Why the Mets' pitching fWAR is so low

Mets pitchers had the lowest FanGraphs WAR in baseball in 2014. What gives?

The consensus among baseball fans—especially Mets fans—is that every beat writer hates their team. Still, just about every writer lauds the young pitching depth that the Mets have amassed. As fans sift through #LOLMets reports of Frank Francisco and what kind of poultry he compares opponents to, whether or not David Wright’s shoulder is still attached to his body, or the next juicy story regarding ownership, one unanimously positive theme is that the team’s young pitching is bright and dandy.

So when the educated fan sojourns to FanGraphs’ wonderful leaderboards, takes a look at the team rankings for pitching, and sees our beloved Mets at dead last, it feels as though someone has stripped us of our last shred of dignity. We had Zack Wheeler, a big gun-slinger with a swanky curveball; Jacob deGrom, the best rookie in the National League; a pretty cost-controlled Niese; and some shiny new righties in the bullpen.

A statistical quirk

Maybe the Mets are not quite the Nationals yet, but dead last? Something’s gotta give. We can’t just discredit a reasonably good estimate of marginal value when the sample size is an entire team (and an entire league), but we know that the Mets' pitching just isn’t that bad. In fact, it's probably pretty good. So let’s take a look at how the Mets did, compared to the league, in various pitching categories:

The Mets got a bit lucky last season, as evidenced by their high left-on-base rate. This metric, also dubbed "strand rate," varies widely from year to year among pitchers, although recent studies indicate that some portion of it is explained by pitchers who don't give up many extra-base hits. Part of that number can be attributed to Jeurys Familia, Jenrry Mejia, and Vic Black, who are very difficult to make solid contact off of; but most of it is likely due to some good luck.

The team also allowed a few too many home runs, although some of the biggest offenders—Dillon Gee, Gonzalez Germen, Rafael Montero, John Lannan, Jose Valverde—don’t figure to make the Opening Day roster next season. For the most part, though, the Mets were about average . . . until you look at their fWAR.

A park-factor problem

First, it’s worth pointing out that fWAR is inherently a marginal estimate, unlike the previously mentioned stats, so while the previous indexed numbers generally show values very close to their mean, fWAR necessarily has a large variance, because it is supposed to represent wins above a certain baseline number somewhat accurately. Still, the Mets' pitchers were ranked dead last by fWAR, and the reason is not immediately obvious from the table above. My best guess? Citi Field’s park factor. It’s likely that any Met’s pitching performance was severely downgraded by the fact that fWAR is adjusted for park factor. In this case, I am nearly certain that park factors over-corrected.

Park factors are known in the public domain as an indexed adjusting factor for each stadium in the league. Because each stadium is in a different climate and has different dimensions, adjusting for park often clarifies the run-scoring environment in which teams play, and can give some added context to a player's or team's performance. The method of calculating park factors, however, is inherently flawed.

Park Factors are a results-based way of assessing the park environment that depends on how the teams who played in that park performed there. This lends itself to fluctuation based on the talent and skill level of the team. For reference, we can look at Fenway Park’s walk factor for 2014: 1.168, the highest in the league. In other words, we’d expect walks to be much more frequent at Fenway Park. This is already a strange proposition to begin with, but what becomes stranger is that the 2013 Fenway Park walk factor was 0.915, 26th overall.

Did the Sox exclusively change the strike zone in their home ballpark in 2014? It seems unlikely. What is true about 2014, however, is that the team’s pitching was an absolute wreck compared to the previous season. It’s likely that the team’s pitching staff had quite a bit to do with its park factor, which negates the entire point of the metric to begin with.

There are also weird trends within the same season, and one needs to look no further than Citi Field in 2014. Citi Field's overall park factor was 0.847, making it the third-worst offensive environment, behind only Petco Park and Safeco Field. But the field's home run factor was much less dramatic, ranking 18th in the league. One would imagine that home runs would be most affected by the ballpark's run-scoring environment, but that doesn't appear to be the case.

According to Citi Field's park factor, triples seem to be most affected by the stadium, with a league-low triples factor of 0.348. However, triples are largely random occurrences that happen due to a runner's speed, defensive hijinks, or both. It would make a bit of sense if Citi Field were perfectly symmetrical and therefore lent itself to fewer odd bounces, but in some areas the outfield walls were a bit wacky. Regardless, an effect this pronounced doesn't make much sense, which offers further proof that current park factors are flawed.

So what gives?

Over at FanGraphs, Tony Blengino recently posted a glimpse of his HITf/x-powered park factors. His rankings, which are based off of data much more relevant to the problem that park factors attempt to solve, suggest that Citi Field’s left field area is one of the most favorable to hitters in the game. This may indicate that the adjustment made to the team’s fWAR should inflate instead of deflate the number, or at least just not change it much.

Anything else?

While the park-factor adjustment may address why the team’s pitching fWAR is so low, another issue remains: The team’s pitching performance last season was still at the middle of the pack. Only xFIP put the Mets at above average, and just slightly so. Matt Harvey, Noah Syndergaard, Bobby Parnell, and Steven Matz could all help remedy that situation for 2015, but with two of those pitchers recovering from Tommy John surgery and the other two lacking positive experience at higher levels, I wouldn't bet on it. These options also leave the bullpen mostly unaddressed.

The Mets’ bullpen was exciting last season and often looked dominant, but it perhaps was not as successful as we might think. Pitching around the league is better, and comparatively the Mets' looks less impressive. The team ranked 14th best in bullpen xFIP and tied for 15th in bullpen SIERA. Again, it seems that the team's bullpen was decidedly middle of the road, and only adding Bobby Parnell after surgery may not be enough to improve its performance substantially.

In addition, many doubts remain about the bullpen's ability to succeed in 2015. We can only hope that the team's strand rate luck doesn't swing the other way, that Michael Cuddyer's defense in the outfield is mitigated by Juan Lagares's prowess, and that the new right field wall won't handicap the team against left-handed hitters.

The Mets will need to make a defensive improvement this offseason without crippling a fragile offense, a tough task to be sure. Whether it comes in the form of another bullpen ace or a realigned middle infield is anybody’s guess, but this is the biggest issue to be addressed this offseason post-Cuddyer—even if the Mets aren't actually the worst pitching staff in the game.