Which Starting Pitchers Aren't Living Up to Their Statcast Numbers?

Since Voros McCracken revolutionized baseball analysis with his Defense Independent Pitching theory, the primary way of assessing whether a pitcher has been “lucky” has revolved around BABIP.

This is still generally a good way of looking at things, as BABIP is incredibly volatile (particularly in small samples) and takes multiple seasons to become reliable. With the recent introduction of Statcast data, we have a new way to see whether a pitcher with poor results on batted balls is giving up suboptimal contact or is just unlucky (and vice-versa, of course).

Last winter, something clicked with FanGraph’s Andrew Perpetua. He found that his iteration of Expected wOBA -- which was based on launch angle and exit velocity -- predicted next-season wOBA better than last season’s actual wOBA. (For xOBA, the correlation coefficient was .38, while it was .31 for actual wOBA, with one implying a perfect relationship and 0 implying no correlation.)

Predicting BABIP from one year to the next remains a fool’s errand -- neither BABIP nor xBABIP predicts the future rate well at all -- but the Statcast data has a much stronger relationship with power (in the form of home runs and slugging percentage) than the past results.

Baseball Savant uses a slightly different calculation for Expected wOBA than Perpetua’s xStats.org, but the results are generally similar.

With that in mind, here are the four pitchers who have most underachieved relative to their Statcast-implied numbers this season, among those 155 pitchers who have allowed at least 250 balls in play. The stats are updated through Wednesday, August 16.

Anibal Sanchez, Detroit Tigers

wOBA: .398

xwOBA: .328

Diff: -0.070

ERA: 6.95

Sanchez’s 6.0 fWAR season in 2013 sure seems like a long time ago.

After following up that campaign with a solid season, Sanchez has battled injuries and ineffectiveness. This has continued into 2017.

Sanchez actually has a healthy 13.2% strikeout minus walk rate, which is mostly the product of a good 5.9% walk rate. However, his FIP is 6.20 in 77 â…” innings, due to a 2.7 home run per nine innings rate, which is the most of anyone who has thrown as many innings as he has.

Over the past three seasons, Sanchez has been an extreme flyball rate pitcher, so his 21.9% home run per flyball has been especially damaging.

The figure screams regression on its own, and the Statcast data seems to support this. He is tied for 69th in exit velocity on non-grounders among the 167 pitchers who have allowed at least 100 air balls, not what you would expect from someone with the 2.7 home run per nine rate.

Also, nine of his home runs have been “just enoughs,” while only four have been “no doubters,” per the ESPN Home Run Tracker. (Home Run Tracker explains that just enoughs are balls that, “cleared the fence by less than 10 vertical feet, or [land] less than one fence height past the fence,” while no doubters, “[clear] the fence by at least 20 vertical feet AND [land] at least 50 feet past the fence.")

Even if we look at things in the positive light, however, Sanchez still comes out as a below average pitcher. His xwOBA is more than four percent worse than league average, while his park-adjusted xFIP is nine percent worse.

Tyler Anderson, Colorado Rockies

wOBA: .392

xwOBA: .324

Diff: .068

ERA: 6.11

Anderson is an interesting case, as it would be tempting to merely chalk up his elevated wOBA to Coors Field.

In reality, that is not the only factor. His wOBA is actually lower at home -- we're talking .377 at Coors, but .403 on the road. Anderson has a 22.4% strikeout rate and a promising 14.2% K-BB percentage. He has been done in by the 15 big flies he has allowed, giving him a 2.1 HR/9 rate.

So far, we can see some similarities to Sanchez, but while Sanchez's career is on a steep downwards trajectory and probably does not warrant the benefit of the doubt, Anderson may actually have some promise.

Right off the bat, he presumably will not continue to allow a strand rate that is lower than 70% and his .337 BABIP is not supported by his .285 xBABIP.

That still leaves the home runs, as Anderson’s FIP- is 116.

According to xStats.org, Anderson “should have” allowed only 11 home runs. This is in between his actual home run rate and what it would be if he had a perfectly average home run per flyball rate (his xFIP is 4.02).

This makes sense, given his park adjusted xFIP is seven percent better than average, while his xwOBA is about three percent worse than the mean.

Bronson Arroyo, Cincinnati Reds

wOBA: .353

xwOBA: .418

Diff: -0.065

ERA: 7.35

Frankly, it does not seem worth anybody's time to go too in-depth on the 40-year-old Arroyo, who is currently on the 60-day disabled list.

His fielding independent numbers were bad (8.1% K-BB%, 134 xFIP-) and their combination with some bad luck/randomness translated to even worse results.

Rafael Montero, New York Mets

wOBA: 0.371

xwOBA: .315

Diff: -0.056

ERA: 5.80

Montero has had a pretty bizarre season.

After a putrid start (8.24 ERA, 5.17 FIP, 5.94 xFIP over his first 13 games, most of them out of the bullpen), he strung together a series of solid starts in June and July. Over his last four outings, the roller coaster has taken a downward turn again.

Over the season as a whole, his ERA- (141) stinks, while his FIP- (106) and xFIP- (108) are OK, and his xwOBA is almost perfectly average.

Unlike the other three players on this list, Montero’s divergence between his actual results and expected results owe more to BABIP than power.

Montero has allowed a .155 ISO, which is right in line with the MLB average of .153. His 13.3% home run per fly ball rate is almost identical to the big league mean, while his 46.9% groundball rate is higher than average.

His biggest problem has been a .373 BABIP, which is way out of whack with his underlying numbers. His line drive rate is only 17.5%, while his average exit velocity is 83.9 miles per hour. Only Rich Hill has allowed softer contact among the 153 pitchers who have allowed at least 200 balls in play.

Montero’s xBABIP is still .307, because he has been “barreled up” at a decent rate, ranking 150th out of 362 pitchers in terms of barrels per batted ball event, minimum 100 batted ball events, per Baseball Savant. (Barrels are classified as well struck balls, with an expected batting average of at least .500 and an expected slugging percentage of at least 1.500.)

Of course, .307 is way below .373, supporting the idea he has gotten somewhat unlucky. For season-long fantasy owners, Montero is still a tricky case due to his volatility, but given his pitcher friendly ballpark and weak division, he may warrant some spot starts down the stretch.