Is Jeremy Hellickson breaking the mold?

It’s no secret that Andrew Friedman, Gerry Hunsicker, and the rest of the Rays’ front office is developing one of the best young starting rotations in baseball down in Tampa. Led by top-pick David Price, 26, a Cy Young candidate two years back with a 5-1 record and a 2.35 ERA on the season, Tampa’s rotation has been the league’s second best this year. The Rays are 19-9 on the season and their starting rotation 16-6 in those games with a 3.51 ERA.

Price has a 2.49 FIP, a career best. James Shields has a 2.94 xFIP, also a career best. Jeff Niemann is one of the best back of the rotation starters in baseball – he has a 3.61 xFIP this season, a 3.73 last year – and while Matt Moore has struggled with a 5.71 ERA on the season, no one is seriously worried about baseball’s top prospect. The Rays’ are taking the AL East by storm, and their rotation is a big part of that success.

Then there’s Jeremy Hellickson, the enigma, Tampa’s third starter and one of their best over the past two seasons.

At 25-years-old, Hellickson is in his second full season as a Major Leaguer. 45 games, 39 starts, and 261.1 innings into his career Hellickson has a 3.00 ERA, a Rookie of the Year award under his belt and twice as many wins as losses. Nearly everyone expected Hellickson would succeed. Baseball America rated Hellickson the sixth best prospect in baseball headed into last season, second among pitching prospects. In fact most expected Hellickson to succeed quickly. The polished, diminutive right hander dominated at each level of the Rays’ minor league system, including a 2.45 ERA in AAA in 2010. Hellickson has pitched exceptionally, but it’s how Hellickson is doing what he’s doing that is so unusual.

In 189 innings as a rookie Hellickson struck out 117 batters and walked 72 – a ratio of just 1.62, not an acceptable level for all but the most unhittable pitchers. Of course an extreme groundball pitcher might get away with a strikeout to walk ratio south of two, but Hellickson’s groundball rate last season was just 35%, far below league average. In those 189 innings he allowed 21 home runs, a rate of one home run for every nine innings. Not good, but it could have been worse. His HR/FB rate was just 8.1%, below league average, and only 18% of baserunners he allowed scored.

Despite posting an xFIP of 4.77, Hellickson allowed just 62 earned runs all season. He went 13-10 with a 2.95 ERA. The baseball media ignored his incredible “luck,” awarding Hellickson the top rookie honors in the American League with 17 of 28 votes. Few complained. The statistical community was largely skeptical, though. Hellickson’s .223 BABIP last season was simply unsustainable. His defense independent pitching metrics told the story of a very lucky and very troubled pitcher who couldn’t finish off batters, throw strikes, or keep the ball in the park. Yes, Hellickson’s was a top prospect and growth as to be expected, but his minor league strikeout to walk ratio of nearly five was gone. Skill growth would almost certainly be undercut by luck regression. Hellickson was headed for a rough 2012 season.

And yet… so far so good for Hellickson. It’s early, but in six starts the Rays’ right hander is 3-0 with a 2.75 ERA. His FIP sits even higher than in 2011, at 5.09, and his xFIP is also higher at 4.86. His fWAR is actually negative. The dichotomy between skill and luck seems to be growing, not diminishing, with time. Greater variance is normal over small sample sizes. The more important point is not Hellickson’s last six starts, but his first 261.1 innings. With more than 70 innings 2010 and 2012, and a full season in 2011, Hellickson’s resume isn’t so short.

As Hellickson’s big league career approaches a significant sample size, the difference between his ERA and xFIP over the course of this big league stint – right now it stands at 1.61 runs – must be taken more and more seriously. Even if Hellickson’s ERA was perfectly predicted by his xFIP over the next year and a half, it would stand 80 points below his xFIP through 500 career innings, a very significant sample size. His BABIP would still sit 30 or so points below league average. The Rays defense certainly helps, but the rest of the Rays’ roster has pitched to a .280 BABIP over the past two-plus seasons. Hellickson is at .231.

The point is that while Hellickson could very well alter his approach, a similar approach would likely yield at least far above average results over a large sample size. This does not mean we should assume Hellickson will not face regression. We’re simply approaching a point at which his success means something, at which the magnitude of his success must be considered statistically significant (if not completely indicative of future similar performance). But is there precedent?

Since 2000, a grand total of 189 starting pitchers have thrown at least 250 innings before their fourth season in the big leagues, a list that includes most successful pitchers in Major League Baseball today. Hellickson’s made 39 starts over that period – near the bottom of the list – but 25 more starts this season would place him firmly in the middle of the pack in mound experience through his early big league seasons. Hellickson’s career BABIP of .231 is the lowest on the list by a whopping 17 points. But assuming some regression – say Hellickson pitches to a Rays’ average .280 BABIP the rest of the year – his BABIP could settle into the .250 range, still basically undheardof. In fact nothing short of a .315-.330 BABIP would place Hellickson in charted territory, around .265. That’s unlikely, but it’s possible.

A dozen other pitchers have pitched to a sub-.270 BABIP through three seasons. Freddy Garcia, Barry Zito, Damian Moss, Chris Young, Chuck James, Matt Cain, Aramando Galarraga, Josh Tomlin, AJ Burnett, David Price, Mark Buehrle, and Jose Contreras. The list is quite diverse. Matt Cain just signed a contract worth in excess of $100 million. Chuck James pitched himself out of baseball. But how well have these twelve maintained their BABIP? Garcia, Burnett, Buerhle, and Contreras have BABIPs now close to league averages. But Garcia had only made 54 starts, Burnett only 40, two of the lowest three figures on the list. Buehrle and Contreras had the highest BABIPs on the list at .269, nearly 40 points above Hellickson. Cain, Zito, and Young have largely maintained their BABIP success. Price is still young, same with Tomlin, and James, Moss, and Galarraga left the big leagues shortly after their third seasons, but all five maintain their BABIPs to this day. This suggests that while regression is possible after 40 or 50 starts, so too is a maintained ability to prevent hits on balls in play.

What does this all mean? First, Hellickson will very likely finish the season with a BABIP in the .240s or low-.250s, even with significant regression. This is not only impressive or rare, it’s effectively unheardof over the past decade. It’s uncharted territory. We don’t have any good comparisons and we don’t really know whether Hellickson can maintain a BABIP in the .230s. No one in recent history has done so, but Hellickson is approaching a large sample size and stands as a remarkable outlier today. Second, there is precedent for maintained low-BABIP among extreme outliers. We’d usually expect regression but a .260-.270 BABIP over 60 or so starts is significant. Whether a .230 or .240 or .250 BABIP is possible is a question of degree. More generally, though, low BABIPs can be maintained (see Matt Cain, Barry Zito, etc.). It’s simply a rare phenomenon. There is still reason to expect regression – we have a small data set and only a few real successes among these low-BABIP pitchers – but Hellickson’s low-BABIP and low-ERA can’t be ignored. If he can’t keep his BABIP in the low .250s or below, his approach will need change. If he can, though, Tampa is going nowhere but up. Price, Moore, and Hellickson could all be top of the rotation starting pitchers for years to come.