January 23, 2006
Probabilistic Model of Range, 2005, Shortstops
Update 1/25/2006: I discovered a data error that will likely change the values in these tables. Look for a new post with new data soon.
Update 1/25/2006: Corrected numbers for this post are here.
It's time to start looking at individual players, seeing how the regulars performed in terms of range in 2005. We'll start with the most important defensive position in terms of range, shortstop.
For each set of fielders, I'll present two models; the original based on all plays over the last four years, and another based on visiting players over the last four years. Players included were on the field for 1000 balls in play in 2005.
Probabilistic Model of Range, Shortstops, 2005, Original Model
Player | InPlay | Actual Outs | Predicted Outs | DER | Predicted DER | Difference |
Omar Infante | 1233 | 171 | 156.99 | 0.139 | 0.127 | 0.01136 |
Clint Barmes | 2209 | 272 | 253.34 | 0.123 | 0.115 | 0.00845 |
John McDonald | 1223 | 163 | 156.30 | 0.133 | 0.128 | 0.00548 |
Wilson Valdez | 1198 | 148 | 141.46 | 0.124 | 0.118 | 0.00546 |
Jason A Bartlett | 1766 | 253 | 244.34 | 0.143 | 0.138 | 0.00490 |
Julio Lugo | 4297 | 512 | 491.44 | 0.119 | 0.114 | 0.00478 |
Yuniesky Betancourt | 1426 | 166 | 159.83 | 0.116 | 0.112 | 0.00433 |
Adam Everett | 3748 | 468 | 456.62 | 0.125 | 0.122 | 0.00304 |
Bobby Crosby | 2163 | 282 | 276.69 | 0.130 | 0.128 | 0.00246 |
Alex Gonzalez | 3291 | 406 | 401.86 | 0.123 | 0.122 | 0.00126 |
Neifi Perez | 3026 | 411 | 407.35 | 0.136 | 0.135 | 0.00121 |
Juan Castro | 1775 | 236 | 234.83 | 0.133 | 0.132 | 0.00066 |
Rafael Furcal | 4111 | 520 | 521.88 | 0.126 | 0.127 | -0.00046 |
Miguel Tejada | 4280 | 531 | 535.03 | 0.124 | 0.125 | -0.00094 |
Jimmy Rollins | 3994 | 460 | 467.81 | 0.115 | 0.117 | -0.00195 |
David Eckstein | 4109 | 555 | 563.86 | 0.135 | 0.137 | -0.00216 |
Omar Vizquel | 4024 | 493 | 502.44 | 0.123 | 0.125 | -0.00235 |
Jack Wilson | 4240 | 538 | 551.57 | 0.127 | 0.130 | -0.00320 |
Oscar M Robles | 1313 | 158 | 162.38 | 0.120 | 0.124 | -0.00334 |
Juan Uribe | 3946 | 492 | 505.26 | 0.125 | 0.128 | -0.00336 |
Cesar Izturis | 2859 | 343 | 353.35 | 0.120 | 0.124 | -0.00362 |
Orlando Cabrera | 3706 | 426 | 442.40 | 0.115 | 0.119 | -0.00442 |
Khalil Greene | 3123 | 363 | 378.13 | 0.116 | 0.121 | -0.00484 |
Cristian Guzman | 3605 | 378 | 395.59 | 0.105 | 0.110 | -0.00488 |
J.J. Hardy | 2805 | 309 | 325.75 | 0.110 | 0.116 | -0.00597 |
Russ M Adams | 3433 | 377 | 398.48 | 0.110 | 0.116 | -0.00626 |
Michael Young | 4398 | 499 | 528.59 | 0.113 | 0.120 | -0.00673 |
Jhonny Peralta | 3736 | 469 | 495.40 | 0.126 | 0.133 | -0.00707 |
Bill Hall | 1447 | 174 | 185.81 | 0.120 | 0.128 | -0.00816 |
Carlos Guillen | 1934 | 230 | 246.47 | 0.119 | 0.127 | -0.00852 |
Angel Berroa | 4438 | 502 | 541.85 | 0.113 | 0.122 | -0.00898 |
Edgar Renteria | 4119 | 416 | 454.05 | 0.101 | 0.110 | -0.00924 |
Jose Reyes | 4308 | 479 | 521.31 | 0.111 | 0.121 | -0.00982 |
Felipe Lopez | 3804 | 413 | 451.25 | 0.109 | 0.119 | -0.01006 |
Royce Clayton | 3711 | 421 | 458.43 | 0.113 | 0.124 | -0.01009 |
Mike Morse | 1437 | 141 | 155.56 | 0.098 | 0.108 | -0.01013 |
Marco Scutaro | 1980 | 229 | 255.87 | 0.116 | 0.129 | -0.01357 |
Derek Jeter | 4231 | 472 | 548.13 | 0.112 | 0.130 | -0.01799 |
Probabilistic Model of Range, Shortstops, 2005, Smoothed Visiting Player Model
Player | InPlay | Actual Outs | Predicted Outs | DER | Predicted DER | Difference |
Omar Infante | 1233 | 171 | 155.90 | 0.139 | 0.126 | 0.01225 |
Clint Barmes | 2209 | 272 | 248.91 | 0.123 | 0.113 | 0.01045 |
John McDonald | 1223 | 163 | 153.31 | 0.133 | 0.125 | 0.00792 |
Jason A Bartlett | 1766 | 253 | 241.27 | 0.143 | 0.137 | 0.00664 |
Julio Lugo | 4297 | 512 | 487.18 | 0.119 | 0.113 | 0.00578 |
Wilson Valdez | 1198 | 148 | 142.28 | 0.124 | 0.119 | 0.00478 |
Yuniesky Betancourt | 1426 | 166 | 159.80 | 0.116 | 0.112 | 0.00435 |
Adam Everett | 3748 | 468 | 455.93 | 0.125 | 0.122 | 0.00322 |
Neifi Perez | 3026 | 411 | 402.01 | 0.136 | 0.133 | 0.00297 |
Bobby Crosby | 2163 | 282 | 277.48 | 0.130 | 0.128 | 0.00209 |
Alex Gonzalez | 3291 | 406 | 400.12 | 0.123 | 0.122 | 0.00179 |
Juan Castro | 1775 | 236 | 233.75 | 0.133 | 0.132 | 0.00127 |
Miguel Tejada | 4280 | 531 | 531.18 | 0.124 | 0.124 | -0.00004 |
Rafael Furcal | 4111 | 520 | 522.12 | 0.126 | 0.127 | -0.00052 |
Jimmy Rollins | 3994 | 460 | 463.44 | 0.115 | 0.116 | -0.00086 |
David Eckstein | 4109 | 555 | 559.91 | 0.135 | 0.136 | -0.00119 |
Juan Uribe | 3946 | 492 | 501.43 | 0.125 | 0.127 | -0.00239 |
Cesar Izturis | 2859 | 343 | 350.28 | 0.120 | 0.123 | -0.00255 |
Omar Vizquel | 4024 | 493 | 504.01 | 0.123 | 0.125 | -0.00273 |
Khalil Greene | 3123 | 363 | 373.38 | 0.116 | 0.120 | -0.00332 |
Jack Wilson | 4240 | 538 | 553.77 | 0.127 | 0.131 | -0.00372 |
Orlando Cabrera | 3706 | 426 | 440.25 | 0.115 | 0.119 | -0.00384 |
Oscar M Robles | 1313 | 158 | 163.05 | 0.120 | 0.124 | -0.00385 |
Cristian Guzman | 3605 | 378 | 395.22 | 0.105 | 0.110 | -0.00478 |
J.J. Hardy | 2805 | 309 | 323.34 | 0.110 | 0.115 | -0.00511 |
Russ M Adams | 3433 | 377 | 395.73 | 0.110 | 0.115 | -0.00546 |
Michael Young | 4398 | 499 | 523.59 | 0.113 | 0.119 | -0.00559 |
Angel Berroa | 4438 | 502 | 529.85 | 0.113 | 0.119 | -0.00627 |
Jhonny Peralta | 3736 | 469 | 492.80 | 0.126 | 0.132 | -0.00637 |
Bill Hall | 1447 | 174 | 183.23 | 0.120 | 0.127 | -0.00638 |
Edgar Renteria | 4119 | 416 | 448.71 | 0.101 | 0.109 | -0.00794 |
Mike Morse | 1437 | 141 | 153.16 | 0.098 | 0.107 | -0.00846 |
Carlos Guillen | 1934 | 230 | 247.31 | 0.119 | 0.128 | -0.00895 |
Jose Reyes | 4308 | 479 | 521.47 | 0.111 | 0.121 | -0.00986 |
Felipe Lopez | 3804 | 413 | 454.10 | 0.109 | 0.119 | -0.01080 |
Royce Clayton | 3711 | 421 | 462.03 | 0.113 | 0.125 | -0.01106 |
Marco Scutaro | 1980 | 229 | 255.28 | 0.116 | 0.129 | -0.01327 |
Derek Jeter | 4231 | 472 | 546.66 | 0.112 | 0.129 | -0.01765 |
Omar Infante is the poster child for a defensive replacement. He has no offensive value, but can flash the leather. Cristian Guzman, who was near the top of the list last year, fell off in 2005. His ability to catch line drives did not hold up, indicating his good range rating was somewhat lucky.
Derek Jeter is back at the bottom of the list after a better showing in 2004. Why the Yankees keep him at that position when there's a better player to his right is beyond me. These numbers also show that the Red Sox will be making a significant defensive upgrade if they end up replacing Renteria with Gonzalez.
As always, comments are welcome.
Answer: Because if you watch Derek Jeter everyday,
you know that it would be a mistake to move him. I
hesitated to even respond to this because these kinds of
studies used against Jeter do little else except generate
tons of malicious comments by self-aggrandizing people,
some who put all their time into knocking Jeter. I
heard a guy today from BIS talking about how they
need to improve how they measure defensive performance. If someone wants to give me the list of
games and plays that show how utterly astounding it is
that he hasn't been moved, please email me, and I'd be
happy to discuss it further. But, you'll have no case,
and you should really move on to something else.
I've never felt Jeter was the best defensive shortstop. But after watching him for a decade, I can't say he hurts the team. Maybe that's not saying much. But with all the other things he brings to the table, I think that's acceptable. Perhaps baseball people put too much into observation and not enough into stats. But find me the stat that would predict, for instance, that Peyton Manning is a choke artist (other than playoff-game rating, and losing big games even in high school, obviously). But we all know he's as good a lock as anyone to fail in the big spot.
I'm getting off track, and not trying to make the Captain Intangibles argument.
But I do agree that A-Rod was a better option when he came to the Yankees. Now, though? Wouldn't playing two years at third base throw off your range and instincts at such a demanding position as shortstop? No doubt, A-Rod is as good an athlete as they come. But once the mistake, if we're calling that in this context, has been made, isn't it too late to undo that?
Joe Torre is nothing if not loyal, the ultimate players' manager.
And no players benefit more from that loyalty than the core of the the Yankees' WS champs -- Jeter, Williams, Posada, Rivera.
He stayed with Bernie in CF long past the point where it was obvious he couldn't handle it any more.
And it's clear since moving A-Rod to 3B that he's more interested in protecting Jeter's ego than he is in what makes the most defensive sense for the team.
The "I watch him play everyday" argument must be the lamest (and most often relied upon by those with no facts to back up their opinions) argument in all of sports. I am SOOO sick of hearing it, especially in regards to Jeter. Do you think the year to year correlation between his incredibly crappy defensive ratings represents random chance?
A format suggestion:
I'd find it very helpful to see a column showing outs minus predicted outs. And a rate version of that would be great too, such as +/- Outs per G or per 150 G.
On the other side, I don't find the DER results very helpful at the individual player level. If you could show DER based on balls hit into that player's zone, that would be very interesting. But the percentage of team BIP that one player turns into outs is not, to me, an intuitive concept. (But it's obviously very subjective.)
David,
When you break down the numbers, could you also break them down by "left side of the infield" and "right side of the infield?" On a number of teams, people argue that sometimes the 1b or the 3b take balls where they should/shouldn't instead of the 2b and SS. It might be useful to know how a team's left and right sides of the infield are doing against average.
All of the top 5 shortstops and 6 of the top 7 had less than 300 actual or predicted outs. None of these guys was a full time SS either secondary to injury, time in the minors or being a backup. This screams sample size issues and probably indicates a need for a higher cutoff for significance. A comparison between past season results might be helpful. Of course 4 of those 7 guys are rookies. Which of course makes it even more unlikely to be acccurate. That 4 of the top seven fielding SS are rookies would be unlikely. Otherwise great stuff!
I had a question about the "Predicted DER". Is it predicted based on... that particular player's past performance? Or is that predicted based on what an average SS in the same situation would do? Or something else entirely? I'm (relatively) new to these tables, and don't quite understand the finer points of them yet...
Can anyone direct me to where to learn more?
On further reflection, the easy rate stat to create is just actual/predicted outs. Infante is 110% and Jeter 86% -- pretty easy to interpret.
Every day you spend watching Jeter at SS means that you aren't watching every other SS. Which means that you can't establish a baseline for good SS defense. Would you argue that Jeter is the best offensive shortstop, because you've seen him hit every day? No. You would compare his stats to the competition and build your argument. So why would your approach be any different for defense?
Granted defensive statistics are IMO still in their infancy. But when every defensive metric says that the guy is bad defensively, there might be something there. I think Jeter's always been a little underrated offensively and overrated defensively. He's still a lock for the Hall. So what's the problem?
That said, it'll be fun watching him play on days when Tiger Wang is on the mound.
Rate2 is a pretty well accepted defensive stat, and Jeter had a 104 Rate2, which is pretty good, and definately above average. He had a Rate2 of 99 in 2004, and all his other seasons were well below that. If you want to chalk it up to ARod helping him out, then go ahead, but it's not like Jeter has played with only chumps at 3B since he started in 1996 (Brosius anybody?). Maybe the guy has actually improved over the past couple year. Who knows?
Yeah, yeah, yeah. Torre is loyal to "his guys" to a fault. The Yankees have also won their division all those years too (except '97 when they were the Wild Card). That sounds pretty good to me, especially since the only other team with such success these past years has only one-fourth as many WS rings.
Plus, where do you move Jeter? A-Rod would be an upgrade (defensively) at short, but then Derek would be a downgrade at third. And all the other positions are filled. (Though the Yankees are going to need a new catcher soon).
I would have moved him to center before they signed Damon.
Anyone find it sad that Jeter is dead last in this metric, yet he won the gold glove. I'm sure this metric isn't perfect, but still, it's pathetic that he won. That award is such a joke.
Great stuff, David! To whoever quoted Rate2, while PMR with GBs only is a *much better* stat, Rate2 is pretty much crap. It has no value when you have something like PMR, or Range.
David, I think you're being a little too harsh when you say that Rate2 is pretty much crap. I have a hard time seeing BP using it so prominently if there wasn't any value to it. That said, without a peek behind the curtain, there's really no way to know exactly how useful it is, nor how to compare it with PMR.
That 5 of the top seven of the list have a lower number of outs is not saying much, but looking at who those guys are is telling. Scouts and analysts (especially Peter Gammons) LOVE Johnny McDonald's glovework, and other analysts have been raving about Lugo, Barmes and Betancourt. No one ever hears about Infante because he's playing in baseball purgatory (it's okay, I'm from the 'D). That shows me that if anything their bats (or in Barmes' case, an 8-pointer and a staircase) are keeping them from getting the chances in the field. The corrolary that their rates are high because they don't get enough chances is wrong, IMO.
I actually believe that Jeter would we an upgrade over ARod at third, just as I believe ARod would be an upgrade over Jeter at short (though not as much as 2 years ago; ARod has slowed a bit more than Jeter). Defensive play at the two positions is not inherently comparable; they require different skills.
Obviously, there's no statistical data to test this, since Jeter hasn't played third.