David Hernandez As A Relief Pitcher

Despite the quotes about Chris Tillman starting the year in the minors, I think the starting rotation for 2010 is largely set as Kevin Millwood, Jeremy Guthrie, Brad Bergesen, Brian Matusz, and Tillman. While there’s virtually no chance that the Orioles will be able to get through the entire season with only those five guys, that does leave a couple pitchers on the outside looking in for now. Neither Jason Berken nor David Hernandez have much to prove in the minors, but while Bekren’s role seems to be that of a 5th starter or swing-man, some people think Hernandez could really excel as a short-reliever. I thought I’d take a look at how that might work.

When Play A Hard Nine analyzed a potential change from the pen to the rotation for Cards pitcher Kyle McClellan, they used some research by Sean (CHONE) Smith that showed what happens to a pitchers stats when switching between the two roles. He found that a starter’s strike-out rate (in this case, K per AB) increased by 15%; his hit rate (non-HR hit per ball in play) fell by 4%; his home run rate fell by 13%; and his walk rate stayed about the same. Let’s apply that to the CHONE projections for Hernandez for 2010, which are for him as a starter:

CHONE SP 148 154 28 69 120 5.44
Adjusted RP 65 60 10 29 59 4.75

I just went with 65 innings out of the pen, though that wouldn’t actually change the relief FIP^. The drop of about 0.7 is close to the 0.8 difference that was found in The Book, which was based on the change in wOBA^ against (which might take into account the drop in hits allowed, whereas FIP does not). Those relief pitcher stats still aren’t that great, largely due to a high walk rate (4.1 BB/9) and a still high home run rate (from 1.7 HR/9 to 1.4 HR/9). Perhaps a look at some slightly more optimistic theoretical projections:

SP Projections 150 155 25 66 130 4.95
Adjusted RP 66 60 9 28 63 4.30

That’s a move from 5th starter to decent middle-reliever. Still nothing to get too excited about. Perhaps moving to the bullpen would allow Hernandez to face more right-handed batters, and he could leverage a potential platoon split. Let’s look at his numbers from the majors in 2009, courtesy of FanGraphs:

Split K/9 BB/9 K/BB HR/9 FIP xFIP^
vs L 5.2 4.7 1.1 2.4 6.92 5.92
vs R 7.2 3.3 2.2 2.5 6.21 5.18

He struck out right-handed batters at a much higher rate, and walked a good deal fewer, but they took him deep at just about the same rate as left-handed batters (and both sides had about a 14.8% HR/FB^ rate, so it’s not as if he got especially unlucky versus righties). The sample sizes here are obviously not that large though.

Was he that consistent in the minors last year (from MinorLeagueSplits.com)?

Split K/9 BB/9 K/BB HR/9 FIP HR/FB
vs L 12.1 2.2 5.5 0.8 2.55 9.4%
vs R 11.7 3.3 3.5 0.6 2.77 7.7%

These are even smaller samples (about 30 IP in each group). For 2008 it was:

Split K/9 BB/9 K/BB HR/9 FIP HR/FB
vs L 9.1 5.4 1.7 0.8 4.32 6.9%
vs R 11.6 3.8 3.1 0.6 2.96 5.2%

This looks a little more like Hernandez’s numbers in the majors in ’09, with the higher strike-out rate and lower walk rate. If his HR/FB rate was the same for both handed batters, then his home run rates (and FIPs) would have been closer as well.

And finally, for his career in the minors:

Split K/9 BB/9 K/BB HR/9 FIP HR/FB
vs L 9.8 4.3 2.3 0.7 3.87 6.7%
vs R 10.4 3.3 3.1 0.8 3.58 7.9%

There’s a split there, but it’s not huge. Maybe half a run, if you even out the HR/FB rates (his general flyball rates were almost exactly the same; around 44% for both). That makes the difference in his xFIPs from the majors in 2009 seem pretty reasonable.

Perhaps Hernandez’s repertoire might play better out of the pen though. Here are his platoon splits, with regards to contact rate. Overall, batters made contact with about 83% of the pitches the swung at against David (which is a little above the league average), and the total rates were about the same for right-handed batters and left-handed batters. There were some larger differences between his different pitches though.

Batter Overall Fastball Change-up Slider Curveball
RH 82% 87% 73% 68% 66%
LH 84% 85% 83% 81% 77%

For comparison, here’s the league totals for right-handed pitchers:

Batter Overall Fastball Change-up Slider Curveball
RH 79% 86% 73% 69% 70%
LH 82% 87% 74% 75% 72%

Hernandez was able to get swings and misses more often than average against right-handed batters on his breaking-pitches, with the fastball rate being a little higher* and the change-up rate being about the same. He wasn’t fooling lefties with the off-speed though. Hernandez actually faced more lefties (260) than righties (202) last year, so if going to the pen would allow him to flip those proportions and he scrapped the change-up in favor of more breaking-balls (split between the two types), his total contact rate would still only drop be less than one percent, to 82%.

* The samples are small and the differences aren’t huge, but I find it interesting that Hernandez had a small reverse platoon split on fastball contact rate.

The issue remains that Hernandez’s fastball is hittable to big leaguers. Maybe if he was moved to the pen and picked up a tick of velocity, he’d be able to blow it by hitters more. Here’s a graph of contact rate against the fastball at various speeds against right-handed batters (blue) and left-handed batters (red).

As you can see, lefties had more trouble putting wood on the ball as Hernandez threw it harder, but righties didn’t quite as much (the lower contact rate for slower pitches is likely due to a tiny sample size – not that the rest of these are particularly large). It’s not a huge difference, but more velocity still did make the fastball harder to hit for right-handed batters. Maybe more movement on the pitch would help as well.

The contact rate was pretty consistent for lefties – a little higher with more movement – but for righties it actually went up quite a bit as Hernandez got more break on the pitch. Perhaps that was because as David threw harder (which helped with some swings and misses), the ball moved less – his average movement for pitches in the low 90s was about half an inch more than for pitches in the mid 90s (with it appearing to be the case that velocity affected contact rate more than movement did). If moving to the pen would add, say, one mph to his fastball (with everything else staying the same), then the contact rate against the pitch would go down by a little over 1.5% – largely because of the improvement versus left-handed batters (since they fared worse versus faster pitches). Switch the proportion of batters faced (202 RHH, 260 LHH to 260 RHH, 202 LHH) and the drop is still 1%. Add 2 mph and it’s a 1.5 to 2% drop in contact rate on the fastball.

Combine the increase in velocity with dropping the change-up and facing more righties than lefties, and maybe Hernandez could increase his swing and miss rate by 2% overall – from 83% to 81%. That would still be worse than average. Given that strike-out rate is pretty strongly correlated with contact rate, as you can see (from qualified starters in 2009):

You could then expect David to up his from 6 K/9 to around 7 K/9. That’s a little bit better than the 15% increase from Sean’s study, but pretty close. (I should note that the projection for Hernandez has him improving to that 7 K/9 level as a starter already.) I guess that that might point towards the FIP conversion from starter to reliever at the top being a good indicator for what to expect if the move comes to pass; and all this was to say that there probably isn’t anything too special about Hernandez that would make the transition extra successful for him.

Overall, I think David Hernandez may end up in the bullpen just due to the number of young pitchers the Orioles have in their system. I wouldn’t expect that to be the cure from some of his issues as a pitcher though. Hernandez just needs to pitch better, no matter what role he’s in.