Giving High-FIP Pitchers a Mulligan

High-FIP pitchers have a pretty rough time of it these days. FIP gets thrown around in common conversation more than ever and, per the nature of FIP (or more accurately, xFIP), high-FIP players must pretty squarely pin their performance woes on themselves.

FIP can be pretty unforgiving. Outside of a seasonal constant, it won’t give a pitcher a break for how the baseball is traveling in, say, 2019. The coefficient assigned to home runs allowed (13) is far greater than the coefficient for any of the other stats incorporated in FIP. One truly wretched appearance can dog a pitcher for a whole season, being only partially diluted by their broader body of work.

To illustrate these points, what follows is a brief glance at those pitchers whose FIP in 2019 left something to be desired, and what exactly their singular worst starts did for their seasonal statistics. As you’ll see, one poor start can have an unfortunate and disproportional impact on a season as a whole.

For context, qualifying pitchers whose season FIP was north of 4.50 are those featured below. They are presented, with those statistics that are incorporated into FIP, in descending order by 2019 FIP. 

There are 16 players on this “leaderboard”, and 61 pitchers qualified by IP overall, meaning roughly one quarter of qualified pitchers owned an FIP north of 4.50 in 2019. Mike Leake, who had himself a Tomlin-esque season, sits atop this table after having given up more than a home run per five innings pitched.

So, a lot of home runs across the board (and these players really can’t be called a complete aberration for 2019 either). Walks also played an antagonizing role in many of these players’ 2019; Dakota Hudson, Julio Teheran, and Sandy Alcantara each allowed over 80 free passes, and 7 of 16 walked more per 9 than was average for a starter last season.

With a summary of each pitcher’s 2019 season in mind, this next table features each player’s most traumatizing chapter.

All but two of these pitchers gave up 3+ home runs in their season’s worst FIP performance. Four of 16 pitchers did not last 5 complete innings. While each of these individual’s strongest performance in 2019 surely worked to counteract the negative fallout from these outings, it’s hard to imagine they would be negated completely, or frankly even nearly.

The next two views offer each player’s seasonal statistics should their worst start of the season have been forgotten. I refer to this as a mulligan, but that isn’t necessarily a completely apt description, as it suggests each pitcher would have had another attempt at these rough starts, as opposed to simply striking them from their records. If this were intended to be even slightly more rigorous-ish, another option might have been to trim each individuals worst and best starts, in order to more fairly truncate potential outlier performances. Another option would have been to winsorize, or replace each personal-worst appearance with that of the second worst, a method not so drastic as excluding that start completely. But, to keep things simple, those starts are simply pulled from the larger body of work.

The first table includes the hypothetical change to FIP for each player in raw terms. The second table includes hypothetical adjustments for each player based on their FIP percentile.

Put simply, sans their worst appearance, these pitchers stack up much better. On average, more than a fifth of a point of FIP (0.21) is shed to go along with those poor starts, a considerable amount. In terms of percentiles, any one of these pitcher’s FIP improves anywhere from 5% to 16% as compared to the rest of 2019’s qualifying cohort. These pitchers made over 31 starts on averaging in 2019; that just one start, or roughly 3% of their appearances, can so impact the bottom line is pretty significant. 

Tangentially, quantifying the impact of a single brutal performance also may underscore various players’ frustration with the Houston Astros. If just one outing can so dramatically impact a season stat line (and particularly dramatically for players with fewer innings pitched), one gets a more concrete view of the frustrations individuals who pitched against the Astros in the 2017/18 seasons must feel now. Ahead of his going on record, Mike Fiers‘ worst (and second worst) outing from 2019 came against the Astros. 

Of course, it’s not just the high FIP players listed above who happen to get shelled once or thrice, and it’s maybe unfair to bemoan only their single-start fortunes. The NL’s reigning ERA leader, Hyun-Jin Ryu, saw a rough performance in Colorado balloon his overall FIP by 0.19 points when all was said and done. 

What this exercise does do is highlight some of the limitations of quantifying performances strictly by averages, especially when severe outliers present themselves. Sample sizes for starters are too small, and the circumstances of outings too varied, to embrace median performance, but searching for nuance in any given player’s average performance seems well worth doing. Because with a cumulative average, that one disastrous performance will always be baked in, regardless of how representative it is of a pitcher’s more usual production.