Two Different Ways to See Melky Cabrera
Maybe the most important thing you learn early on in any basic stats class is that you can’t just throw out data. At least, not just because you want to, and not just because the data doesn’t fit. Just about all information is legitimate information, and you simply have to decide how heavily to weight it. Take Chase Headley, for example. No one figures he’s going to hit 31 dingers again, like he did in 2012, but the reality is that Headley did have a seven-win season just a few seasons ago, and we can’t justifiably ignore that. It’s a part of his record, and it hints at his true-talent level, or at least where it was in San Diego that one time. Because of that year, Headley gets a more favorable projection, and I don’t think you can argue that away.
If you’re going to eliminate data from a sample, you need to have a damn good reason. You need to be able to prove that the data is irrelevant. If you’re a research scientist, maybe the data came out of an experiment run you know you messed up. You accidentally buffered a solution to the wrong pH. As baseball fans, we’re not research scientists, but we’re still always looking for reasons to eliminate data. This is basically the same thing as having a disagreement with a given player projection. Overall, the projections do well, because they don’t eliminate data. But we’re always trying to beat them. Last year gave reason to eliminate prior data from J.D. Martinez. The White Sox saw reason to eliminate prior data from Zach Duke. And now the White Sox have also signed Melky Cabrera for three years and something like $42 million. Cabrera’s another interesting case, like Duke — he didn’t just overhaul his mechanics, but there’s something about his record that makes you wonder how much you should care about his 2013.
A tumor! A tumor in his back. Melky Cabrera, in 2013, had a tumor in his back, and while it was eventually discovered and removed, it was apparently causing him great discomfort through the year, especially in his lower body. Perhaps not coincidentally, Cabrera had a terrible season in 2013, and then the tumor was taken away, and Cabrera did well again. This goes beyond, say, “he struggled through hamstring discomfort” or something. “…when healthy” tends to be an over-optimistic suffix, since no one plays a whole season at 100%, but there’s a difference between ordinary aches and pains and having a tumor in your back. Isn’t there? It seems like there should be.
Deciding what to do with Cabrera’s 2013 goes a long way toward determining your opinion of the contract. Let’s run through some math. Here’s a table of Cabrera’s last four years. I’m including WAR/600, and a basic weight for purposes of calculating a 2015 projection. This is a very basic projection, and it doesn’t account for age, but just focus on the result here for a minute.
| Season | WAR/600 | Weight |
|---|---|---|
| 2011 | 3.1 | 0.3 |
| 2012 | 5.4 | 0.4 |
| 2013 | -1.5 | 0.5 |
| 2014 | 2.5 | 0.6 |
| Projection | 2.2 | – |
Cabrera’s 30 years old, so he shouldn’t age too steeply all of a sudden. If you treat 2013 like it was just an ordinary season that happened, two seasons ago, you see Cabrera as almost exactly an average player. And, as it happens, Cabrera is getting paid as if he’s almost exactly an average player. He’s getting paid, roughly, to contribute about six wins to the White Sox over three seasons. What that implies is that the White Sox aren’t just tossing out Cabrera’s 2013. But look what happens if you were to do that:
| Season | WAR/600 | Weight |
|---|---|---|
| 2011 | 3.1 | 0.3 |
| 2012 | 5.4 | 0.4 |
| 2013 | – | – |
| 2014 | 2.5 | 0.6 |
| Projection | 3.5 | – |
If you just throw it out entirely, Cabrera looks a heck of a lot better. That’s 59% better. Last season, he was a little better than average. In 2011, he was a little better than average. In 2012, he was a borderline superstar before getting suspended. Now, that’s another variable, and some of you might have a desire to throw out steroid-fueled data, but we know relatively little about what performance-enhancing drugs mean for offensive performance. Cabrera just hit well in 2014, presumably off of the stuff. We could observe that Cabrera was crap in 2013, hardly able to move, and the reason for that feels pretty significant.
That 2013 makes Cabrera a little bit unpredictable. Because of that year and because of his medical condition, you could simultaneously argue that Cabrera’s a little overpaid, but also he has big upside. Some of Cabrera’s indicators remained unchanged. In 2013, his walks were normal, and his strikeouts were normal. His batted balls were normal. Cabrera simply lost power and defense. We know he’s not a good defender when healthy, but he should be better than he was, and he should hit for more power than he did. In Cabrera’s last three healthy seasons, he’s been a good player. There’s just also a fourth season, gumming up the works.
Where I think you come down is that Cabrera is worth something in the neighborhood of $42 million over three years, but he’s got a little extra volatility. It’s worth noting somewhere in here that Cabrera apparently turned down a four-year offer, potentially from Seattle, because he wanted to play closer to the east coast. So there was a bigger commitment out there, but it wasn’t sufficiently big to change Cabrera’s mind. The White Sox seem to be focusing mostly on Cabrera’s 2014. Call this a third way to see Melky Cabrera. We know that something happened that had an effect on Cabrera’s 2013 performance. Yet we also know that something happened that might’ve had an opposite effect on Cabrera’s 2012 performance, so the White Sox probably just see Cabrera as a line-drive sort with just enough contact and power to offset uninspiring defense. It’s not an All-Star package, but Cabrera isn’t yet old and Cabrera also isn’t Dayan Viciedo. I can’t tell you how badly the White Sox wanted to move forward with Not Dayan Viciedo. So, think of this as mission accomplished.
The White Sox have achieved almost all their offseason goals. They’ve now built a team talented enough to hang around the race for at least the bulk of the season ahead. A month ago, the White Sox didn’t look like a contender, so in that sense it’s odd to see what they’ve decided to pull off, but on the other hand, Jose Abreu, Chris Sale, and Jose Quintana are good now. So why not seize the day? The day is being very aggressively seized.
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.
If you throw out 2013 as an outlier, should you not use similar caution with his potentially roid driven 2012 year ?
“In 2012, he was a borderline superstar before getting suspended. Now, that’s another variable, and some of you might have a desire to throw out steroid-fueled data, but we know relatively little about what performance-enhancing drugs mean for offensive performance.”
“We know that something happened that had an effect on Cabrera’s 2013 performance. Yet we also know that something happened that might’ve had an opposite effect on Cabrera’s 2012 performance…”
Read the article.
Speaking of what is known about PED for offensive performance, this website has a lot of good info: http://steroids-and-baseball.com/
Not all evidence backs that PED enhances offensive performance. This website, by a respected sabermetrician, captures a lot of information regarding why PED is not likely to improve offense. In fact, amazingly and ironically, in our times of sabermetrics, I find that a lot of people use their eyes or a “smell” test as their reason for why they believe PED helped offensive production, much like how we like to deride the prior generations for using their eyes or a “smell” test about how good a player is.
Does he use batted ball velocity data? Because that would be the key.
You also have to control for when someone starts using PEDs (which is, of course, data none of us have access to):
A guy who starts using HGH at the age of 33 to stave off declining bat speed may still be worse than his average performance from 23-32.
A guy who isn’t quite talented enough for the MLB (which I believe accounts for most PED users, personally) and uses PEDs to be able to stick in the bigs will still be worse than his more talented peers, including the clean ones.
Meanwhile, a good contact hitter who, say, starts taking testosterone when he’s 26 and in his physical prime, would likely dramatically increase his performance, just because the additional raw strength to go with the preexisting contact skills should increase his BABIP and ISO, since he’ll simply hit the ball harder.
Saying of something ‘maybe it doesn’t affect performance’ is a meaningless statement. There are mountains of evidence that steroids do enhance performance in multiple sporting contexts. THAT is evidence even more difficult to ignore. It continues to amaze me that we have weasel arguments on this in baseball when in other sports the impact of steroids and other enhancers is beyond any doubt (with the actual outcome of HGH use the most problematic so far). Hitting a baseball is a complex act, and guys results will, yes, vary. In aggregate there is really no doubt that doping helps.
But let’s ignore the larger context for the moment. The hypotheses of this post from the get-go is ‘we can’t ignore demonstrated, quantified facts.’ The argument that followed from that is, let’s try a counter-factual and throw out Melky’s worst year. From that standpoint, it is at least as legitimate an experiment to throw out Melky’s BEST year. When we do this, Melky’s projections aren’t nearly as pretty, are they?
It is far more relevant to exclude Melky’s 2012 than his 2013. We know he was on stuff, and we know he had a performance spike he never showed before or since. The _actual_ outlier year is 2012, since we know also that guys with injuries have performance drops, so 2013 wasn’t an outlier, just the normal hazards of sports performance. (I don’t for a second buy Melky’s ‘it was the tumor, guys’ excuse for 2012 either, since many guys who have been on the juice have a collapse year when they go off abruptly, at least as relevant an argument for what happened to him. Guys who dope have a lot of excuses, always about outcomes that don’t favor them: ever notice that?)
All of this ‘somebody’s saber buddy who doesn’t know jack about medical pharmacology thinks’ weaseling is disingenuous. At best. And ignores plenty of evidence outside of baseball’s bubble excuse-world that PEDs have major impacts. But to me, the worst problem in this post is simply the reasoning. 2012 is far more of an outlier, so if we are going to exclude a year as a counter-factual hypothetical, that would be the year to exclude. Except that then Melky doesn’t look so good, which outcome would seem to counter-argue where this post wants to go. Picking the less likely outlier when the outcome works for you as opposed to the more likely outlier when the outcome works against you is other than intellectually honest. Want to be honest?: show both outcomes and let the audience choose. That didn’t happen in this post, did it . . . ?
Here’s a nice alternative: exclude 2012 _and_ 2013. Melky’s projections don’t look so good then—and are relatively meaningless excluding facts which we are advised to include at the start in this discussion. It’s an interesting rhetorical approach: advise the audience not to do something, then do exactly that later on to get to where you want to be. I mean, who could criticize a guy for that, he gave fair warning it was a bad idea?
He addressed that directly in the article.
It’s hard to figure out when his roid use started, and whether it had a significant impact on his hitting. After all, prior to his Giants stint as well as the 2014 season, Melky showed he was an above-average player.
I think there is some circumstantial evidence to say it started before the 2011 season.
In 2010 he was traded to Atlanta, where he was fat and lousy. He was non-tendered after the season.
That off-season he “re-committed” to getting in shape and was training with A-Rod…
He showed up in KC in good shape and had a nice season. The next season he was very, very good, but got busted.
All that said, I think this is a good contract, it seems like it has the volatility risk baked into the dollars and years.
Nobody ever got in shape without drugs.
Right, because he’s already been ‘roided up, so he doesn’t have any further need to ‘roid up.
MLB is so weird about all this. OTOH, a whole generation of players is now suspected of having been ‘roided up, when only a few, uh, some, ‘mokay quite a few but not all were; and not even discounting someone like Bonds, who’d be in HoF no matter if ‘roids never existed, can seem to overcome this inane institutional ‘forgetting’ the HoF voters think they’re getting away with.
But OTOH, ChiSox weren’t by any means the only franchise that seriously considered signing Melky The Lab Experiment back on for another voyage or 3. ‘Alright then, we’re giving you 43 mil and you’ll be set for life able to buy not just real estate but people and certain island nations; but, understand, you’re not ever getting into Coopertown, not matter what.’
“Nobody ever got in shape without drugs.”
Lots of people get in shape without steroids. However, those who have tested positive for steroids a year after getting into shape probably used them to get into shape for god’s sake.
I agree that there is circumstantial evidence that Melky was on the stuff in KC too. Anybody who thinks he got to 2.5 WAR is fooling themselves or hoping he stays on. Sheesh.
Yeah, the guy trains with A-Rod in the offseason, proceeds to have the best season of his career in terms of power and BABIP, and then is popped for using testosterone, that he got from Biogenesis.
Seems pretty straightforward to me, and it was a good decision by Melky.
Why read the article when you can just come snark in the comments!
I’d say that comments drawing a line from Melky’s illegal PED use and his simultaneous spike in performance are simply acknowledging reality.
The whole “how much did the PED’s help” will of course always be unanswerable, for everyone from Melky to Lance Armstrong to Marion Jones to Michelle Smith to Barry Bonds. All we can know with absolute certainty is that it was illegal/unethical.
The more interesting question for me is, Could the PED’s have contributed to the tumor’s formation? (Just as the question has arisen of whether Lance Armstrong’s cancer could’ve been instigated and/or exacerbated by injecting & ingesting numerous illegal PED’s over several years.)
There’s no doubt about Armstrong. He said, without qualification, that he couldn’t have done without the stuff. We can never know ‘how much’ PEDS helped a given individual, only that ‘a lot’ is the median answer.
There have been waaay to many accredited posters here at Fangraphs in recent years who don’t like that answer, and pile more than questionable rhetoric on top of it in attempts to make it disappear.
I think steroids turned Barry Bonds from Willie Mays into Babe Ruth.
He already had 400-400 before he started.
The thing about this whole conversation is that steroids don’t only affect you while you’re using them. They help you build muscle, and that muscle stays even after you’re not using them anymore. It can atrophy, of course, but muscle built with the help of steroids doesn’t atrophy any more quickly than muscle built without it, and so it should be a long, gradual decline, not a sudden fall-off when he stops using steroids. So I think the best way to think of it would be as analogous to the ageing process, in which the player has aged considerably since his steroid-fueled years, but they still happened with the same player – so we should just weigh them less heavily.
Except that completely discounts the known destabilization of an individuals hormone system due to taking massive doses on exogenous steroids. One’s own endocrine system _quits producing the natural stuff_ because it is fooled by the excessive amounts in the body. And there is ample evidence that the endocrine system does NOT bounce back immediately upon going off exogenous inputs but stays suppressed for extend periods of time, sometimes several years, or even being permanently destabilized to varying degrees. The suppression varies considerably, and is also affected by what all else a guy was using, how much, how long, his age and health, &etc.
There is ample reason to believe that numerous guys will severely crash in hormonal output and hence physical performance for an extended time after going off massive steroid injections abruptly. This is in fact why there are tailored PED cycles of lower dosages and mixes specifically to ward off big collapses. But big collapses are exactly what we should expect to see is guys hyper dose and go off suddenly. Like about the time they sign that big money contract and don’t want to get busted and lose big chunks of ill-gotten gains.
We have seen a very large number of instances of this kind of pattern in the last 25 years of baseball history. Good-not-great guy gets big body bulk and boffo performance results that he didn’t project for pre-change in obvious visuals. Gets big deal. Utterly sags in performance for _several years_ like somebody’s weak sister with mono. Blames ‘this/that mysterious, strength-sapping injury, but it’s all fixed now’ except it isn’t. Takes the money to the bank like clockwork. Has disingenuous arguments slapped up on various putatively analytical baseball websites. Yeah, we’ve seen too much of that kind of cycle . . . .
I guess we need to know more just how much steroid ingestion can drive individual numbers in an individual season. Certainly he should not be overly penalized for playing through pain, as he did in 2013.
I like this signing and I like what Chicago is doing as I think their window will be pretty wide but I’m still not convinced they have enough to compete in the AL in 2015.
Maybe, but maybe they have enough to do exactly what the Royals did in 2014. Same division dominated by a daunting Tigers team, same AL full of expected powerhouses like the Angels and unexpected ones like the A’s and Baltimore. But the Tigers got injured and the AL West beat up on itself and the AL East imploded and it turns out there are now two Selig Lottery Tickets per league each year. The White Sox only need one. And even if they fall short like the Mariners did, that’s enough to drive butts into seats late into the season (as it did in Seattle). For a White Sox team in a city where everyone has anointed another, different MLB team as the one to watch in 2015, that’s got to be especially important.
Question for the panel: if Scherzer signs anywhere besides Detroit, how good are the Tigers actually?
Tigers should still should be best team in AL Central on paper, but stuff happens and they have questions (like every team). Tigers are is a position where its going to be hard to upgrade, budget seems to be maxed out and what do you offer in a trade anymore?
IMO, not that good. VMart will be worse. JV doesn’t look like a good bet to be much better. Miguel will be a year older, and is likely out of his Barry Bonds phase. JD Martinez will likely be a little worse (at least) due to regression. No more Porcello.
The only upside, as I see it, is in Castellanos and a full season from Price/Anibal.
Max would be a big (5 win) boost for them.
Boy, that JV extension looks stupid right now. He would just be entering FA this offseason, and if they hadn’t unnecessarily locked him up, they could swap Max in for JV and have another 3 wins for free.
“I can’t tell you how badly the White Sox wanted to move forward with Not Dayan Viciedo.”
I’m not sure this point can be emphasized enough. The primary objectives here were to write a different name on the card for LF and upgrade the #2 spot in the lineup. Even 2013 Melky Cabrera hits significantly better than 2014 White Sox #2 Batter. That doesn’t necessarily justify limitless spending on its own, but the move does accomplish those goals.
The thing about this White Sox team is that the upside is pretty high, and there are a lot of low hanging wins if not-that-unlikely things go their way. If Hahn managed to build an above average bullpen, AND Sale pitches a full season, AND Rodon can deliver half a season of above average starts, AND Abreu’s regression is minimal, you’re probably looking at a 90-92 win team.
The point is, they may only be a 80-82 win team on paper, but there’s a bunch of things that could go right — and you can easily see how they are poised to go right — that could buy a dozen more wins.
I agree. They aren’t instant favorites to win a championship or anything, but they have successfully moved the needle from “sucking” to “not sucking”
Yeah, I totally agree. If you look at the projections on their depth chart/standings page, which still project them to be an under .500 team, it looks like they present pretty reasonable downside projections (excepting long-term injury to presumed starters).
For example, the projection has Melky Cabrera at 1.7 WAR and had Viciedo at 0.7 WAR. It’s possible that Melky is only one win better than Viciedo, but last year, Viciedo was worth -0.7 WAR and Melky was worth 2.6 WAR.
And what’s the deal with Quintana? He’s expected to contribute nearly 3 fewer wins. His home run rate was really low last year, but last year’s xFIP and SIERA don’t indicate nearly as much regression as the Steamer projections for his ERA this year (3.95)
Eaton and Abreu have the same deal, in their case due to BABIP regression. It’s totally plausible that they’ll both contribute fewer wins, but aren’t they also on a part of the player development curve where we’d expect them to get slightly better? .350+ babips might be unlikely, but .330 seems plausible, since they hit to all fields and don’t hit a lot of fly balls–or in Abreu’s case, a lot of fly balls that stay in the park.
Wait, so you think that a projection system that literally predicts a massive, across-the-board regression for every single White Sox player is a “downside” projection.
I’m not sure what you’re implying. Are you arguing that the fact that the projection system assumes everyone will get worse should make me more confident that each individual player will get worse? If anything I would be inclined to think the opposite. If you find a bunch players with a common attribute who all beat their projection, then there must be some attribute common to all those players that the projection system isn’t accounting for.
Further, I think about the projections this way: there are a range of potential outcomes for every player. A projection is going to give you one number, without giving you an idea of the error associated with that prediction. You have to evaluate error yourself based on a number of factors. We know that, across large samples of major leaguers, BABIPs tend to regress toward some average number. The same goes for individual players: their BABIPs tend to regress toward their career average (I’m not sure what approach the projections in the depth chart assume, but I think it’s the former). On the other hand, we also know that players with certain profiles have consistently higher BABIPs. At this point in their careers, is it more logical to assume that players like Eaton and Abreu are going to be career .330 BABIP hitters or career .300 BABIP hitters? I think you can make an argument either way, but since both had BABIPs over .350 last year, should we expect .317 and .310 represent the peak of the distribution of outcomes, or somewhere around the 1st quartile. I would argue the latter, and I would call that a “downside” projection.
What about 2010? Ignoring that one too?
2010 is far enough away that it isn’t very useful to include in projections, so, yes.
Then again, in the entire four seasons prior to the last four seasons, Melky was worth virtually nothing over replacement (0.5 WAR total, as in ~0.125 per year) over 2114 plate appearance, no tumor in sight.
He’s had a strange and surprising career.
Well, he was probably brought up before he was ready for the majors, so the early part of his career can probably be explained by learning on the job.
The latter part was at least partly fueled by PEDs.
It is both surprising and unsurprising. Unique, I will give you.
That’s falling straight into the trap Sullivan is describing, though.
This. And it is exactly this that should put massive error bands on any projection of his future performance.
Well, to be fair, 2010 was never really indicative of his true talent, so much as it was indicative of his method of dealing with the grief of being separated from his True Love, Robinson Cano.
Melky, Peralta, Cruz. It’s kind of interesting that Baseball has taken a huge dump on Bonds and Clemens, but these PED users still get paid the big bucks even after it’s proven they did the same thing.
The ones doing the dumping aren’t the same ones handing out contracts.
They kind of are the same people. The owners are the one’s giving out the big contracts and the ones who didn’t do enough to clean up the game.
The ones you blame for not doing enough are not the ones dumping on Bonds and Clemens. This is not supposed to be a trick question. The media and the veteran players have been the most vociferous critics of PED users, and they have nothing to do with what contracts current players receive.
I seem to recall those that hand out contracts didn’t sign Bonds when he still projected to be a tremendous on-base machine, at the very least. His last season slash line: .276/.480/.565
No one cared about the money Bonds and Clemens were making. If Johnny Peralta hits 70+ HR’s in a year, I imagine the people will take notice.
Implying pre-roids Bonds was the offensive equivalent of Jhonny Peralta…
No one implied. Someone inferred.
What about 2005-2010? Surely those 800 games give us a bit of information about what his TTL is as well? And the answer to that is: pretty freaking bad.
http://www.fangraphs.com/statss.aspx?playerid=1109&position=OF
http://www.fangraphs.com/statss.aspx?playerid=302&position=OF
http://fivethirtyeight.com/features/what-powerlifting-tells-us-about-the-effects-of-peds/
http://sports.espn.go.com/mlb/gammons/story?id=1938586
“Prior to 2000, Bonds hit three homers longer than 450 feet; in the last five years (2000-2004), he has hit 26.”
Also, having such little different weights between years is a bit curious. In my experience, most people tend to make the differences significantly larger to account for the higher relevance of more recent performance.
I’m not saying PEDs definitely did not help Cabrera, but his best offensive season in San Francisco was fuelled by a .379 BAbip (as opposed to his career figure of .311). His second-best season by OPS also saw his second-best BAbip.
Would PED-induced increases in strength and recovery time necessarily lead to higher BAbips?
BABIP and PEDs are correlated if I am not mistaken
(hit the ball harder, more hits…)
But is that really true? It’s difficult to know, because when we think “PEDs” we think of players whose extra successes were based on hitting balls *out* of play. Bonds, Sosa and McGwire, for example, only saw modest spikes in their BAbips during their power years, certainly nothing compared to their increased proficiency at hitting the long ball.
Melky has actually continued to have BAbips over .300 with the Blue Jays, compared to a .290 career BAbip prior to 2011. So that doesn’t strike me as proving that his BAbip spike in 2011-12 was necessarily down to PEDs. If anyone has seen research on this subject I’d be interested in seeing it.
PS If Melky was hitting the ball harder, wouldn’t that show up in his HR/FB% or line-drive percentages? of But his LD% has been higher over the last two years than it was in ’11-’12. And his HR/FB% was as high this year as it was in 2012.
Yes, RJ, it’s really true. We know that line drives and balls hit with high velocity are more likely to go for base hits.
Are Cabrera’s batted ball velocities available somewhere? It would be interesting to see if there was a spike in 2011-2012.
I’m looking at his outcomes on each hit type, and what really stands out in those two years is his success on fly balls. He had OPSs of.829 and .992 in ’11-’12; he’s never been above .675 in any other year.
Probably because he hit them so much harder that year, haha.
Cabrera could go out there and be the most average of all the averages and I’d still be happy that I’m not watching Viciedo strike out as hard as he possibly can.
As a Jays fan, Melky Cabrera is an interesting player.
I actually disagree with the author about how to judge the deal based on looking at the past four years.
I think that last year is a pretty accurate picture of what you can expect a good Melky Cabrera to look like.
The author says that he is 30 so he shouldn’t decline too steeply, but that also means that I don’t see him improving a lot either. As he gets older, his (bad) defense will only get worse and his (limited) power won’t improve dramatically. Which means that his upside comes with his bat, but he already hits around .300
So, we’ll just go with that then.
That’s kind of what I was thinking. He’ll probably be around a 3 win player next year, with a little bit of decline each following year.