This started as a comment, but given that I get paid a whole cabbage for every post and only a leaf for a comment, here we are.
Hopefully needless to say, I don’t have a horse in this race. While I really like Barton the player, he has to get some negative points for thinking that God appeared in the body of Eric Chavez’ kid brother delivering the message to bunt more. Either way, I don’t care about these guys, I just like the numbers.
Who is more likely to succeed against a left-handed pitcher?
To make a somewhat educated guess, I looked at two things. One, what their handedness split looks so far and two, how good of a hitter is each one projected to be this year. To start with the latter, this is what a consensus wOBA projection for 2014 is (ZiPS, Oliver, Steamer):
Daric Barton – .317
Alberto Callaspo – .310
So far, both have hit better against LHP, but there is a big difference – Barton is a left-handed hitter himself, while Callaspo switch hits. We have much clearer idea of what to expect from a single-handed MLB hitter, than we do for switch hitters, so let’s start with Barton.
Left-handers have larger splits than right-handers. The reason being, as I wrote here:
As for the fact that the left-handers constantly have higher splits than right-handers, the explanation is rather simple. There are two truths about hitting in baseball. One, it is easier to hit opposite-hand pitching than same-hand pitching. And two, it is easier to hit right-handed pitching than left-handed pitching.
As with anything else, the group data can help us estimate the individual’s true talent level. In case of Barton, we know the following of his MLB career:
- He had 562 Plate appearances against LHP
- He had 1459 Plate appearances against RHP
- His wOBA against LHP is .365
- His wOBA against RHP is .316
That is a huge 13% reverse split. Â Now, in case of Daric Barton we already have a reason to doubt that this is a very good representation of his true level talent, because we can check his Minor League stats*. But, let’s imagine for a second we had nothing else to go on, but these numbers above and the data on generic performance of left-handed batters in MLB.
In the article I linked to, you can check why we regress with exactly the number we regress with. Here, let’s just say that Daric Barton had a harmonic mean of 811 plate appearances in the Majors and that his observed level of talent (-13%) has to be regressed with 2261 plate appearances of average league (LHB split) production, which is 9.8%. When we do that, we come to the conclusion that our best estimated guess going forward is that Barton’s true talent level is:
811*(-13%)+2261*9.8%)/(811+2261) = 3.7%
So, we expect him to be a better hitter against RHP than LHP, albeit significantly less so than an average MLB LHB.
* In 3000+ PA in the Minors, Barton was better against LHP as well, albeit only marginally so (1.9%). If we include this data in our calculations, he is still expected to hit better against RHP, with the difference dropping to 2%
So, what is he expected to hit in 2014? His overall wOBA is projected to be .317 and such a projection is based on the same mix of pitchers faced as so far. Out of that we can easily reverse engineer that Steazipsoliver expects him to hit:
- wOBA .320 against RHP (.319 if you include the MiLB data)
- wOBA .309 against LHP (.312Â if you include the MiLB data)
With Callaspo we have a more difficult case, what with him being a switch hitter. Switch hitters have pronounced splits, just like the hitters always batting from the same side, but sometimes it is hard to say which side is the better side, when we don’t have enough data. In other words, we know we have to regress them, we have a pretty good idea by how much, but sometimes we are not sure in which direction. Also, contrary to the LHB or RHB, we are talking about two different physical skill-sets, not unlike forehand and backhand by tennis players.
As far as switch-hitters go, Callaspo is one of the easier ones, because he has more than 3000 PA in the MLB, another 3000+ in the MiLB and he has constantly shown pretty much the same split, 8-9% favoring the matchups against the LHP. So, to sum up Callaspo:
- He had 925 Plate appearances against LHP
- He had 2310 Plate appearances against RHP
- His wOBA against LHP is .337
- His wOBA against RHP is .309
Regardless whether we regress him as an average switch hitter (with 1620 PA of 6.8% split) or as a right-handed batter (1669 PA of 6.1% split) and regardless whether we use his MiLB data or not his calculated true talent level split comes out at pretty much 7% better against LHP.
Using same reverse engineering on the projection system data, our best guess what to expect Callaspo to hit in 2014 is:
- wOBA .303 against RHP
- wOBA .327 against LHP
Now, this is not exact science, but it is the most educated guess we can make going forward. Callaspo should hit for about 15-20 points of wOBA higher than Barton against LHP. Over a course of estimated 200 PA against LHP for A’s 1B, that is just about three runs advantage. Assuming that Barton plays defense as he did so far, and assuming that UZR gives a somewhat accurate picture, Barton is supposed to save about three runs playing first base in the games where the A’s face LHP. So starting Callaspo over him should make sense if the A’s assume that he can play as an average fielding first baseman.
For the TL;DR crowd:
- Barton’s observed split results in the MLB are misleading and not in line with his MiLB results
- After properly regressing, we should expect Barton to be marginally better hitter against RHP
- After regressing both players and reverse engineering their composite projection wOBA, we should expect Barton to have  wOBA of ~ .310 against LHP, Callaspo to have ~ .327
- Over ~200 PA that A’s 1B will have against LHP, bat-wise Callaspo should outearn Barton by ~3 runs
- Over that same period with the glove, Barton should save ~3 runs above average fielding MLB 1B player
- Breakeven for Callaspo vs. Barton versus LHP  is for Callaspo to be an average fielding first baseman.
So, MELVIN CAN’T GO WRONG!!
It’s basically like going with whichever guy gives you the better BJ that morning.
I don’t believe Callaspo can be average defensively. Barton also has more upside. Barton is 29 while Callaspo is 31, so Barton is the smallest bit closer to his prime. Unless there’s a special need for contact, I’ve convinced myself I would rather have Barton. Melvin, please take note.
Also, I love reading your pieces, ec.
Yeah, from what I read he is at least now struggling with the position.
Thank you.
This is more or less what I was thinking. On one hand, we don’t want to bias ourselves by focusing on upside. But I think there is some value in picking the high variance option, since you need some players to have big seasons to make the playoffs (Donaldson last year, Reddick and Cespedes in 2012). Callaspo at 1B is just so uninspiring.
This is great EC. A quick question does the likelihood of replacing Callaspo for defense late in games minimize the defensive cost for him getting the bats against a LHP?
Thanks. Not significantly. I am trying not to build false precision into this, and I think the overall estimation is imprecise enough to not change significantly with what would probably be some 10% change in innings played.
You sure you don’t want to use 7 decimal places?
I appreciate the comment about false precision, it really seems like a wash, unless Callaspo is really bad at the sack, which I think has a likelihood that melvin should head.
not sure if:
*in the sack
or
*melvin should heed.
If he can’t play first, the whole deal is definitely moot and he should be no where near the line up. Unless, I suppose Lowrie and play there and every one shifts over?
Thanks for writing this.
I remembered the relevant chapter of The Book about how observed splits have to be regressed by a large amount, but was far too lazy to apply the computation to this case.
Glad you liked it.
Yeah, that was done by Andy, who is actually the best of the three when it comes to regression and math in general. I exchanged some e-mails with him when I retackled the issue for THT.
Left out of the discussion has been Freiman:
Consensus wOBA projections: .308 (Steamer .316, Oliver .317, ZiPS .291)
Split:
I’m still too lazy to do the math right now, but he has a huge observed split in a very small major league sample, so the result will be something a little bigger than league average.OK, let’s see if this is right:– Observed: .196/.351 = .558, so a huge 44.2%.
– PA: 148 and 42, harmonic mean is 65.
– Regressed: (65 * .442 + 1669 * .061)/(65+1669) = .075.
– .313 vs. RHP, .290 vs. LHP given same mix of plate appearances.
Defense: Averagish (UZR says a little above, DRS says a little below).
So almost exactly the same as Callaspo in all respects, and thus also about equal in value overall to Barton against LHP. And he just turned 27.
[EDIT: Actually, a little worse, since Callaspo’s overall projection was apparently based on a seeing a disadvantageous mix of pitchers, and Freiman’s based on mostly seeing lefties.]
We can probably all agree(?) that given his lack of experience at higher levels and his struggles last year against RHP (and the fact that he has options remaining) the best decision is to have him playing full-time in Sacramento. But if he performs well at Sacramento for a couple of months, he could easily be the best option for this role by midseason.
Interesting read.
Question:
Why are you not factoring in Callaspo’s position versatility at 2B and 3B into the equation?
(Don’t disappoint me MikeV…)
The equation is only regarding the 1B PA distribution. It was triggered by comparing Barton vs. Callaspo in the other thread.
Wooo!
Thanks, and go As.
I was wondering how much Callaspo’s positional flexibility would weigh into the Barton/Callaspo decision as well. With Little Nicky Punto on the roster, Callaspo isn’t nearly as necessary as a non-1B backup, but I suppose it could be a tie breaker
I was wondering this as well.
There’s also a considerably larger financial commitment to Callaspo, which for the A’s actually means quite a bit more.
What’s the implication there for analysis? Is it that unlike with a non-switcher we should completely ignore Callaspo’s numbers against righties when projecting his numbers against lefties?
Also, great post.
Thanks GM. I really struggle to find an approach that works uniformly on switch-hitters. Perhaps it is really that we should not be talking about the split, but about two separate skill sets. However, I would think that pitch recognition should be mostly unaffected and I just don’t know how to factor that in properly.
Again going by hazy memories from The Book, most switch hitters do have a stronger side, which becomes apparent when their number of observed PA is sufficiently high, but as with regular hitters it takes a lot of data to figure out which it is.
I wonder if you would see any difference in observed splits of switch-hitters if you grouped them by which hand they throw with. You would definitely still get some individuals wrong, but it might be (incrementally) better than assuming that they are all the same.
Yeah, I played with that, too
I would guess that a bigger difference would be whichever one they started with, although that might correspond with throwing hand.
Probably. Also harder to find the data. Any way we look at it, there is just more noise than with single-handed batters.
I enjoy reading this kind of stuff and learning, like I enjoy following the bouncing ball. But I’m afraid that 3 minutes after the song is over I can’t even remember the tune
(Please excuse uncorrect touristic uncharacteristic capitalization; I am using the dictation machine.)
heh
This just confirms to me that both options are incredibly mediocre.
Agreed.
Yes.
Quick, somebody teach Jaso how to hit righty.
Yeah, throw out Barton’s lack of options and it’s a non-starter.
EC, should any weight be given at all to the fact that Barton’s reverse split has pretty much disappeared since he suffered that injury in 2011? I realize that breaking down small samples even further is frowned upon, but because of a major injury, it might be good to differentiate between Barton 2010 and before (when he demonstrated most of his reverse splits ability), and Barton 2011 to the present.
In both his major league and minor league performance the last two years, his reverse split has, well, reversed itself.
That’s a tough one. Injuries are often good enough reason to “start a new bucket”, but not always. Tough call.
Also:
Moss
Mean projected wOBA .344
Observed 290 PA with .308 wOBA vs. LHP, 1260 PA with .349 wOBA vs. RHP.
Observed split: 11.7%
Harmonic mean of PA: 471
Projected split: (471 * .117 + 2261 * .098) / (471 + 2261) = .101
Projections: .350 vs. RHP, .315 vs. LHP.
The conclusion seems to be that against LHP Barton should play 1st and Callaspo DH, with Moss on the bench.
Per lenscrafter’s earlier comment, the other conclusion would be that we really need another decent right-handed bat.
Good catch
While Gentry’s out, probably. When he’s back, I would want to use the DH to get all four outfielders in the game against LHPs.
And also pick up another decent right-handed bat.
or just keep Taylor
Thanks, and go As.
Does not qualify.
this
oh right I forgot I’m wrong again nevermind fuck taylor
Thanks, and go As.
Why do you think we are wrong?
He’s 28 and never gotten a shot. Maybe he should get one.
Doesn’t really matter though, I doubt he breaks camp with the team.
Thanks, and go As.
again?
wait… are we bringing cust debates back??
:p
Man, that’s some heavy lifting there, elcroata. I very much appreciate that the weighty analysis by you and others here allows me to seem well-informed when I use it with others.
In other news, it’s not every U.S. state that sees high-ranking elected officials get indicted for arms trafficking. A liberal Democrat, no less. California, fuck yeah!
is introducing legislation, making a proclamation, or introducing some one to someone else in exchange for campaign money a crime?
link
I did say indicted, now. He might not be guilty. Though “The criminal complaint alleges that Yee later repeatedly took bribes and participated in explicit discussions about facilitating a gun-trafficking deal” sounds like a little more than just legislating/proclaiming/introducing.
I do believe that the feds tend to have a raging hard-on for political corruption cases. But they also will walk away from cases they don’t think they can make stick (like Perata and Willie Brown). Maybe it’s an over-charge to get a plea, but Yee’s guilty of something more serious than being a political whore for campaign donors.
“his alleged crimes come down to taking campaign contributions in return for issuing proclamations, using campaign funds to set up a meeting and taking campaign funds for writing a letter.”
Ed, either you’re ghostwriting Willie Brown’s column or he’s trolling FK for bon mots.
Yee is renowned among journalists for sending out tons of press releases and offering his thoughts on pretty much everything under the sun. It seemed sad today that we couldn’t call him for a comment on this.
you have a new guy named kevin working there?
Not that I know of, but might be an intern or TV or online.
I don’t think he is an intern. He left a job at KTVU so that wouldn’t make sense.
TV makes sense though.
Every journalist in the city was at the federal building that evening.
Brandon Moss, incessant chatterbox. The story about Barton in the woods is pretty funny.
Tagline?
Done. Hard to imagine a better tagline for this blog.
Right to the heart of matters really
Nice article about locally produced hot sauces, peppered (sorry) with references to HEAT, the hot sauce shop on MLK near Trader Joe’s in Berkeley, whose proprietor looks an awful lot like Tim Lincecum. In any case, and perhaps against all odds, HEAT continues to thrive as a retail hot sauce outlet. When I was there last month Tim Lincecum told me they had a fantastically busy Christmas season and were doing well.
well, i’m piqued!
also of note, TL-alike’s shirt.
See–if they can make a go of it, I think the vegan shop can make it work. Which by the way, I went to a few weekends ago and found their prices to be totally reasonable which was a nice surprise. I imagine I’ll be a frequent shopper.