It’s SABR time: Fiddling with FIP

If you’ve been reading Halos Heaven articles any time over the past year, you’re bound to have stumbled onto the use of the stat FIP (fielding independent pitching). If you have, you’re probably wondering what exactly it is and why in the heck it’s being used so often. Here’s a tip of the proverbial cap and a primer to the most important advanced pitching statistic in existence, FIP.

Earned run average, or ERA, has long been a traditional stat used to judge pitcher performance, but it doesn’t actually accomplish this goal. Why? Well a number of reasons, namely:

a) It does not account for performance of the defense. Last year, the Chicago Cubs had one of the best defenses (and excellent defensive positioning from bench coaches) in baseball, which served to convert balls hit at fielders into outs at the best rate in the league. This meant that pitchers’ ERAs were remarkably lower than their true talent level because their defense significantly aided them in run prevention.

That can work to a pitcher’s disadvantage as well, as Ricky Nolasco was significantly hurt by his defense during his tenure with the Twins, putting up FIPs of 4.30, 3.51, 4.30 and ERAs of 5.38, 6.75, and 5.13. The culprit? One of the league’s worst defenses.

b) It does not account for inherited runners. A starting pitcher can go 6.2 scoreless innings, get pulled, and the reliever replacing gives up a home run which inflates the starter’s ERA. This is another key reason why ERA is a mostly empty stat, completely meaningless for relievers who pitch a small number of innings and therefore are heavily influenced by luck on balls in play.

c) It usually does not account for luck on BABIP, or batting average on balls in play. Against the Seattle Mariners on May 3rd, Blake Parker gave up 4 earned runs. However, the hits were all weakly hit, well placed groundball singles or miscues by fielders that were ruled as hits (as opposed to errors). His replacement, Bud Norris, came in and gave up a few more hits, and those hits knocked in runs that were charged to Parker. Parker’s ERA following the game was over 4, while his FIP was well under 2. This caused a conundrum after the game for those who believed ERA, because Parker had almost always looked brilliant in previous outings. Parker got BABIP’d, and his ERA has ‘reverted to the mean’, but has taken six weeks and about 20 outings for that to happen. FIP isn’t perfect, but it is almost always a better indication of true talent level than ERA.

So how exactly does FIP correct for ERA’s mistakes?

FIP only takes into account the actions that the pitcher can directly control: home runs, strikeouts, walks, and hit batters. For balls in play, FIP assumes average luck and average results, since most research has shown pitchers have little to no control over the results of balls in play.

The formula:

FIP = ((13*HR)+(3*(BB+HBP))-(2*K))/IP + constant

The constant allows us to use FIP on the ERA scale, so we can use in direct comparison. For reference, the average ERA across MLB is 4.36 this year. The average FIP is also 4.36.

Of course, FIP isn’t perfect, and some pitchers have batted ball profiles that allow them to consistently outperform their batted ball results. From Fangraphs, here is why some pitchers can outperform their FIP.

For example, pitchers with the ability to limit the running game or generate fly balls at the expense of line drives or ground balls are more likely to beat their FIP than the average player. This doesn’t mean that every lefty fly ball pitcher will do so, but simply that holding runners and generating a type of batted ball that falls for hits less frequently are legitimate skills that might allow you to limit your runs allowed.

If you have to bet on a pitcher’s ERA or their FIP, FIP is the better bet, but FIP tells you about a subset of a pitcher’s results which means that it is possible that it is missing something important about that pitcher’s profile that allows them to run consistently high or low BABIPs.

As an addendum to this, Fangraphs’ WAR (or fWAR) is FIP-based for pitchers, while Baseball-Reference’s WAR (bWAR) for pitchers is ERA-based. Over time with extremely large sample sizes, FIP and ERA eventually converge, though there are always exceptions to the rule.

FIP is your friend. Use it!