Trading Jeremy Hellickson

Coming into this offseason, the Tampa Bay Rays seemed most likely to move veteran starter James Shields. Yet according to Jon Heyman, a different pitcher, Jeremy Hellickson, has been the Rays’ most sought-after starter during early trade talks.

In 402.1 career innings, Hellickson has posted an incredible career ERA of just 3.06. The 25-year-old right-hander is eligible for the league minimum salary next season and is under club control through the 2016 season.

Why would Tampa Bay ever trade him?

Well, we should all know by now that earned run average is not the best measure of actual pitcher performance. There are factors outside a pitcher’s control that affect ERA, namely defense and random variation (batting average on balls in play, BABIP).

Here are where Hellickson’s BABIPs in 2011 and 2012 ranked among other qualified starters:
{exp:list_maker}2011: .223 (the lowest)
2012: .261 (t-6th lowest) {/exp:list_maker}If people versed in sabermetrics were to look at each of the seasons independently, they would almost inevitably conclude that Hellickson had been lucky. This point is hammered home when we consider a statistic that excludes all balls in play, fielding independent pitching (FIP).

Hellickson’s FIP in 2011 was 4.44 and that number rose in 2012, to 4.60.

During the past two seasons, Hellickson’s park-adjusted ERA was between 20 and 24 percent lower than average, while his park-adjusted FIP sat 15 to 17 percent higher than the major league average.

In both seasons, Hellickson had a larger gap between his ERA and FIP than any other qualified starter in baseball: His ERA was 1.49 points lower in 2011 and 1.50 lower, in 2012.

Has Hellickson just been extremely lucky two seasons in a row or has he found the magical BABIP elixir?

I feel like this question has been beaten to death, with strong opinions and arguments on both sides, but I’d like to re-open the discussion with these rumors swirling around the pitcher.

I don’t know if Hellickson has found a way to out-pitch his peripherals by somehow yielding weak contact, in turn resulting in a lower than average BABIP and lower ERA. All I can give you are the facts.

ERA estimators

Defensive independent pitching metrics (DIPS) have been shown to do a better job at estimating future ERA (in the short-medium term) than actual ERA.

DIPS measures are primarily based around the two true outcomes for pitchers, strikeouts and walks. As with FIP, there is a large discrepancy between Hellickson’s ERA and DIPS measures, mainly because there is a large discrepancy between Hellickson’s ERA and what his strikeouts/walks would indicate.

Hellickson’s career strikeout-to-walk ratio is 1.97. In 2012, his K/BB was slightly higher (2.10) than that number, but it still ranked as the 16th-worst among 88 qualified starters. If you trust certain data, that doesn’t bode well for Hellickson’s future.

Should we trust DIPS or ERA to project Hellickson’s ERA going forward?

Based on solely on his 2012 FIP, we’d expect Hellickson’s 2013 ERA to be 4.60, a significant rise from last season. Using a version of FIP that is re-weighted to be more predictive, rather than descriptive, pFIP, we get a slightly more favorable ERA projection of 4.50.

A starter who throws 175+ innings (as Hellickson did last season) with a 3.10 ERA is substantially more valuable to a major league team than a starter who throws the same number of innings but has an earned run average of 4.50.

Last season, Hellickson’s 3.10 ERA ranked as the 16th-best among pitchers who threw at least 120 innings, ahead of names like Cliff Lee, Stephen Strasburg and Madison Bumgarner.

Using pFIP as the projection tool for that same set of pitchers, Hellickson’s projected 4.50 ERA would rank as the 102nd-best (out of 121 pitchers) for next season, behind Bruce Chen.

Tampa Bay’s defense and ballpark

Strong team defense is supposed to have a negative effect on BABIP. A better defense usually results in a lower-than-average team BABIP. According to research done by Matt Swartz, defense has a 13 percent effect on BABIP.

In 2011 and 2012, Tampa was one of the best teams at converting balls in play into outs. The Rays’ team defensive efficiency (DER) was the best (or tied for the best) in baseball in both seasons.

DER is essentially (1- BABIP); thus, maybe we should assume that Hellickson’s abnormally low BABIPs are at least somewhat due to the fact that Tampa has been so good defensively over the last two seasons.

I would not say that assumption is set in stone though, as the correlation between an individual pitcher’s BABIP and BABIP for the rest of the team isn’t as high as one would expect. There’s actually a chance that Hellickson has had close to the same negative effect on Tampa’s team BABIP as the team’s defense has had on his.

Tampa Bay has played home games in Tropicana Field since the team’s inception in 1998.

We have ESPN Park Factors dating back to 2001. Tropicana has consistently suppressed both hits (importantly balls in play) and runs in almost every one of those seasons. Consider the scenario where Hellickson gets traded and moves away from the Trop: Maybe his BABIP, and more importantly ERA, will rise. The interesting question is by how much? Or even more importantly…

Should a team that plays in a hitter’s park with a bad defense risk trading away quality talent for Hellickson’s services?

Maybe the first step to answering that question would be to look at Hellickson’s career home/road splits:
{exp:list_maker}Home: Career ERA/BABIP/FIP: 2.84 / .245 / 4.03
Road: Career ERA/BABIP/FIP: 3.22 / .243 / 4.95
{/exp:list_maker}These splits complicate the mystery of Hellickson even further. His career home ERA is better than on the road, but he’s pitched better at home,in terms of FIP, and his career BABIP is actually better away from the Trop.

Are these splits reason to believe that Hellickson has magical BABIP-suppressing powers, or are we simply being confounded by sample size?

Sample size

I feel like I’ve typed this sentence a thousand times, but BABIP for pitchers is affected by random variation more than any other relevant pitching metric.

Derek Carty showed that it takes more than 3,700 balls in play before a pitcher’s BABIP “stabilizes.” Carty uses the word “stabilize” to describe the point at which the correlation between past BABIP and future reaches 0.5; thus, when projecting future BABIP for that pitcher we can use half of his past BABIP and regress the other half back toward the league average.

That’s a lot of balls in play. It’s eight seasons worth of balls in play for a starter. Year-to-year BABIP fluctuates all over the place.

Take for instance 1999, when Pedro Martinez‘s BABIP was .323> The next season it dropped almost 90 points, all the way to .236. Heck, Hellickson’s BABIP was 38 points higher in 2012 than it was in 2011. Who knows if next season it will rise another ~40 points, or instead drop down below where it ended in 2012.

Thus far, Hellickson has induced only 1,190 balls in play. It’s tough to come up with anything conclusive based on a sample size that is only one-third of the way toward revealing true talent level.

Conclusion

So far, I’ve only discussed ERA and defense-independent metrics, and which is a better indicator of Hellickson’s future. But I’d like to look briefly at a different metric than FIP or ERA to describe Hellickson’s performances over the past two seasons.

Baseball Prospectus publishes a statistic known as Fair Run Average (or FRA). FRA does a good job of describing for us how many runs a pitcher deserved to give up.

Hellickson’s FRA in 2011 was 4.93, well below average and well over his actual runs allowed per nine innings (3.05). Again last season, Hellickson’s FRA was bad, 4.54, versus the 3.46 that he actually allowed. These numbers would indicate that Hellickson hasn’t been nearly as good as his runs allowed per nine innings would indicate.

In fact, Baseball Prospectus uses FRA for its pitching Wins Above Replacement Player (WARP) calculations.

Hellickson’s career WARP is 0.8, which would indicate that despite compiling more than 400 career innings, Hellickson has been a replacement-level pitcher (or even worse if you consider playing time). Using pFIP, I found an RA/9 projection of 4.83 for Hellickson next season, which jibes much better with his career FRA (4.67) than it does with his career RA/9 (3.27).

Is Hellickson a replacement-level pitcher who is due for serious regression, or is he a top-20 pitcher who can be had for pennies on the dollar compared to free agents like Zack Greinke?

I honestly don’t have an answer. But if a team decides it wants to get into deep trade talks with the Rays about Hellickson’s services, it should be aware of the facts that I’ve laid out.

Hellickson’s been a fascinatingly successful pitcher, and could continue to be, but his future is still uncertain.

References & Resources
All statistics come courtesy of Baseball Prospectus and FanGraphs

Print Friendly
 Share on Facebook0Tweet about this on Twitter0Share on Google+0Share on Reddit0Email this to someone
« Previous: BOB:  New free agent rules primer
Next: Playing at the extremes »

Comments

  1. Jon Roegele said...

    Hellickson’s such an interesting topic, Glenn.

    We discussed this briefly on Twitter the other day, but I thought I’d bring it up here to get TB fans’ takes.

    As you and Brian pointed out, it’s hard to argue with leaving well enough alone when his actual results have been so good. He works out of the zone a LOT, so ends up walking quite a few guys but is likely generating weak contact by doing so as well.

    The thing that I find interesting is that in his excellent 2010 (albeit SSS in the major leagues), he worked far more in the zone, had a higher K%, lower BB%, better FIP….but worse ERA.

    I was just surprised that he seems to have changed from that philosophy, when it was actually working well. But since his ERA has improved, and it’s the Rays….well they know what they’re doing.

  2. Glenn DuPaul said...

    @Jon

    I agree that it is strange that his best K/BB season was his worst ERA year, SSS aside.  I’d love to know if the Rays and him discussed a change in philosophy (working off the plate to induce weak contact), but that’s not knowledge we really have.

    It’s a thought I’ve had before though on the opposite side of the spectrum, with Zack Greinke.  I wondered in a BtB piece from a long while back, of whether or not the fact that Greinke is concerned with keeping HRs low, BBs low and Ks high, thus staying more around the plate, leads to harder contact.

    The post has a lot of issues, and I’ve learned a lot since, but here’s the link ( I still think it’s an interesting idea): http://www.beyondtheboxscore.com/2012/4/26/2972729/zack-greinkes-fip-complex

  3. Jon Roegele said...

    Glenn, yes I really find that to be an interesting idea.

    I keep thinking that batted ball angle/velocity data would be such a rich set of data for many things, including BABIP research. I would imagine it would provide so much more context if that was known. And I don’t doubt that the Rays have this sort of data, and likely understand weak contact and how to generate it, etc.

  4. SeattleSteve said...

    I think the rays would be wise to trade hellickson with his value possibly at a high. They match up really well with Kansas City. Hellickson and Jeff nieman for butler, Myers and a B type prospect would be awesome. Should be an interesting offseason for the rays.

  5. Hank Sager said...

    I haven’t done the digging to look at it, but I wonder what if any change the new turf in the Trop has had on their defensive performance analytics.

    This wouldn’t impact Hellickson’s numbers particularly as he’s more a flyball pitcher, but it seems this may have had an effect on how true GB’s play. It seems to me – though eyeball evidence is suspect I admit – that that surface results in a slowing of GB pace and bigger hops.

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>

Current day month ye@r *