation of pitching performance.

xFIP is based on FIP (Fielding Independent Pitching). The formula for FIP is:

(HR*13+(BB+HBP-IBB)*3-K*2)/IP, plus a league-specific factor (usually around 3.2) to round out the number to an equivalent ERA number.

xFIP simply replaces the actual number of HR allowed with the number of HR that would have been allowed assuming that the pitcher’s HR per fly ball rate was league average. Typically, the league average is around .11 per fly ball. Knowing how to calculate xFIP is great is you want to use projections for the various inputs to project xFIP. On the other hand, if you just want to know a player’s xFIP to get an idea of how he’s performed so far this season (or in past seasons), then you can simply look it up in the player pages at Fangraphs.com.

Keep in mind that xFIP is essentially a measure of how well a pitcher has pitched. If you want to use it to forecast how well a pitcher will pitch in the future, remember to make adjustments for park and defense…unlike ERA, xFIP does not implicitly factor those in by including results of balls in play.

One popular way to use xFIP is to identify pitchers whose xFIP is either much better or much worse than their ERA. In theory, these are pitchers who have likely gotten very lucky, and are likely to have a reversal of fortune.

Here are some of those for 2010 whose xFIP indicates have pitched much better than their ERA shows. The number next to their name is the difference between the ERA and xFIP for that pitcher so far this season.

Here are some whose xFIP indicates that they’ve gotten lucky to some extent. That doesn’t mean they’re not great pitchers, just that they shouldn’t be expected to maintain as low an ERA going forward as they’ve achieved so far:

Keep in mind that the difference between xFIP and ERA is only partially created by ‘luck’. It is also impacted by park and defense…so some of these are likely to continue, just not to such an extreme degree. In addition, certain pitchers seem to have a tendency to underperform (or outperform) their xFIP.