November 05, 2008
Probabilistic Model of Range, 2008, Shortstops
The first position to examine is the most important of the fielders working behind the pitcher, the shortstop. As a reference, the first table looks at the position on a team-wide basis:
Team Shortstop PMR, 2008, Visit Smooth Distance Model, 2008 data only
Team | In Play | Actual Outs | Predicted Outs | DER | Predicted DER | Ratio |
Brewers | 4354 | 551 | 526.00 | 0.127 | 0.121 | 104.75 |
Giants | 4232 | 492 | 469.81 | 0.116 | 0.111 | 104.72 |
Marlins | 4338 | 517 | 501.00 | 0.119 | 0.115 | 103.19 |
Angels | 4374 | 524 | 510.64 | 0.120 | 0.117 | 102.62 |
Cardinals | 4597 | 580 | 566.37 | 0.126 | 0.123 | 102.41 |
Red Sox | 4232 | 480 | 472.55 | 0.113 | 0.112 | 101.58 |
Phillies | 4396 | 557 | 551.85 | 0.127 | 0.126 | 100.93 |
Braves | 4383 | 566 | 561.40 | 0.129 | 0.128 | 100.82 |
Diamondbacks | 4224 | 469 | 465.74 | 0.111 | 0.110 | 100.70 |
Cubs | 4156 | 498 | 495.33 | 0.120 | 0.119 | 100.54 |
Astros | 4292 | 500 | 497.32 | 0.116 | 0.116 | 100.54 |
Athletics | 4285 | 477 | 474.54 | 0.111 | 0.111 | 100.52 |
Rangers | 4667 | 538 | 536.31 | 0.115 | 0.115 | 100.31 |
Dodgers | 4265 | 546 | 544.65 | 0.128 | 0.128 | 100.25 |
Indians | 4513 | 542 | 540.73 | 0.120 | 0.120 | 100.23 |
White Sox | 4409 | 548 | 546.97 | 0.124 | 0.124 | 100.19 |
Royals | 4413 | 508 | 507.30 | 0.115 | 0.115 | 100.14 |
Rays | 4264 | 490 | 490.56 | 0.115 | 0.115 | 99.89 |
Orioles | 4540 | 537 | 539.06 | 0.118 | 0.119 | 99.62 |
Rockies | 4535 | 587 | 589.61 | 0.129 | 0.130 | 99.56 |
Pirates | 4683 | 577 | 580.37 | 0.123 | 0.124 | 99.42 |
Blue Jays | 4215 | 476 | 479.71 | 0.113 | 0.114 | 99.23 |
Twins | 4607 | 578 | 584.70 | 0.125 | 0.127 | 98.85 |
Yankees | 4349 | 491 | 499.00 | 0.113 | 0.115 | 98.40 |
Nationals | 4417 | 526 | 538.02 | 0.119 | 0.122 | 97.76 |
Mariners | 4512 | 480 | 493.64 | 0.106 | 0.109 | 97.24 |
Padres | 4419 | 520 | 536.39 | 0.118 | 0.121 | 96.94 |
Reds | 4299 | 468 | 485.83 | 0.109 | 0.113 | 96.33 |
Tigers | 4536 | 519 | 544.40 | 0.114 | 0.120 | 95.33 |
Mets | 4335 | 498 | 524.64 | 0.115 | 0.121 | 94.92 |
Notice the Mets are last. While recent articles mention Jeter as the worst fielder in the majors, in 2008 he wasn't the worst shortstop in New York:
Individual Shortstop PMR, 2008, Visit Smooth Distance Model, 2008 data only (1000 balls in play)
Player | In Play | Actual Outs | Predicted Outs | DER | Predicted DER | Ratio |
Marco Scutaro | 1352 | 173 | 157.04 | 0.128 | 0.116 | 110.16 |
Omar Vizquel | 1863 | 210 | 193.60 | 0.113 | 0.104 | 108.47 |
Mike Aviles | 2277 | 271 | 252.42 | 0.119 | 0.111 | 107.36 |
Maicer Izturis | 1307 | 151 | 144.80 | 0.116 | 0.111 | 104.28 |
Jed Lowrie | 1142 | 123 | 118.01 | 0.108 | 0.103 | 104.23 |
J.J. Hardy | 3804 | 477 | 460.72 | 0.125 | 0.121 | 103.53 |
Erick Aybar | 2437 | 305 | 295.72 | 0.125 | 0.121 | 103.14 |
Alex Cora | 1137 | 140 | 135.75 | 0.123 | 0.119 | 103.13 |
Cesar Izturis | 3136 | 408 | 396.19 | 0.130 | 0.126 | 102.98 |
Jack Wilson | 2231 | 285 | 276.88 | 0.128 | 0.124 | 102.93 |
Bobby Crosby | 3740 | 423 | 411.71 | 0.113 | 0.110 | 102.74 |
Jason Bartlett | 3208 | 380 | 372.71 | 0.118 | 0.116 | 101.96 |
Hanley Ramirez | 3986 | 472 | 462.98 | 0.118 | 0.116 | 101.95 |
Juan Castro | 1331 | 153 | 150.32 | 0.115 | 0.113 | 101.78 |
Jimmy Rollins | 3537 | 451 | 443.43 | 0.128 | 0.125 | 101.71 |
Luis Rodriguez | 1191 | 143 | 140.93 | 0.120 | 0.118 | 101.47 |
Yunel Escobar | 3344 | 440 | 434.04 | 0.132 | 0.130 | 101.37 |
Nick Punto | 1646 | 227 | 224.05 | 0.138 | 0.136 | 101.32 |
Orlando Cabrera | 4218 | 527 | 521.06 | 0.125 | 0.124 | 101.14 |
Adam Everett | 1183 | 156 | 154.33 | 0.132 | 0.130 | 101.08 |
Miguel Tejada | 4062 | 472 | 469.63 | 0.116 | 0.116 | 100.50 |
Jhonny Peralta | 3963 | 469 | 467.52 | 0.118 | 0.118 | 100.32 |
Michael Young | 4165 | 489 | 487.98 | 0.117 | 0.117 | 100.21 |
Ryan Theriot | 3615 | 425 | 424.27 | 0.118 | 0.117 | 100.17 |
Julio Lugo | 1947 | 216 | 217.95 | 0.111 | 0.112 | 99.11 |
Angel Berroa | 1730 | 225 | 227.09 | 0.130 | 0.131 | 99.08 |
Derek Jeter | 3815 | 429 | 433.24 | 0.112 | 0.114 | 99.02 |
Stephen Drew | 3820 | 422 | 429.34 | 0.110 | 0.112 | 98.29 |
Cristian Guzman | 3640 | 441 | 449.15 | 0.121 | 0.123 | 98.19 |
John McDonald | 1387 | 150 | 154.82 | 0.108 | 0.112 | 96.89 |
Yuniesky Betancourt | 4173 | 446 | 460.45 | 0.107 | 0.110 | 96.86 |
Troy Tulowitzki | 2730 | 354 | 365.56 | 0.130 | 0.134 | 96.84 |
Edgar Renteria | 3696 | 428 | 449.40 | 0.116 | 0.122 | 95.24 |
Jose Reyes | 4196 | 480 | 504.15 | 0.114 | 0.120 | 95.21 |
Khalil Greene | 2841 | 327 | 345.54 | 0.115 | 0.122 | 94.64 |
Tony F Pena | 1808 | 199 | 211.52 | 0.110 | 0.117 | 94.08 |
Brendan Harris | 1480 | 159 | 170.68 | 0.107 | 0.115 | 93.16 |
Jeff Keppinger | 2636 | 274 | 296.50 | 0.104 | 0.112 | 92.41 |
David Eckstein | 1445 | 149 | 163.53 | 0.103 | 0.113 | 91.11 |
Jose Reyes converted 24 fewer balls into outs than expected while Jeter was just down four. It was actually one of Derek's better years.
Omar Vizquel remains impressive at an advanced age. He didn't play a whole season, so he didn't get a chance to wear down, but if any team is looking for a great late inning defensive replacement, Omar is it.
Mike Aviles came up as a great find for the Royals in terms of his batting, but he also performed well with the glove.
Troy Tulowitzki went from first to almost worst in 2008. His injuries seemed to limit his range.
Jimmy Rollins won the Gold Glove today, but J.J. Hardy deserved it based on this data.
Feel free to comment and criticize.
Yet another data point for the Royals being idiotic to move Aviles off shortstop.
Michael Young wasn't his usual self with the bat, but he wasn't bad defensively either.
Tejada was average defensively, which will help him sustain his value as his bat declines.
Is there anyone who can point me in the direction of a comprehensive explanation for why this is the best way to rank fielders?
I don't claim it's the best. First, I'm only measuring range. Second, like any small sample size statistic, it's subject to random variation.
What is does try to do is take into account how difficult or easy it is for a ball to be fielded.
So, given your last statement, you are stating that this is better data than the fielding bible. Back it up. Convince me that Hardy is better than Rollins. Because according to this data Hanley is better than Rollins?!?!? We all know that isn't true. But give an explanation. I'm open to debate, and I'm not tied to one guy. Just saying: back it up.
Where would Everett have placed?
"Notice the Mets are last. While recent articles mention Jeter as the worst fielder in the majors, in 2008 he wasn't the worst shortstop in New York:"
So does that mean the Mets are a minor league team?
(sorry, couldn't resist).
No doubt Marco Scutaro is the best SS in MLB. Reminding me of the year Kevin Millar came in as the best fielding 1B.
Alex Cora's rank among the top ten is surprising to me. Much of the consensus that I've heard among fans and media is that he's lost a step in the field.
I agree with Mike. I'm not doing this to be argumentative, but I am demanding of the information that I use to evaluate players. I would like to see some explanation for how this is any better than anecdotal evidence.
The data is the same data that the fielding Bible uses, the data supplied by Baseball Info Solutions. I build a probability model based on six parameters of a ball in play. The are direction, distance, type (ground, fly, etc) batter hand, pitcher hand, and park. I then ask, for each ball in play, based on those paramenters, what is the probability of that ball being turned into an out. Adding all those up gives me expected outs.
What make my system different:
1) Based on reader suggestions, I eliminate home team fielders, since they can have too much of an impact on the model. (There is a smoothing factor that includes them.)
2) I treat the home field as part of the model, rather than applying park factors to a larger model.
I hope this answers your questions.
I'm not at all lazy, Thumble. You don't have to try to be such a troll about it. That was uncalled for.
I just think that if someone is going to do a systematic breakdown, it would make everything more accessible if they linked to an explanation for the stats. I also want to thank David for the info that he supplied. I am always eager to learn more about the various metrics.
"So, given your last statement, you are stating that this is better data than the fielding bible. Back it up."
David stated no such thing. How in the world do you get that from his statement "I don't claim it's the best."?
For an explanation of the system, click on the PMR link in the navigation box above. David has been doing this for 4? years now.
Question:
When you are calcing the PMR for specific positions, do you use the DER of that position in a given zone or the overall DER for that zone (to determine the expected outs)?
Thanks.
There really isn't a zone in this system. There are slices of the field, and in some cases distances, but it's the DER of the fielder for the six parameters.
Right, so an individual shortstop is compared to the probability of all SS making a play in vector 40 (or whatever number)? Not compared to all outs made in vector 40.
Jason,
The author does link to an explanation via the tag at the end of the article. All the information you needed is readily accessed using the tags or the archives located on the right side of the website.
It probably took you longer to post your comment/request than it would have taken to retrieve the information yourself. If that behavior doesn't fall within the boundary of intellectually lazy then we should remove the word from our lexicon.