Phillies Grab Whatever Roberto Hernandez Is

I remember the way things used to be. Used to be, writing about baseball analysis was pretty easy. Inflated or deflated BABIP over here. High groundball rate over there. This guy has a big difference between his ERA and his xFIP. That guy is miscast as a role player. That’s not the way things are anymore. Many of the principles were fine, and you still see a lot of the same ideas, but over time things have grown more complex, more nuanced. We’re moving beyond pointing out weird things, and we’re moving toward trying to explain weird things. It’s all in the name of identifying what is and isn’t sustainable. All of us want to be fortune tellers.

The Phillies signed Roberto Hernandez today, as a starting pitcher. He’s going to get a base salary around $4.5 million, and with incentives he can top out around $6 million. If writing were the same as it was a few years ago, I could just write a few paragraphs about how Hernandez put up a 4.89 ERA and a 3.60 xFIP. On that basis alone, hey, look, bargain! But because of what writing and research have become, now you also get that intro paragraph.

What’s known is that, in his return to the majors, Hernandez put up some neat peripherals in Tampa Bay. What’s also known is that he allowed too many runs, because he allowed too many home runs. Just about the entire story here is trying to figure out whether or not Hernandez is severely homer-prone. And, I don’t know. I’ll just put that out there now. I don’t know what his dinger rate is going to look like in 2014.

But, for the record, Tim Lincecum is coming off a 124 ERA- and a 94 xFIP-, and he signed for two years and $35 million. Dan Haren is coming off a 125 ERA- and a 97 xFIP-, and he signed for one year and $10 million. Hernandez, as a starter, is coming off a 131 ERA- and an 89 xFIP-, and he signed for one year and no more than $6 million. The deal looks at least reasonable, and because it’s a one-year commitment it’s a virtually harmless roll of the dice. Something of a sabermetric roll of the dice, being attempted by the Phillies. Ruben Amaro was just talking about pitcher wins the other day. He signed a guy with a miserable ERA who also tied for 27th in baseball among starters in park-adjusted xFIP. Amaro didn’t sign Hernandez because of his xFIP, but it’s funny when certain organizations end up with surprising types of players.

Here’s something to know about Hernandez. Since 2002 — as far back as the data goes — 111 starters have thrown at least 1,000 innings. Hernandez has put up an FIP- ten points higher than his xFIP-, which is the second-biggest positive gap in the pool, below only Brett Myers. One of his selling points is that he can keep the ball on the ground, but because of the homers, Hernandez’s “effective” grounder rate is quite a bit lower. It seems to be Hernandez is unusually homer-prone, and last year was a major flare-up.

Again, since 2002, there have been 1,580 individual starting-pitcher seasons of at least 100 innings. Hernandez’s 2013 owns the second-highest rate of homers per fly ball, above 21%. Basically, twice the league average. Mark Trumbo just put up a rate of about 21%. The question is, all right, so what does that mean going forward? We can look at the four worst rates from the season before. Ervin Santana regressed to something much closer to average. Tommy Hunter regressed to the average, but as a reliever. Jason Marquis kept allowing a ton of homers. Henderson Alvarez rebounded to allow the lowest rate of homers per fly ball in all of baseball.

So, there’s your answer. Hernandez, this coming season, will not regress at all, or he’ll move to the complete opposite extreme, or he’ll regress to the average, or he’ll regress to the average upon changing jobs. That covers literally every base. It feels like we’ve learned a lot, but really, we haven’t learned anything. Not as far as next season is concerned.

The safest assumption is that Hernandez will allow dingers, but fewer of them. He allowed 24 last season, throwing about 2,400 pitches. In other words, 1% of Hernandez’s pitches wound up as a dinger, and you can see why this is a stat with so much volatility. We’re talking about roughly one pitch per game, and even good pitchers are frequently making mistakes in terms of location. The average mistake rate is way higher than 1%, and when there’s a mistake the opposing hitter still has to do what he can do to cash in. Sometimes the pitches get punished; sometimes they get swung through or popped up. Some missed timing here, some fortuitous swings or wind gusts there, and Hernandez can look a lot better. It’s clear that xFIP leaves a lot out. It’s clear that xFIP is still extremely useful, and worth keeping in mind. Better to have a guy coming off a high ERA and a low xFIP than a guy coming off a low ERA and a high xFIP. For 2014 I’d rather have Roberto Hernandez than Hector Santiago.

It’ll be interesting to see if Hernandez holds on to some changes with the Phillies. The Rays had him throw a low more changeups to same-handed hitters, and a lot more sliders to opposite-handed hitters, both of which defy convention. Against righties, Hernandez posted his lowest career xFIP. Against lefties, Hernandez blew away his previous career-high strikeout rate. Lefties also torched Hernandez for 17 of his 24 homers allowed, so that could be a real concern, but as we’ve discussed, we don’t know how much of one. We just don’t. Probably can’t, at least not from here.

So the Phillies will take the chance, for a year and a little money. They’ll take the chance any sabermetric writer would’ve recommended a few years ago. Absolutely, it could work out real well. Absolutely, Hernandez could keep allowing homers, because he has a problem we can’t quite put our fingers on. For a team in the Phillies’ position, better to take a cheap chance than a more expensive one. Maybe Hernandez won’t be good, but then the Phillies probably won’t be good, either. As free agents go, the right thing to target is affordable volatility.



Print This Post



Jeff made Lookout Landing a thing, but he does not still write there about the Mariners. He does write here, sometimes about the Mariners, but usually not.


Sort by:   newest | oldest | most voted
Yirmiyahu
Member
2 years 8 months ago

I refuse to believe that the Phillies made a smart, sabermetrically-inclined move to find a bargain. Therefore, Hernandez’s high HR/FB% is going to be sustained and he’ll put up a 5ish ERA.

Ruki Motomiya
Guest
Ruki Motomiya
2 years 8 months ago

Hernandez worries me because of his career 12.5% HR/FB (IIRC, league average is 10%). He’s always put up a good deal of dingers. Combine that with declining stuff and he is a very worrying pitcher. But for 4.5 mil, it is worth the gamble, especially for a team like the Phillies.

Pops
Guest
Pops
2 years 8 months ago

The gamble last year in the form of John Lannan didn’t work out… at all, for the Phillies. I’m skeptical about this signing as well… even with Hernandez’s ground ball tendancies, in that small ballpark.

AMB
Guest
AMB
2 years 8 months ago

While CBP ranked 1st last year in HR Park Factor data it ranked between 10th-16th from 2008-2012.

While certainly a park that favors home runs it is far more neutral than people think especially since they moved the left field wall out after the park had been open for a few years.

Drew C
Guest
Drew C
2 years 8 months ago

I think the Lannan deal result had a lot to do with his health.

Pale Hose
Guest
Pale Hose
2 years 8 months ago

I remember the way things used to be. Used to be writing about baseball was pretty easy. The players used to have two names, one first and one last. Now, some players have two names.

Vegemitch
Member
Vegemitch
2 years 8 months ago

Pepperidge Farm remembers

jruby
Member
Member
jruby
2 years 8 months ago

I’ve always been somewhat fond of Hernandez, but even if I hadn’t, as a Phillies fan, I like this because it keeps them from doing something else ridiculous with this money, like putting it towards a third of a guy the front office thinks will take them from 85 to 90 wins and the rest of us realize will take them from 73 to 75 wins and cost them 10 wins worth of prospects/silly payments over the next five years.

Not that I’m, ahem, disillusioned or anything…

Brandon Firstname
Member
2 years 8 months ago

The reason why HR/FB is removed from xFIP is that it just isn’t reliable because homers are a rare occurrence and therefore have huge error bars in small samples.

When your sample size gets bigger though (Hernandez’s is getting bigger), the predictive nature of your HR/FB goes with it. The most reasonable expectation is to expect Hernandez to regress some of the way, but not all of the way.

Then again, that last statement goes for pretty much everyone. And the bigger the sample, the less you regress. You obviously shouldn’t trust the HR/FB of some rookie.

Brad Johnson
Member
Member
2 years 8 months ago

Right, so you take Hernandez’s career HR/FB rate, the average HR/FB rate, and assign weighting. That’s standard procedure. So just to put some numbers on it, if his average is 12.5, league average is 10.0, and we assign a 50% weight to each, then we expect a 11.25 HR/FB ratio. Which happens to be a lot lower than 20.

LoveTheJays
Guest
LoveTheJays
2 years 8 months ago

Can anyone point me to a resource that shows all the formulas used to calculate sabermatics metrics?

Have a statistical background and have been a huge ball fan my whole life, just getting interest in a deep dive into this sabermatics stuff now.

Seems like lots of good talk about stuff on this site so:

A couple things I never really understood fully / have questioned are BABIP and xFIP. They seem incomplete and the reverence given to them seems to potentially introduce false analysis.

For BABIP is the fact that when you hit the ball harder you are more likely to get a hit not factored in? Maybe a guys BABIP is down this year not because he is unlucky but because he isn’t hitting well and not hitting the ball hard and so doesnt get hits. Most likely it is some combination of the two while sometimes a guy is simply unlucky and sometimes he simply didnt hit the ball hard all year. Is there a how hard was the ball hit factor to weight into this and along those lines weight in the distributions of the trajectory the ball was hit on as well?

For xFIP is how the ball is thrown not a significant variable for how many homers will be hit on fly balls? Maybe a pitcher has a higher than average homer to fly ball ratio one year because he kept putting straight fastballs right down the hump and the batters were mashing the homer balls while some other guy was throwing late breaking stuff that causes the batter to miss and hit a shallow fly ball.

I see the value in these things as indicators but they seem like indicators warranting more investigation rather than truth and there may be potential to make the metrics more accurately predictive by weighting in other variables or perhaps combining with other metrics.

It gives me a bit of a chuckle reading about small sample sizes being entire or even multiple seasons and therefore performance in a particular entire season can be written off due to randomness. Someone saying oh he had terrible defensive stats last year but that is a small sample size. That is many events for good statistical analysis. Metrics need to be improved to be a better predictor if it requires 3 years of data to have confidence in a result – not suggesting this is easy to generate solid concise metrics that are better with smaller samples by any stretch. And failing the metrics some comprehensive and common sense analysis comes into play. Seems often that entire years are written off as outliners and statistical randomness when it seems clear there are variables from year to year not necessarily given their due course in the generation of the metrics.

coldseat
Guest
coldseat
2 years 8 months ago

Increasing sample size and utilizing your scouts help fill in the gaps you mentioned, but as this article emphasizes, no one knows &the rest are just playing the “I’m smarter than you” game like its 2008 on Wall Street.

Tyler
Guest
Tyler
2 years 8 months ago

http://www.fangraphs.com/library/

You can find definitions/formulas here by clicking on any term you would like to have defined.

coreyerb
Guest
2 years 8 months ago

BABIP, at least, often is misused in analysis as you mentioned. xBABIP is an attempt to weight for batted-ball types like you suggest. It’s exciting because it comes closer to offering predictive value to an otherwise descriptive stat (again, you’re right that if a batter has a BABIP of .190 isn’t necessarily due for regression toward the mean – cf. Scott Hairston this year: .185 BABIP but that wasn’t far from his expected BABIP since he hit fewer line drives and more weak fly balls than he ever has, so he should have a low BABIP). Check out this article:

http://www.fangraphs.com/fantasy/eating-crow-xbabip-and-the-shift/

I’d like to see xBABIP used more. If there’s been an article with an updated formula with 2013 weights, I must’ve missed it.

Brad Johnson
Member
Member
2 years 8 months ago

Jeff’s intro paragraph gets at some of the concerns you’ve listed here. For BABIP, if I see that Freddie Freeman has a BABIP that smells high, I go into his batted ball outcomes, take a look at his average fly ball distance, and figure out if defenses are/aren’t shifting him in some obvious way (via spray charts and video). And I’ll also use xBABIP like Corey mentioned. In the past, I just would have said, “high BABIP will regress, maybe not all the way to league average, but close.”

For xFIP, I do a similar analysis. I see the discrepancy. Then as Jeff did with this article, I check in on some related stats. Did his velocity drop, does a heat map of his ISO by location reveal any problem areas, is a specific pitch to blame, is there a reason that pitch might be to blame? Etc etc. With Lincecum, I found that his change-up was mashed in 2012 and his fastball was mashed in 2013. There was no evidence to suggest either pitch should be expected to be mashed going forward, which led me to expect a slight decline in home run rate.

Jason Collette
Member
Member
2 years 8 months ago

Simply put, whatever he elevates within the zone gets destroyed. When he’s on top of his sinker, it’s fun to watch him pitch for short spans. His changeup is a good pitch because he has the exact same arm action for his sinker so batters can be fooled with it.

That said, it didn’t work with The Trop & a very good infield/positioning behind him so I’m doubtful it is going to work in a worse situation. He’d love to eat innings and whatever else you feed him.

John C
Guest
John C
2 years 8 months ago

The Phillies are just hoping he can channel his inner Fausto Carmona.

Bengals Fan
Guest
Bengals Fan
2 years 8 months ago

Ruben Amaro and Roberto Hernandez star in the newest play to hit broadway…”Fausto”.

raltongo
Guest
raltongo
2 years 8 months ago

per trade rumors:

“Phillies have an agreement with an unnamed starter…more to come…”

“It has been determined that his name is Roberto Hernandez”

haha

Metsox
Guest
Metsox
2 years 8 months ago

At least he gets to pitch on the road in Miami, Flushing, Atlanta, and Washington….

Josh Z
Guest
Josh Z
2 years 8 months ago

What is his HR/fly ball ratio when midges are present?

Object moved to here
Guest
Object moved to here
2 years 8 months ago

Robusto Hermona

Steve-o
Guest
Steve-o
2 years 8 months ago

10 paragraphs to say ” I dunno”

Ruki Motomiya
Guest
Ruki Motomiya
2 years 8 months ago

But who does the methodical pace benefit?

Antonio Bananas
Guest
Antonio Bananas
2 years 8 months ago

Can I get a link to RAJ talking about pitcher wins? For the lulz

RJ3
Guest
RJ3
2 years 8 months ago

Amaro said he has not engaged Kendrick’s party in talks for a multi-year extension.

“It’s amazing,” Amaro said. “We talked about it this morning. He has 64 wins. Do you know how many wins Garza has? Sixty-seven.”

Amaro added: “He’s benefited from some run support and defense. But wins are wins.”

Read more at http://www.philly.com/philly/blogs/sports/phillies/Phillies-continue-to-mine-mid-level-pitching-market.html#FuoacCD3AuzSmW8o.99

NEPP
Guest
NEPP
2 years 8 months ago

Sometimes its hard to believe someone could have that simplistic a grasp on baseball and be a GM of a billion dollar franchise but here we are…referencing Wins as if it means anything at all when comparing two pitchers.

vivalajeter
Guest
vivalajeter
2 years 8 months ago

RJ3, I initially thought it was a lame attempt at humor. Then I saw that you included a link, and I couldn’t believe he actually said that!

On an unrelated note, I was also surprised to see that Amaro was white. For whatever reason, I always thought he was black.

Zen Madman
Guest
2 years 8 months ago

Steamer projects Fausto at 2.9 fWAR on the strength of a 3.50 FIP and .62 HR/9. Oliver projects him at -.4 fWAR with a 5.00 FIP and 1.48 HR/9. So taken together, the projection systems also answer the question with a big shrug of the shoulders and a “Maybe this?” response.

wpDiscuz