I don’t know how to do this better, but considering many of these players missed significant time in either 2011 or 2012, and defensive metrics are already suspect when looking at single years, I’m not sure how meaningful this is.
Additionally, I honestly just don’t believe that actually losing weight would not be a benefit, so if there is a story here I think it is something you somewhat touched on but did not say explicitly: players are liars.
I would fully expect Peralta’s UZR numbers to take a significant hit in 2013, just due to natural regression. Does anyone really believe he was the third best defensive SS in the league last year? Didn’t think so. So, when the numbers aren’t as sparkly, let’s not attribute the difference to his weight- or lack thereof.
Several years ago, Peralta came into camp touting off season laser surgery for his eyes as a projected benefit to his hitting in the upcoming season. He’s definitely a guy who looks for new angles to improve.
No kidding!!! I’m a 220lb financial risk manager and that’s considered beefy for this profession.
Comment by kevinthecomic — February 1, 2013 @ 4:51 pm
“DRS characterizes him as being about a league-average defensive shortstop. UZR characterizes him as being an above-average defensive shortstop.”
They don’t work very well then.
Sometimes it seems that stats like UZR and DRS might be revealing something that can’t be easily observed with the naked eye. Sure, stats are great for that, but these defensive metrics are obviously wrong often enough that I wonder if we’re not just better off without them.
I would like to see how this correlates with age. It seems to me that the players who lost weight (and gained weight, for that matter) were all a year older that year than they were the previous year. They probably should have been at least slightly worse defensively in year 2. The fact that their defensive performance was roughly the same may indicate that the weight loss worked to counter the effect of aging. In other words, it might mean that losing weight worked because the players would have been worse due to age had they not lost weight.
An ambitious study with a tough to prove goal but an admirable attempt. One of the greater issues being the difference in the physical demands of each position i.e: comparing Miggy to Heyward or Inge to Fowler seems a bit unfair.
Still, nice work though. Certainly worth considering
What I’ll never understand is athletes making millions of dollars needing to be told to get in shape. Shouldnt Peralta assume that, oh I don’t know, that because he’s a SS maybe he should be relatively agile?
TroutFan and Oscar, there’s nothing wrong with reporting a lack of definitive findings or a negative finding. In fact, it’s essential to any experimental process that we determine which hypotheses are correct, which are incorrect, and which need more study.
Just because the “data mining” or “statistical analysis” doesn’t lead to some ground-breaking discovery doesn’t mean it’s not worth reporting what you find — it can inspire others to approach the same question from a new angle.
I agree. I think the lost weight will definitely help his range. Especially with the little things, like turning double plays. If he’s making more difficult plays, he might make more errors though. Even if that’s the case, it would be nice to see the effort.
In addition to the caveats you cited, here’s more: Your comparator groups.
Among the players who lost weight who you cite, only one was a middle infielder shortstop or even a middle infielder and he saw a marked improvement in UZR, Ian Desmond. The rest of the group are largely corner outfielders and infielders and may not make suitable comparators to a shortstop.
Your weight-gain group is meant to test whether there is an opposite effect but that would only hold some validity if those players went from being out of shape to in-shape. But some of these players did not — they went from being skinny to adding some muscle — Andrus is a prime example.
Finally, while Peralta’s total defensive rating changed only marginally in 2012 his UZR range rating improved a much more substantial amount that was offset by a worsening in error rating. If the point of his weight loss was to improve his range and if you believe such small samples of UZR, then his weight loss achieved its intended effect.
Comment by Jonathan Sher — February 2, 2013 @ 8:27 pm
@Baltar – Depends on when in the season. If you’re talking about Ryan Raburn, about 180 lbs of completely dead weight.
Earlier, I was looking at fielding runs based on Revised Zone Rating and Out-of-Zone plays using a method that Colin Wyers developed a few years back. I don’t know that this method is any better than any other defensive metric, but the results for Peralta made sense. According to this, he was one of the best in baseball in the zone (steady) and one of the worst outside the zone (no range). Combining the two, he was -3 runs below average. http://www.detroittigertales.com/2012/12/in-previous-posts-i-looked-at-revised.html
So do advanced fielding stats have value or not?
According to the numbers Peralta was indeed a solid SS last year, yet many commenters disregard them because they don’t fit their narrative. When some argued during the MVP battle that Trout ain’t Willie Mays and throws like Johnny Damon, the pro-Trout group point to the positive fielding numbers to fit their narrative. Can’t have it both ways…if UZR can’t be trusted yet, why use them?
Comment by tbonemacd — February 3, 2013 @ 12:15 am
For what it is worth, if a player or group of players is -3 in UZR one season and -3 the next season, that implies a loss in UZR talent. That is because -3 for one season or a partial season needs to be regressed, perhaps half-way towards a mean (say, zero), which puts their year X UZR true talent at -1.5. So a -3 next year means a loss of 1.5 runs in UZR talent.
That, BTW, is entirely consistent with aging. Players lose 1-2 runs a year in UZR talent at almost any age (due to aging – UZR peaks quite early, probably in the early 20’s).
What are you talking about? The samples are the same size. Both years would get regressed equally. Why would you assume that a -3 in year 2 is somehow more indicative of true skill than the -3 in year 1? If a group of drivers of unknown skill got into 10% more accidents than average in a given year, then the next year they also had 10% more than average, you would claim that they actually lost skill between years 1 and 2? Interesting. By that logic, a team that goes .450 win pct. in year 1 is a lot better than the same team that goes .450 in year 2 (assuming same roster and no aging)? That’s not the way it works. You have more data, so all you can say is that the true talent is closer to .450 than you first thought (after year 1). However, you cannot say that the players got worse from year 1 to 2.
Data mining is great. But this article has no ACTUAL statistical analysis. No strength of correlation, no nothing. It’s an arbitrary sample of some numbers and then no attempt to prove, disprove, demonstrate, or anything. I wouldn’t even be posting if it wasn’t a frustrating trend for Sullivan and a lot of the others on this site. Compare this to a Dave Cameron article, or Tom Tango or Dave Studeman or Mike Fast or Pizza Cutter or etc etc.
There’s no original insight, there’s no rigor. It’s fluff. And after the twentieth one of these in two weeks, I get annoyed. This is not the scientific method.
Seems to me like if you wanted to try getting any kind of meaningful analysis with this you would have to get much better information on players weights and break them up into groups based on their height to weight ratios.
First off you don’t know which of your players kept the weight off for more than a few weeks into the season and which put it back on or even gained more weight.
Second, you don’t know which of your players would have been considered overweight to begin with. You have young players on there who may have actually been considered underweight who merely gained muscle weight. That’s fairly different than an overweight player with too much fat.
It seems kind of lazy to try to do this kind of analysis without really gathering the necessary data and just lumping together what little you found without thought of whether what you were comparing is actually comparable.
Minus three for Peralta passes the smell test. He’s probably not close to as good as UZR thinks he is, and he’s certainly much better than fans who only look at range think he is. Fringe average overall.
Comment by The Party Bird — February 3, 2013 @ 10:19 pm
I look at Jhonny’s effort to lose weight and effort to try to increase his range (I don’t think “range” can be taught) and say that he sees a really good thing happening in Detroit and he wants to his best to be a part of it. If his hittiing last year was an anomoly, and I think it was, he will be a valuable part of a championship team. If he bats 7, there should be men on base when he comes up, RBI’s should go up also. We certainly don’t have any shortstops in the pipeline if we read ann article about Cale Iorg, he can’t hit my wife’s weight, let alone mine, which would be a good average!! That makes Jhonny look all the better unless someone is willing to part with a good shortstop for Porcello or Smyly…I say let’s keep all three!
Comment by Alaruss1942 — February 4, 2013 @ 8:26 am
We agree that this isn’t ‘statistical analysis,’ in the sense that there is no attempt to do a rigorous, parametric correlation of something like weight-change and UZR.
But my point is that one needs to start somewhere. Jeff’s article could be the first step in a conversation about where to take this for further research. And frankly, given that the information on a player’s weight is self-reported, I’m not sure there’s much better that can be done. But an article without a firm conclusion one way or the other could be seen as an invitation to do further research.
Comment by Chris from Bothell — February 4, 2013 @ 4:36 pm
I wonder what type of effect weight loss/gain has on older players and their injury probabilities. Packing some extra pounds can do some real damage to the ol’ knees, especially for older players who play physically demanding positions. If losing 18 pounds makes it less likely that Peralta suffers an injury, then even if there is no effect on his overall defense, the ability to play in more games creates a positive effect overall.
Comment by AdmiralWhiskers — February 4, 2013 @ 4:47 pm