What Jhonny Peralta Tells Us About Defensive Metrics

Five years ago, the Cleveland Indians decided that Jhonny Peralta just wasn’t capable of playing shortstop at the Major League level anymore, shifting him to third base to allow Asdrubal Cabrera to move back from second base to shortstop, the position he had primarily played in the minors. Peralta had never put up particularly good defensive numbers at shortstop, and with a thick lower half, he certainly looked more like a third baseman than a middle infielder.

After roughly a year at third base, while still hitting like a shortstop, Peralta was traded to Detroit. The Indians weren’t going to pick up his $7 million option for 2011, and the Tigers were looking for an infielder to give them some depth on the left side of the infield. Peralta played third base for a week with the Tigers, but then incumbent Brandon Inge returned from the disabled list, and the Tigers moved Peralta back to shortstop.

Since that move, Peralta has played the position exclusively, spending four years at shortstop between Detroit and St. Louis. And along the way, a funny thing happened; UZR fell in love with Jhonny Peralta’s defense.

Over the first 6,000 innings of Peralta’s career at the position, UZR had rated him as a -28 defender, or -6 runs per 150 games. He was decidedly below average, even though he was very good at error avoidance (+8 runs) and turning double plays (+5 runs). Peralta’s lack of range (-41 runs) was a legitimate problem, though, and what got him moved off shortstop to begin with.

In 2010, in his first full year back at shortstop, UZR rated Peralta as an average defensive shortstop. In 2011, that jumped to +10, putting him in the same range as defensive specialists Brendan Ryan and Alexei Ramirez. It was, at this point, that Peralta became something of a poster boy for the flaws of defensive metrics. After all, everyone knew that Peralta was actually not a good defender. The Indians had already moved him off the position. He had no range. Ranking Peralta as an elite defensive shortstop was evidence that UZR was not to be trusted.

Instead of regressing back to his prior mean in 2012, however, his UZR actually got even better, going up to +11.5, and again ranking third best among Major League shortstops. Among others, he rated ahead of Clint Barmes, Brandon Crawford, and Elvis Andrus, each of whom were essentially in the big leagues because of their defensive abilities. Again, the rating was seen as evidence that UZR wasn’t reliable, and Peralta was held up as an example of the system’s limitations. Perhaps playing next to Miguel Cabrera was throwing the system off, giving him credit for making plays that a better third baseman would have gotten to instead.

In 2013, Peralta’s UZR finally did regress some, though the system still thought he was an above average defender, putting him at +5 runs per 150 games. However, a 50 game suspension for his ties to Biogenesis led to the Tigers trading for Jose Iglesias, essentially ending his career as the Tigers shortstop; when he returned for the postseason, they experimented with using him in left field, because no one in their right mind would choose Peralta over Iglesias at shortstop, even though UZR was on year three of ranking Peralta as one of the better defenders in the game at the position.

With Iglesias in the fold, the Tigers bid adieu to Peralta this winter, not even making him a qualifying offer that could have returned a draft pick as compensation for letting him leave. With a poor defensive reputation and the Biogenesis connection hanging over his head, there were low expectations for what the market would offer him; MLBTradeRumors guessed 3/$30M, Jim Bowden guessed 2/$20M, and Jon Heyman guessed 2/$16M, with an agent and GM polled in that same piece only coming in slightly higher. Everyone basically agreed that he’d get something like $10 million a year for two or maybe three years.

Instead, the Cardinals gave him $53 million over four years. The Cardinals, one of baseball’s most respected organizations, bet big on Peralta’s ability to play shortstop. If he was actually a below average defender, or was going to need to move to another position in a couple of years, the deal would have been a ridiculous overpay. The only way to justify that price for Peralta is to argue that the defensive metrics were correct and that the public perception of Peralta’s defensive value was wrong.

So, fast forward to 2014. Peralta is no longer playing next to Cabrera, but is instead lined up besides one of the better defensive third baseman in baseball. He’s no longer playing behind the Tigers dominant pitching staff and weak-contact generators like in-his-prime Justin Verlander. He’s changed leagues, parks, and teammates, and yet again, he’s rated as the third best shortstop in baseball by UZR, second best if you use UZR/150.

Yes, Jhonny Peralta is rated higher by UZR this season than defensive wizard Andrelton Simmons. And that little factoid is an endless source of entertainment for people who want to again remind you to not put too much trust in defensive metrics. After all, no one actually believes that Peralta is a better defender than Simmons, so a result like this makes it easy to question the entire system.

Only the arguments for Peralta actually being a bad defensive shortstop are getting harder and harder to find.

Small sample size? Well, not anymore. We’re now on year four of Peralta being rated as an excellent defensive shortstop by UZR. Since the start of the 2010 season, he’s played over 4,600 innings at the position, and he has a UZR/150 of +10 runs per season over that stretch. Over the entirety of his career, spanning nearly 11,000 innings at shortstop, he’s rated as slightly above average. While defensive metrics absolutely do need larger sample sizes than offensive metrics, Peralta is well out of range of the small sample arguments.

The Cabrera factor? Well, UZR first rated him well while Cabrera was still playing first base, and his rating actually didn’t change much in the first year where he played next to a third baseman who could barely move. And now that he’s playing next to a much better defender, his defensive numbers have actually improved. There’s no real evidence for the idea that playing next to Cabrera artificially inflated Peralta’s defensive numbers.

The Tigers’ pitching staff? Well, nine players have played shortstop for the Tigers since the start of the 2010 season, and of those nine, only two of them have posted better defensive numbers than Peralta. That includes a few months of time from Jose Iglesias, who passes every eye test for elite defensive ability, and yet graded out slightly worse during his few months behind the Tigers pitchers than Peralta did during his tenure at the position. And playing behind a Cardinals pitching staff that has been decimated by injuries, resulting in a carousel of ever-changing arms who do not specialize in weak-contact, has not caused Peralta’s ratings to fall.

Peralta may or not be as much of a defensive asset as UZR claims; the model is still an educated guess, essentially, and it can certainly get things wrong. But the evidence is really starting to stack up against the Peralta-is-a-bad-defender idea. Even if you don’t buy into UZR at all, and you think that more granular measures that incorporate things like batted ball speed and launch angle would prove that UZR was overrating him, then you still have to explain how the Cardinals — one of the most aggressive teams in collecting and analyzing that kind of data — were willing to dramatically outbid the expected price for him as a free agent.

To believe that Jhonny Peralta is a poor defensive shortstop, you have to think that UZR is not just limited in value in smaller samples but completely useless regardless of the amount of data it has, while simultaneously believing that the Cardinals either don’t know how to evaluate defense or don’t know how to properly value free agents. Good luck defending those ideas against the mountains of evidence to the contrary.

So often, defensive data is assailed for not matching what the eye test leads us to believe. Peralta has been the poster boy for this disconnect, but at this point, there’s more evidence that the data was right and the eye test was wrong than vice versa. That doesn’t mean that the data will always be right whenever there’s a disconnect, or that we should just put full faith into single season defensive numbers, but perhaps Peralta can be a reminder that the eye test might be just as flawed as the metrics. There’s nothing wrong with being skeptical of defensive statistics; you should also be skeptical of things that are obviously true even when the data asserts otherwise, however.

UZR is imperfect. It will get things wrong. It might even still be getting Jhonny Peralta wrong, though that’s getting less likely by the day. But Jhonny Peralta should no longer be considered the poster boy for why UZR is unreliable. If anything, he should perhaps be the poster child for why we shouldn’t put unfailing trust in our own abilities to evaluate a player’s defensive value by watching him play.



Print This Post



Dave is the Managing Editor of FanGraphs.


Sort by:   newest | oldest | most voted
catswithbats
Guest
catswithbats
1 year 10 months ago

Peralta is the main reason I’m still skeptical of UZR. I can’t fathom him being a better defensive shortstop than Andrelton Simmons or Elvis Andrus. I feel lime my whole life has been a lie.

Brian
Guest
Brian
1 year 10 months ago

Peralta doesn’t have a better UZR career than Simmons. He just has a better UZR this year, the same way Lonnie Chisenhall and Seth Smith are outhitting Miguel Cabrera and Ryan Braun. So I don’t think UZR says Peralta is a “better” defensive shortstop than Adrelton Simmons.

I watch Peralta play almost every day, and I was extremely skeptical of his fielding at first. He just didn’t LOOK like a shortstop. But I gotta say, he’s starting to pass the eye test for me as well. He really is a vacuum. Not saying that’s any kind of be-all-end-all kind of argument, but the more I see him the more I agree with UZR. (Or am I merely seeing him in a new light because I knew his UZR was so high?)

tz
Guest
tz
1 year 10 months ago

Well, among active shortstops, Peralta is #4 in fielding percentage, behind Tulo, Rollins, and Hardy. I’d have to assume UZR dings you pretty bad for not making routine plays (errors or not), so his steadiness there probably offsets any shortcomings in range.

And lateral quickness doesn’t always coincide with running speed. Cal Ripken Jr. was very quick laterally, but not exactly the speediest baserunner.

Dre
Guest
Dre
1 year 10 months ago

agree with brian. He really makes a lot of plays. Not many of them are web gems but most plays that can be made are made. He has a really quick first step, and is usually is very good position. I also think he has great hands, making tough charging plays look easy. He’s better than Kozma, imo, and Kozma’s defensive numbers weren’t terrible.

shthar
Guest
shthar
1 year 10 months ago

‘This year’ is the one that counts.

Big in Japan
Guest
Big in Japan
1 year 8 months ago

I came on here to say the same thing. I’m a Cardinals fan who lives outside the blackout zone, with MLB.TV and a Roku. So I’ve probably watched 100+ games this season. And I watched Pete Kozma last season under similar circumstances. And I gotta say, Peralta gets to almost all the balls Kozma used to get, and he’s stronger on the web gem plays (e.g. flipping backhand to start a double play, ranging deep to the 2B side of the bag,) than Kozma ever was, plus his arm is way better than I thought it would be – definitely strong enough. His accuracy on long throws isn’t always perfect, but Matt Adams is a *wall* at 1st base, and you never see Peralta make a throw on a lost cause.

I think Peralta is showing the value of intelligent positioning, because he rarely dives, yet he always seems to get to the balls he should. And I think his values is enhanced by the Cardinals stellar infield defense – he can rush-bounce a throw to 1st, or play up the middle and leave Carpenter on an island, and he rarely has to hold the runner at second when Yadi is playing, due to the lack of steals. And Wong has been excellent at starting/finishing double plays, so we’re really strong throughout. That probably allows Peralta to position himself more aggressively, and get to more ground balls as a consequence.

If you had no idea what a shortstop was “supposed” to look like, and you watched Jhonny Peralta play, you’d see he’s obviously well-above average at the position. Better than Kozma, who was among the top 5 in the league last season, IIRC.

Zack Cozart
Guest
1 year 10 months ago

I am the best shortstop of 2014. No one fields better than me, and I’m even more clutch than Tulo. Eat my dust, Larkin!

Don’t believe me? Just look at the numbers.

Iron
Guest
Iron
1 year 10 months ago

Stupid sarcasm aside, Cozart is a very good defensive shortstop.

Matt
Guest
Matt
1 year 10 months ago

Someone should make a filter that checks your name against a list of all players and execs and deletes any comments that match. I would call it the “Deleting Shitty Jokes” filter.

Joe Morgan
Guest
Joe Morgan
1 year 10 months ago

Don’t be too hasty. I love reading and posting here.

Bob
Guest
Bob
1 year 10 months ago

The Amaro bot/really bored person is getting old though

joser
Guest
joser
1 year 10 months ago

You’re talking about the one that is making decisions in Philadelphia? A lot of fans there agree with you.

Mike
Guest
Mike
1 year 10 months ago

I’m not sure why so many think Peralta fails the eye test. I’ve watched a good amount of his defense, albeit this year only, and think he looks solid.

CFIC
Guest
1 year 10 months ago

yeah, he looks good by the eye test too. he just doesn’t have the standard shortstop look. he’s not the most rangey, but he’s really good.

Jhonny Manzana seed
Guest
Jhonny Manzana seed
1 year 10 months ago

I swear, it’s because he’s got fat cheeks. And viewers think people with fat cheeks are overweight. And people who are overweight likely aren’t good a shortstop. #eyetest

Bartolo Colon
Guest
Bartolo Colon
1 year 10 months ago

I get that all the time.

Carlos Williams Carlos
Guest
Carlos Williams Carlos
1 year 10 months ago

Do we really need a Colon joke after any mention of any player’s weight? Call me a killjoy but I’m over this one.

Pudge, Jeurys and Mexecutioner
Guest
Pudge, Jeurys and Mexecutioner
1 year 10 months ago

You’re over this joke like Colon is overweight?

Har har har

Mr Roboto
Guest
Mr Roboto
1 year 10 months ago

Killjoy was here.

leon
Guest
leon
1 year 10 months ago

Yeah, when we talk about the ‘eye test’ I think you have to make a distinction between, ‘I’ve watched this guy play a bunch of ball, and he looks good/bad’ and ‘this guy has/doesn’t have the body type I associate with an elite defender at this position.’ I think a big part of the knock on Peralta was how he looked. Jean Segura was a guy who got pigeonholed the opposite way. The book on him was the bat was fine, but he’d end up moving to second base because he’s too little. Turns out he can’t hit, but he plays shortstop just fine. He’s got the quickness and the arm, just because it’s in a 5’8″ package doesn’t mean he’s destined for second base.

Dan
Guest
Dan
1 year 10 months ago

I’m sorry, but I find the part about the Cardinals front office to be a weak argument. Half the articles on this site are centered around analyzing whether a team is making the right decision or not. Doesn’t it seem lazy, for lack of a better term, to pull the whole ‘Well, major league baseball front offices are smart, so we should trust them’ argument? Couldn’t you just use that every time a team makes a transaction?

HMS
Guest
HMS
1 year 10 months ago

You still see the argument trotted out that only neckbearded whirlyheads who don’t know anything about baseball look at advanced defensive metrics. Just the other day I read an article trying to claim that Billy Beane has just gotten lucky in the years when the A’s have done well, and Luhnow’s miserable three years in Houston are a true measure of what happens when major league teams base their decisions on data. I’d hope most reasonable people realize that all major league front offices look at so-called “advanced” statistics like UZR, but as long as sportswriters and fans are claiming otherwise it’s worth pointing out that a successful organization like the Cardinals looks at them and reaps the benefits.

CFIC
Guest
1 year 10 months ago

…when the Astros are in the position they are in because of adhering to traditional thinking

arc
Guest
arc
1 year 10 months ago

It’s a reasonable prior.

Thomas Grantham
Guest
Thomas Grantham
1 year 10 months ago

Dave,

Do you think Cabrera improved his defense or that he was simply playing under his true talent level?

Zach
Guest
Zach
1 year 10 months ago

I think in a situation like this, where the average scouting opinion contradicts the numbers, you have to use the numbers to inform the scouting. If Peralta has been an above-average or well above-average shortstop despite lacking in tools, the best scouts should be able to see that. The numbers aren’t magically generated, they’re reflecting real things that happen on the field.

With a guy like Peralta, it’s not good evaluation to just say that he’s a good shortstop because he puts up a high UZR every year. You have to use him as a test case to figure out why he’s putting up the defensive numbers that he is, and then figure out if he reflects a flaw in UZR or if he has a subtle skillset that allows him to significantly over-perform his tools.

I was a Tiger fan and watched Peralta play on a regular basis with my untrained scouting eye. What I saw was that he had a decently strong, very accurate arm that allowed him to play deep and make a play on just about every ball he got to. He always fielded a ball in a textbook position to make a strong throw. You wouldn’t ever see him put himself in a position where he had to make a spectacular throw like a Jeter or Asdrubal Cabrera. I think the average eye sees that as a detriment, but a good scout should recognize good footwork. I also believe he positioned himself more toward third than the average shortstop. I thought he covered a good amount of ground and made a play on nearly every ball going to his right, but he almost conceded anything more than a couple steps to his left.

I would have graded him as average or above-average, but would have a hard time believing metrics that say he’s better than some of the fundamentally sound and plus-tools guys that UZR has ranked above or equal to. I also think it’s possible that a team that likes to shift could get more defensive value out of him if a second baseman is able to cover more ground up the middle, but I don’t think that was happening with the Tigers. But my assessment may be (probably is) totally wrong, and I’d like to see a scouting breakdown from a professional scout.

Jim Price
Guest
Jim Price
1 year 10 months ago

Sorry I’m only an amateur definitely not pro scout. Its a bit misleading to say Peralta is lacking in tools. He has good balance, good hands, gets rid of the ball quickly and accurately. He is top of the scale on making the routine plays. He doesn’t look quick but he seems to get to every ball he should AND he’s in position to make a good throw. I think a good comp to Peralta is Jose Uribe, also doesn’t look like a plus defender but gets in done. He might still be with Tigers if he hadn’t forced them to find a replacement in the middle of a pennant race.

Matt
Guest
Matt
1 year 10 months ago

Looking at the Inside Edge numbers, I think his UZR is heavily influenced by making almost all of the easy plays. His numbers for “likely” and “routine” plays are better than Andrus, and almost identical to Simmons. But he’s bad at the plays that should be 50/50, and horrible at the ones that only Simmons can get to.

So not having watched him, looks like he’s great at the easy plays, but once it gets just a bit past his range, he never gets those. And since the vast majority of plays are routine ones, that’s likely why his advanced metrics show him as a solid plus defender, but since he never gets balls that are tough to field, nobody will every think of him as a good defender.

Zach
Guest
Zach
1 year 10 months ago

I think it brings up an interesting debate that either we put too much emphasis on range, or UZR is weighing the ability to make the routine play too highly and conventional thought wins out. It’s cases like Peralta that can help us really learn something about fielding.

AK7007
Member
AK7007
1 year 10 months ago

+1 vote for over-emphasis on range. That’s not to say that Jeter-esque range is a good thing, just that you don’t necessarily need to be the rangiest guy on your team in order to succeed. From my limited exposure to scout-speak, the language tends to filter players, and range is a way to filter players that are eligible to play short.

brad
Guest
brad
1 year 10 months ago

I think what it shows is that defensive value, like offensive value, can be compiled in many different ways. Making all the routine plays but no web gems might be like being a slappy mostly singles hitter, but more or less anyone has some value if they manage to hit .330.
Peralta is more or less the anti-Jeter, defensively. Jeter could make the spectacular play, but as a Yankee fan soon no longer hearing the phrase “past a diving Jeter” in every game will be very, very strange.

Famous Mortimer
Member
Member
1 year 10 months ago

Perhaps the number of “difficult” plays that go past him would be worth looking at. If his skill is in positioning himself so he just gets fewer difficult plays, then that’s a plus.

Andrew
Guest
Andrew
1 year 10 months ago

since he’s better at the routine plays, which happen more often, wouldn’t that make him a better defender? less range, but better defender. since when does range automatically equal a good defender?

bob
Guest
bob
1 year 10 months ago

Unless he is also worse at less routine plays (which by definition often require more range). Then the question is how good is UZR as a stat with one side arguing routine plays should be deemphasized relative to uzr’s use of them.

shthar
Guest
shthar
1 year 10 months ago

We’ve been hearing that argument for 100 years.

It’s usually put forward by fans of players who are getting older and slower.

CFIC
Guest
1 year 10 months ago

this is his value, the executes all the plays he SHOULD make. there is a certain value to that. not everyone does that, for whatever reason. Peralta seems to be cool as a cucumber.

Matt F.
Guest
Matt F.
1 year 10 months ago

Since the start of Inside Edge stats in 2012, Peralta is fourth among qualified SS in 90-100% balls, 5th in 60-90, second to last in 40-60, last in 10-40, and tied for last in 1-10. http://www.fangraphs.com/leaders.aspx?pos=ss&stats=fld&lg=all&qual=y&type=3&season=2014&month=0&season1=2012&ind=0&team=0&rost=0&age=0&filter=&players=0&sort=15,d

This seems to match what our eyes are telling us. He makes nearly all of the routine plays, and very few of the tough plays.

wade
Guest
wade
1 year 10 months ago

UZR may like Peralta, but Strat still gives him a “4” range rating at SS

Hiram Hover
Guest
Hiram Hover
1 year 10 months ago

I’m a little surprised you don’t make any reference to DRS as an alternate measure of his defense – a way to either confirm or qualify what UZR is seeing.

He ranked 4th (of 15) among qualifying SS in UZR, from 2011-13 with Detroit, but 9th in DRS, at a whopping +1.

This year, DRS loves him as much as UZR–he’s 2 or 3d by both measures.

Alvaro Pizza
Guest
Alvaro Pizza
1 year 10 months ago

DRS considers him a +15 SS this year and +19 in 4990 innings since 2009.

Total Zone considers him a +8 this year and +30 in 4990 since 2009.

UZR considers him a +9.4 SS this year and +37.2 since 2009

Who cares about the eye test.

tz
Guest
tz
1 year 10 months ago

I think the big thing is the “since 2009” part.

If you look at the UZR breakout, he’s always been very good at not making errors, and average-above average at turning the double play. But Peralta’s range component went from below-average while with Cleveland to slightly above average since then.

So either he’s gotten more agile over the years, or his positioning improved a lot since leaving the Indians.

Alvaro Pizza
Guest
Alvaro Pizza
1 year 10 months ago

I have not seen him that much, but I don’t think he is more agile in his 30’s than in his 20’s, so it has to be better positioning.

Rick
Guest
Rick
1 year 10 months ago

The reason I am skeptical of UZR is because it has Dustin Ackley as an elite left fielder, and better than Cespedes. I like Ackley’s defense, especially last night, but I gotta think Cespedes’s arm alone put him ahead of Ackley.

AK7007
Member
AK7007
1 year 10 months ago

Covered on these very electronic pages: http://www.fangraphs.com/blogs/yoenis-cespedes-run-saving-right-arm/

The differences in UZR values between the two are tiny, I don’t understand how it would make you skeptical. They are similarly valuable in the field, but arrive there differently. So, basically – learn more about UZR before you get skeptical.

Bip
Member
Member
Bip
1 year 10 months ago

How many outfield assists do even the best arms usually get? Not that many, I think. I don’t think a great arm adds all that much value, compared to something like outfield range.

Anon
Guest
Anon
1 year 10 months ago

My guess is the value from a great arm comes more from baserunners taking fewer extra bases rather than assists.

Alex Gordon
Guest
Alex Gordon
1 year 10 months ago

I beg to differ…

LHPSU
Guest
LHPSU
1 year 10 months ago

Cespedes has been known to take some pretty terrible routes though.

arc
Guest
arc
1 year 10 months ago

That is not a good reason.

Jason B
Guest
Jason B
1 year 10 months ago

Totally agreed. It’s akin to saying, “OBP tells me that Mike Napoli is a better hitter than Trout or Stanton. Ergo I am skeptical of OBP.”

Jeff Francoeur
Guest
Jeff Francoeur
1 year 10 months ago

Have you seen my arm?

chuckb
Guest
chuckb
1 year 10 months ago

Outfield arms are nowhere close to being as important as range. Good range, bad arm outfielders are much more valuable than bad range, good arm outfielders.

BRH
Guest
BRH
1 year 10 months ago

The following is not intended to be a jab at UZR:

Measuring “range” of an infielder involves comparing the total ground covered by Fielder A vs. Fielder B on a similarly batted ball. So lateral quickness is obviously important, but so is reaction time – getting a read off the ball. But when measuring “range” the position of the fielder when the ball is struck should not be relevant. Since UZR rewards a fielder for turning a batted ball into an out, and punishes a fielder for not turning a batted ball into an out, isn’t it safe to conclude that UZR may ultimately be rewarding the fielder more for his positioning rather than his range?

I believe Cal Ripken was woefully underappreciated as a defensive shortstop when he played. He basically had five years of Tony Fernandez and a decade of Omar Vizquel taking every Gold Glove. Having watched all of them play, I have no doubt that Fernandez and Vizquel had considerably more range than Ripken – meaning they both covered more lateral ground than Ripken could. But Ripken turned more balls in play into outs because he was phenomenal at positioning. And so Ripken was almost certainly a more valuable defensive player than the other two more flashy, quick, and athletic shortstops.

I can’t put Peralta in Ripken-in-his-prime territory, but I am suggesting that Peralta, through experience, and perhaps better advanced scouting, has become a better defensive player because he is more often in a better position to make plays. Defensively, Peralta may be Derek Jeter with a brain.

Vidor
Guest
Vidor
1 year 10 months ago

Cameron seems to be advancing two arguments here. One is “Peralta must be a good shortstop, because UZR has said so for a long time, and UZR couldn’t possibly be wrong for years.” The other one is “Peralta must be a good shortstop, because the Cardinals say so.” This is the weakest argument I have ever read on Fangraphs.

Brian
Guest
1 year 10 months ago

Maybe you should take a crack at reading the article again.

arc
Guest
arc
1 year 10 months ago

You are functionally illiterate.

Mike
Member
Mike
1 year 10 months ago

obviously he haxx UZR

B.J. Upton
Guest
B.J. Upton
1 year 10 months ago

UZR is a joke. Everyone knows I’m the best defensive CF in the game. And a great lead off man.

Fats
Guest
Fats
1 year 10 months ago

i think its because he basically looks fat. simple as that, not eye test for defense but eye test for his body type. its the reverse of speedy outfielders having bad defensive number, and people disagreeing with that because how could some so fast not have range?! how can someone who looks so fat have range?!

tz
Guest
tz
1 year 10 months ago

Which is hilarious simply because:

Peralta: 6’2 / 215
Trout: 6’2 / 230

I know these are listed weights, but I think Peralta is in deceptively good shape.

Cabrera
Guest
Cabrera
1 year 10 months ago

All weight is not created equally.

DNA+
Guest
DNA+
1 year 10 months ago

1. If there is a discrepancy between what people see and UZR, more UZR data can not be used to confirm the UZR data. You need something independent of UZR.

2. The fact that the savvy Cardinals bought into Peralta does not mean that the Cardinals buy into Peralta’s defense. It may be the case that the Cardinals, with their more complete knowledge, have decided that defense itself is currently overrated by the SABR crowd. Their valuation of Peralta may actually be primarily to do with his hitting.

3. Some teams shift on almost every batter these days (e.g. Yankees). UZR must be worse than useless for infielders on these teams.

DNA+
Guest
DNA+
1 year 10 months ago

I should clarify that my point 1. is only applicable if you think there is a problem with the UZR methodology and not the sample size.

21_22
Guest
21_22
1 year 10 months ago

i was thinking the same thing, self consistency of a method isnt validation of accuracy of that method. the biggest problem with UZR is that we have no way to understand how to

MLB Rainmaker
Member
Member
MLB Rainmaker
1 year 10 months ago

First, I’m not a fan of UZR. Models are only as good as their inputs and their assumptions, and given that much of the data for UZR is manually inputted and the data doesn’t account for defensive positioning, the error bars in the output are significant.

Specific to Peralta, it seems like his fielding has improved over time — Over about 3000 innings with the Tigers Peralta had 18 errors, in the same time with the Indians he had roughly three times that. Aside from the range argument, that is a compelling difference.

Regardless, it seems like you cherry-picked this argument to reinforce UZR as a defensive metric vs. writing thoughtfully about Peralta, which is disappointing.

Bloggo
Guest
Bloggo
1 year 10 months ago

This flies in the face of everything Fangraphs stands for, but based solely on the eye test, Peralta looks decent at SS. As a Cards fan, I kinda expected him to be a butcher based on what I had read of him in the offseason. Instead, I really haven’t noticed much of a downgrade from Kozma defensively, and his defense was literally the only reason he was in the majors.

I know nothing about how UZR is calculated, but based on my completely uneducated guess, Peralta is, in fact, an above average defensive SS (or at the very least, a completely competent one).

redsoxu571
Guest
redsoxu571
1 year 10 months ago

I think people are a bit off point here. First, you shouldn’t use UZR or UZR/150 to attempt to paint a complete defensive picture, just as you shouldn’t use any single stat or metric to evaluate anything in baseball. We also have DRS available to us…considering how little else there is that is conveniently out there, I find a combination of the two to be very helpful.

The bigger mistake, though, is a thought by people that a single outlier against perception invalidates a metric. I’ve placed a certain amount of trust in UZR ever since I saw that it rated Manny Ramirez as the worst fielder in the game; by my eyes and the information I’d taken in over time, he should be seen as such, and the system nailed it! If you look at the full rankings over many seasons, or even over a given season, you’ll see that UZR correlates strongly with perceived defensive abilities.

That doesn’t make it infallible, but it does say that we should spend more time investigating the outliers than we should questioning the metric as a whole. It’s good. It’s useful. The point of having reasoning, arguments, evidence, and so forth is that we aren’t slaves to the results spit out at us, but those results are mighty powerful. Use them.

DNA+
Guest
DNA+
1 year 10 months ago

I think UZR is worse than useful, I think it is probably misleading. It is used to both measure individuals fielders, and compare between fielders. Assuming you accept the methodology, we have no way to assess the size of the error bars, because they won’t tell us that information. We have no way to know what magnitude of difference in UZR is actually meaningful when comparing two players. Also, because we don’t know the data that goes into UZR we don’t know how any single player’s score came about. These are two leaps of faith that I see no reason to take.

MLB Rainmaker
Member
Member
MLB Rainmaker
1 year 10 months ago

^^Dead on

BenRevereDoesSteroids
Guest
BenRevereDoesSteroids
1 year 10 months ago

So I read the article, and I still don’t know what Jhonny Peralta tells us about defensive metrics.

Mike Green
Guest
Mike Green
1 year 10 months ago

This article is a bit of an oversimplification (and one that I don’t think MGL or John Dewan would be comfortable with). Peralta was measured at about -5/150G by DRS and at about -10/150G by UZR from 2005 to 2007. The consensus of the metrics was clear-he was a below average defensive shortstop.

From 2011 to 2013, Peralta was 0/150G by DRS and at about +10/150G by UZR. A 5 run improvement per 150 at Peralta’s age would not be typical but would hardly be extreme, but a 20 run improvement would not be. Do I think that Peralta is the Jose Cruz Sr. of shortstops? Probably not.

As for the 2014 data, DRS loves him even more than UZR but it’s a very small sample. Maybe the Cardinals have taught him something but until there is more data, I’d be inclined to say “sample size and leave it at that”.

In sum, in a case like this, you are better off to look at more than one metric.

Colin
Guest
Colin
1 year 10 months ago

Peralta is just evidence that perception of a player can carry through complete changes in that players ability. Peralta was known as a poor defensive SS, therefore he continued to be that in the eyes of many even when it wasn’t true anymore.

Frankly he has looked like a top defensive SS to me the last few years. He never makes a mistake, low errors, makes all the throws, is a total vacuum for balls he can get to. That itself has a ton of value that people are underrating.

Ben
Guest
Ben
1 year 10 months ago

Worth noting that MGL, who created UZR, has acknowledged that in light of all the shifting over the past few years, it is not as reliable a measure of infield defense as it once was.

As he said in a recent podcast, “To tell you the truth, because of the shifting that’s being done, UZR in the infield is becoming a little bit obsolete.”

This quote is at ~ 15:50 on the linked podcast, the discussion about the effect of shifts starts about 30 seconds earlier than that.

http://www.replacementlevelpodcast.com/2014/07/08/episode-36-mitchel-lichtman/

I’m not sure how much this affects the analysis, and I don’t want to mischaracterize what MGL was saying or put words in his mouth, so I apologize in advance if I’ve misunderstood.

But given that UZR throws out all shift plays, it seems that at least for the past two years as shifting has dramatically increased, we should take the UZR ratings of infielders with a grain of salt.

Ben
Guest
Ben
1 year 10 months ago

I just want to add that I trust UZR more than any other defensive metric we currently have, and really appreciate all the work MGL has done and continues to do to help people like me understand baseball.

He does a great job in this podcast explaining why some of the reasons that people mistrust UZR and other defensive metrics are misguided (such as, an individual can have a surprisingly bad or good fielding year, just as with hitting or pitching; we don’t discount OBP because tells us that Dee Gordon has been better at getting on base than Buster Posey so far this year).

isavage30
Guest
isavage30
1 year 10 months ago

I can buy Peralta as an average-ish shortstop because as they have always said, “he makes the easy plays.” And the majority of the time I saw him was with Cleveland, when UZR disliked him as well. But I find it hard to believe they he could ever be a well above average shortstop. He is not a quick person. When he was at 3b in Cleveland in 2010, I figured he was responsible for like a 1/4 of Justin Masterson’s ERA, seemed like anything hit weakly down the line was a hit. Maybe the working out and the steroids improved his range. But then, the best argument against all of this would be one that’s avoided in the long discussion, DRS, which continued to rate him as pretty much precisely an average shortstop (until this year, when it thinks he’s amazing.)

The Brantley comparison on offense in a small sample size was an interesting one, because Brantley is a guy who UZR seems to have a remarkable dislike for. Just as I can’t conceive of Peralta as an above-average SS, I can’t fathom that Brantley is somehow a below average left fielder, even though he moves well, seems to get good jumps, and has an accurate arm and plays balls off the wall well and nabs a lot of guys trying to stretch singles. But here we have a large sample size where UZR says Brantley is well below average, so maybe, as with Peralta, the eyes are wrong. Except, according to DRS, he’s been above average, which, again, matches the eye test. It’s totally unfair and not logical I know, but pretty much because of those two players, I trust DRS a lot more than UZR.

Famous Mortimer
Member
Member
1 year 10 months ago

Are you not just saying “this set of data agrees with my prejudices / perception, so I’m going to say this set of data is better”?

isavage30
Guest
isavage30
1 year 10 months ago

Yes, that’s exactly what I’m saying. Though I’m not sure that actually watching an irresponsible amount of Cleveland Indians baseball over the past 10 years and seeing these guys play is necessarily “prejudice”. Perception, yes.

tz
Guest
tz
1 year 10 months ago

PEDs help more than just pure brute strength, so maybe that’s part of Peralta’s transformation (assuming he had been using them for a while before he got caught).

UZR is a model that is built upon a number of assumptions that may have a large amount of “standard error” within their parameters. For example, velocity and spin of the batted ball can’t be accounted for, so two similarly classed balls hit in the same zone will be graded the same even though one could be a 90% chance of making the out and the other just a 20% chance.

I would look at multiple metrics like you have.

Also, a question. Does anybody know if there is a play-by-play tracker of how each defensive metric scores each play (UZR, DRS, +/-, Inside Edge probability, etc.)? It would be great to break down what causes the disparities in these metrics?

MrKnowNothing
Guest
MrKnowNothing
1 year 10 months ago

If he does things differently but succeeds, then so be it. And if UZR acknowledges that approach as working, so be it as well.

If a guy came to bat with his helmet on backwards, holding the bat from the wrong end, and swinging one handed but hit .400 with 60 homeruns, no one would care that it was “wrong” so long as it worked.

Burke
Guest
Burke
1 year 10 months ago

It’s hard to believe his range would improve so dramatically with age. I imagine there is a considerable body of evidence supporting the assertion that range decreases, rather than increases, with age. I would happily stack that evidence up against UZR; his alleged improvement in range is the only reason his defensive metrics have improved.

Seedless Plums
Guest
Seedless Plums
1 year 10 months ago

PEDs do help with quickness and endurance, I’ve heard.

Robert Hudgins
Member
1 year 9 months ago

I’ve watched him play all season. My eye test tells me he is a good shortstop. Perhaps many of the commenters, and the author, should watch him more. His range is indisputably subpar, but he is sure-handed, quick with his release, and positions himself well.

wpDiscuz