The Shortstop Position: Weighing the value of defense.

keninten · May 11, 2014

Unless they want to sign Drew, who else could step in at SS this year or even next. He might be there until Marrero gets his shot or a trade is made.

joe dokes · May 11, 2014

teddywingman said:
Watching Xander Bogaerts this year--I keep thinking about this nebulous idea--let me see if I can frame it with some clarity.

He's got talent crackling off his spit, so he probably has the ability to become an average defensive shortstop; but he's certainly not one yet.

The Red Sox seem to be all-in with this year's hand. Let Bogaerts "marinate" in the show, and see once and for all if WMB can lay off outside shit and smash 30.

After watching last year's team, and the noodle bat tandem of Iggy and Drew, I have become a believer that winning in baseball's current landscape hinges upon making every possible out, an out. Last year's team was incredible at this; the best defensive Sox team of my lifetime.

This year's team is an interesting contrast: as we've seen from Victorino's absence, the early Sizemore-in-center experiment, spotty fill-ins at 3rd, and anyone but Napoli at first, plus the X-man himself.

It's early and all: for the 2014 team and the young career of Xander Bogaerts. But does he really have to be a shortstop?

My oppinion--and I'm looking for numbers and sharper observations from folks who disagree--but a strong SS who manages to field 1 extra out per game might be worth as much or more than a poor fielding SS---even if they're raking like Teddy Ballgame.

Alternate title: Bogaerts: Lugo two, electric boogaloo

Devil's advocate (mostly because I'm never sure how to shake out the offense/defense equation)....in this depressed offensive environment, a SS who will hit so much better than all other SS (at least that's the idea) only has to play roughly average defense. 10 years ago, 850 OPS players were a dime a dozen, so you could afford a real noodle bat with defense. As offense becomes scarce, maybe the strong bat becomes more playable. I also think that with Middlebrooks and Victorino back and Bradley seemingly installed as the #1, non job-shared CF, the overall defnse isn;t as cringe-worthy.

bosockboy · May 11, 2014

Drew was FAR from a noodle bat.

NDame616 · May 11, 2014

teddywingman said:
Fine. But what was his average though 16 post season games?

Anyway, it's not really the point.

SOME might call this cherry picking....

His OPS was 4th in the AL and his slugging was 3rd in the AL.

Was he knocking on the door of being our cleanup hitter? No. But throughout last season he was a respectable hitter for an AL SS.

FinanceAdvice · May 11, 2014

As a dilitante on true baseball knowledge but for the love of the game, let me put in my two cents. Far be it from me to argue with the scouts and management for they know infifinitely more than I and can see him (XMan) on a full time basis. However, I felt at start of season and even now that they should have re-signed Drew and moved XMan to third and try to move WMB (perhaps hard to do) or put him in Pawtucket. My reasons.

Look at Drews's numbers in 2013 vs. ALL SS in AL

Avg .253 12th
OBP .333 6th
SLG .443 3rd
OPS .777 4th
HR 13 4th

Secondly, Bogaerts seems to be more of a fit on 3rd. at short he doesnt seem to have the range. I read some articles that projectd him at 3rd base because of his growing size. Although small sample for this year, Bogaerts is not that far from WMD' numbers.
Bogaerts line of .252 .364 .336 .700 with 1 HR
WMB line of .211 .338 .368 .707 with 2 HR

So for me youd have an upgrade defensively at 3rd and short over the situation currently and youd have more of an offense, but I grant you that Sox may lose some power without WMB.

SoxJox · May 11, 2014

"who manages to field 1 extra out per game"

So, this got me to thinking...how likely is that?

So I put this together:

[There is supposed to be a table here, but the table generator over here doesn't seem to be liking it].

Edit: So, in any case, the table summarizes fielding stats for SSs who won WS the last 25 years.

They averaged 4.2 outs per game (1.49 Putouts + 2.79 Assists). None of them led the league...or were even close to being "top" defensive SSs - at least by this measure.

The league leaders over the same time span averaged ~~5.69~~ 5.75 (updated table below) outs per game.

So to manage fielding 1 extra out per game would represent about a 25% increase from an historical perspective and would put them in close company with the league leaders - most of whom have not made it to the WS.

I know, blah blah blah about sample size, yada yada. Not intended to be a thorough analysis. Just a quick simple look.

Also, interesting to note that only 2 Gold Glove SS won a WS in the past 25 years - Rollins in 2008 and Jeter in 2009.

SoxJox · May 11, 2014

teddywingman said:
That's interesting. When I threw out the extra out per game I was thinking that it was probably an exageration.

How about comparing SS who's team reached the postseason vs. those who didn't?

That would require looking at SSs for 96 teams making the playoffs the last 25 years, and the 652 SSs who didn't make it over the same time span (accounting for Marlins joining in '93, Diamondbacks and Rays joining in '98, and no playoffs in 1994.

I think I'll beg off from that one. :smithicide:

Edit: I should say "primary" shortstop, recognizing that few play every game.

Ed Hillel · May 11, 2014

Xander is not a major league caliber SS.

LuckyBen · May 11, 2014

Ed Hillel said:
Xander is not a major league caliber SS.

Step away from the ledge, he's a rookie.

Ed Hillel · May 11, 2014

LuckyBen said:
Step away from the ledge, he's a rookie.

No ledge, just a statement of fact for now. I do believe he's ultimately a corner infielder, though. Range isn't really a learned skill in the infield.

LargeFather · May 11, 2014

Ed Hillel said:
No ledge, just a statement of fact for now. I do believe he's ultimately a corner infielder, though. Range isn't really a learned skill in the infield.

This may depend on WMB development. Not a stretch to project X to 3rd (Cecchini notwithstanding) and Marrero as SS

EricFeczko · May 11, 2014

Ed Hillel said:
Xander is not a major league caliber SS.

Poor transfers and bad range on slow grounders up the middle are more common than you think.

In terms of what he's done so far, he's actually somewhere between below-average and average defensively in the majors (-3 DRS, -1.1 UZR, .968 FP).

Unless you think half of starting SS in the league are not major league caliber....

EricFeczko · May 11, 2014

LargeFather said:
This may depend on WMB development. Not a stretch to project X to 3rd (Cecchini notwithstanding) and Marrero as SS

Additionally, X hasn't finished growing yet. If he gets bigger physically, his already below-average range may be further reduced.
I hope I'm wrong, but I think WMB has finished development. Through 186 games, he's the definition of a league-average third baseman. He accomplishes average through lots of strikeouts, poor BABIP, and a lot of power.

EDIT: I mean bigger physically, not better.

SoxJox · May 11, 2014

Here's the table (had to split in 3), after appreciated help from Rev (and even then it wouldn't display until I put it in 3 separate posts) [tablegrid= Shortstops Winning World Series in Last 25 Years (Table 1) ] Player Age Pos G GS CG Inn Ch PO A E DP Fld% Rtot Rdrs Rtot/yr Rdrs/yr RF/9 RF/G lgFld% lgRF9 lgRFG Awards Outs/G 2013 Boston Red Sox (97-65, AL) Stephen Drew 30 SS 124 122 118 1093.1 516 176 332 8 82 0.984 -3 -2 -3 -2 4.18 4.1 0.973 4.15 4.13 4.10 2012 San Francisco Giants (94-68, NL) Brandon Crawford 25 SS 139 122 107 1101 607 195 394 18 74 0.97 8 12 8 13 4.81 4.24 0.97 4.26 4.21 4.24 2011 St. Louis Cardinals* (90-72, NL) Ryan Theriot 31 SS 91 87 75 755 388 117 254 17 54 0.956 -10 -8 -16 -12 4.42 4.08 0.971 4.41 4.4 4.08 2010 San Francisco Giants (92-70, NL) Juan Uribe 31 SS 103 96 85 864 373 134 233 6 38 0.984 -3 -5 -4 -7 3.82 3.56 0.971 4.31 4.27 3.56 2009 New York Yankees (103-59, AL) Derek Jeter 35 SS 150 147 121 1260.2 554 206 340 8 75 0.986 4 3 4 3 3.9 3.64 0.972 4.36 4.31 GG 3.64 2008 Philadelphia Phillies (92-70, NL) Jimmy Rollins 29 SS 132 132 127 1168 593 193 393 7 71 0.988 5 18 5 18 4.52 4.44 0.976 4.43 4.4 GG 4.44 2007 Boston Red Sox (96-66, AL) Julio Lugo 31 SS 145 139 129 1228.1 593 214 360 19 70 0.968 0 0 0 0 4.21 3.96 0.97 4.47 4.42 3.96 2006 St. Louis Cardinals (83-78, NL) David Eckstein 31 SS 120 119 106 1029 547 178 363 6 87 0.989 9 4 10 5 4.73 4.51 0.972 4.41 4.38 4.51 2005 Chicago White Sox (99-63, AL) Juan Uribe 26 SS 146 143 135 1293.1 687 249 422 16 98 0.977 8 6 7 5 4.67 4.6 0.972 4.6 4.55 4.60 2004 Boston Red Sox* (98-64, AL) Pokey Reese 31 SS 71 57 40 507.2 281 85 190 6 37 0.979 17 12 39 29 4.88 3.87 0.972 4.56 4.53 3.87 2003 Florida Marlins* (91-71, NL) Alex Gonzalez 26 SS 150 150 140 1315.2 679 237 426 16 109 0.976 1 -3 1 -3 4.54 4.42 0.973 4.52 4.4 4.42 [/tablegrid]

SoxJox · May 11, 2014

[tablegrid= Shortstops Winning World Series in Last 25 Years (Table 2) ] Player Age Pos G GS CG Inn Ch PO A E DP Fld% Rtot Rdrs Rtot/yr Rdrs/yr RF/9 RF/G lgFld% lgRF9 lgRFG Awards Outs/G 2002 Anaheim Angels* (99-63, AL) David Eckstein 27 SS 147 146 132 1276.1 616 205 397 14 91 0.977 10 9 4.24 4.1 0.975 4.56 4.51 4.10 2001 Arizona Diamondbacks (92-70, NL) Tony Womack 31 SS 118 114 101 1013.2 485 152 311 22 74 0.955 9 11 4.11 3.92 0.968 4.47 4.42 3.92 2000 New York Yankees (87-74, AL) Derek Jeter 26 SS 148 148 136 1278.2 610 237 349 24 78 0.961 -23 -21 4.12 3.96 0.973 4.71 4.66 3.96 1999 New York Yankees (98-64, AL) Derek Jeter 25 SS 158 158 154 1395.2 635 230 391 14 87 0.978 -11 -10 4 3.93 0.97 4.73 4.66 3.93 1998 New York Yankees (114-48, AL) Derek Jeter 24 SS 148 148 140 1304.2 625 223 393 9 82 0.986 2 2 4.25 4.16 0.972 4.71 4.66 4.16 1997 Florida Marlins* (92-70, NL) Edgar Renteria 20 SS 153 149 145 1328.2 674 242 415 17 95 0.975 -3 -3 4.45 4.29 0.973 4.58 4.54 4.29 1996 New York Yankees (92-70, AL) Derek Jeter 22 SS 157 156 149 1370.2 710 244 444 22 83 0.969 -14 -12 4.52 4.38 0.971 4.66 4.64 4.38 [/tablegrid]

SoxJox · May 11, 2014

[tablegrid= Shortstops Winning World Series in Last 25 Years (Table 3) ] Player Age Pos G GS CG Inn Ch PO A E DP Fld% Rtot Rdrs Rtot/yr Rdrs/yr RF/9 RF/G lgFld% lgRF9 lgRFG Awards Outs/G 1995 Atlanta Braves (90-54, NL) Jeff Blauser 29 SS 115 114 104 997 503 151 337 15 62 0.970 2 3 4.41 4.24 0.968 4.51 4.49 4.24 1993 Toronto Blue Jays (95-67, AL) Tony Fernandez 31 SS 142 142 135 1235.1 702 279 410 13 90 0.981 3 3 5.02 4.85 0.97 4.69 4.65 4.85 1992 Toronto Blue Jays (96-66, AL) Manuel Lee 27 SS 128 126 111 1080 525 187 331 7 67 0.987 12 13 4.32 4.05 0.969 4.71 4.69 4.05 1991 Minnesota Twins (95-67, AL) Greg Gagne 29 SS 137 127 88 1067.2 567 181 377 9 69 0.984 8 9 4.7 4.07 0.971 4.67 4.66 4.07 1990 Cincinnati Reds (91-71, NL) Barry Larkin 26 SS 156 154 141 1344 740 254 469 17 86 0.977 10 9 4.84 4.63 0.967 4.59 4.56 4.63 1989 Oakland Athletics (99-63, AL) Mike Gallego 28 SS 94 85 58 730.1 421 152 255 14 68 0.967 -2 -3 5.02 4.33 0.971 4.79 4.74 4.33 Average 26.84 126.88 123.24 111.08 1081.30 545.24 188.84 343.44 12.96 73.08 0.976 1.56 2.60 2.38 2.82 4.43 4.20 4.20 [/tablegrid]

EricFeczko · May 11, 2014

teddywingman said:
It's been clearly established that UZR samples of this small of a sample have no real value. In fact even a full season isn't considered to be definitive.

To my eye, he's well below average, consistently.

*facepalm*

This isn't the right place for a long discussion on the difference between "reliability", "validity", and "measurement". However, your statement confounds the three. Samples of this small have no value in predicting future value. In other words UZR or DRS over a 30 game sample is not stable nor representative of the true talent of a player. However, they do measure what a player did over the course of 31 games, and do so in very different ways. By three separate measures, X has been below-average to average relative to his peers, which suggests that he has not been a travesty a shortstop.

I'll agree that one's eye is a better measure of true talent at this point, but the eye is biased in many ways. I'll agree to disagree with your particular eye here.

absintheofmalaise · May 11, 2014

teddywingman said:
It's been clearly established that UZR numbers of this small of a sample have no real value. In fact even a full season isn't considered to be definitive.

To my eye, he's well below average, consistently.

You do understand that what we are currently seeing is what he would normally be doing on an AAA field. Growing pains aren't always fun to watch, but it you read the other thread on XB on the main board, you will see that pretty much everyone who has experience with this sort of thing seems to understand that this is all part of him learning to play the position at the MLB level.

EricFeczko · May 11, 2014

absintheofmalaise said:
You do understand that what we are currently seeing is what he would normally be doing on an AAA field. Growing pains aren't always fun to watch, but it you read the other thread on XB on the main board, you will see that pretty much everyone who has experience with this sort of thing seems to understand that this is all part of him learning to play the position at the MLB level.

This. What people are seeing by eye needs to be adjusted because memory is biased towards the outliers: really bad plays and really amazing plays.
He's not the type of defensive player that will make amazing plays on a regular basis, he is young enough that he will make some really bad ones (tonight being a prime example). However, I can count about 15-16 games where he's made every play to SS that he should make.

EricFeczko · May 11, 2014

teddywingman said:
It's been clearly established that UZR numbers of this small of a sample have no real value. In fact even a full season isn't considered to be definitive.

To my eye, he's well below average, consistently.

I was replying to someone who said that he isn't a ML-caliber SS. To which, I responded that he's been below-average to average.

teddywingman said:
All I said is that he's been below average but that his UZR numbers don't mean shit. Both of which are facts.

Except neither are correct facts. I'll agree that, worst-case, he's been below average, however you didn't say that at all. As I keep trying to inform you, UZR/DRS/FP numbers aren't reliable at SSS, but they are still valid measurements of what a player has done.

SoxJox · May 11, 2014

[tablegrid= Shortstops: League Leaders in Outs/Game in Last 25 Years ]Putouts Assists NL AL NL AL 2013 Andrelton Simmons (ATL) 240 Pedro Florimon (MIN) 245 2013 Andrelton Simmons (ATL) 499 Alexei Ramirez (CHW) 433 2012 Starlin Castro (CHC) 266 J.J. Hardy (BAL) 244 2012 Starlin Castro (CHC) 465 J.J. Hardy (BAL) 529 2011 Starlin Castro (CHC) 267 Alcides Escobar (KCR) 271 2011 Starlin Castro (CHC) 446 Alcides Escobar (KCR) 459 2010 Ian Desmond (WSN) 221 Yuniesky Betancourt (KCR) 256 2010 Brendan Ryan (STL) 430 Alexei Ramirez (CHW) 499 2009 Hanley Ramirez (FLA) 221 Elvis Andrus (TEX) 261 2009 Miguel Tejada (HOU) 475 Orlando Cabrera (2TM) 428 2008 Hanley Ramirez (FLA) 236 Orlando Cabrera (CHW) 242 2008 Miguel Tejada (HOU) 442 Orlando Cabrera (CHW) 472 2007 Troy Tulowitzki (COL) 262 Jhonny Peralta (CLE) 249 2007 Troy Tulowitzki (COL) 561 Jhonny Peralta (CLE) 452 2006 Rafael Furcal (LAD) 269 Orlando Cabrera (LAA) 252 2006 Rafael Furcal (LAD) 492 Michael Young (TEX) 492 2005 Rafael Furcal (ATL) 255 Julio Lugo (TBD) 310 2005 Jack Wilson (PIT) 523 Miguel Tejada (BAL) 479 2004 Cesar Izturis (LAD) 234 Derek Jeter (NYY) 273 2004 Jack Wilson (PIT) 492 Miguel Tejada (BAL) 526 Jack Wilson (PIT) 2003 Orlando Cabrera (MON) 258 Angel Berroa (KCR) 264 2003 Rafael Furcal (ATL) 481 Miguel Tejada (OAK) 490 Cesar Izturis (LAD) 2002 Juan Uribe (COL) 261 Alex Rodriguez (TEX) 259 2002 Juan Uribe (COL) 504 Miguel Tejada (OAK) 504 2001 Rich Aurilia (SFG) 246 Alex Rodriguez (TEX) 279 2001 Orlando Cabrera (MON) 514 Alex Gonzalez (TOR) 509 Orlando Cabrera (MON) 2000 Neifi Perez (COL) 288 Royce Clayton (TEX) 265 2000 Neifi Perez (COL) 523 Miguel Tejada (OAK) 501 1999 Neifi Perez (COL) 260 Miguel Tejada (OAK) 292 1999 Neifi Perez (COL) 481 Mike Bordick (BAL) 511 1998 Neifi Perez (COL) 272 Omar Vizquel (CLE) 273 1998 Neifi Perez (COL) 516 Mike Bordick (BAL) 446 1997 Edgar Renteria (FLA) 242 Nomar Garciaparra (BOS) 249 1997 Royce Clayton (STL) 452 Derek Jeter (NYY) 457 1996 Barry Larkin+ (CIN) 230 Kevin Elster (TEX) 285 1996 Jay Bell (PIT) 478 Mike Bordick (OAK) 476 1995 Royce Clayton (SFG) 223 Mike Bordick (OAK) 245 1995 Royce Clayton (SFG) 411 John Valentin (BOS) 414 Jose Vizcaino (NYM) 1994 Barry Larkin+ (CIN) 178 Greg Gagne (KCR) 189 1994 Jay Bell (PIT) 380 Gary Disarcina (CAL) 358 1993 Jay Bell (PIT) 256 Mike Bordick (OAK) 280 1993 Jay Bell (PIT) 527 Cal Ripken+ (BAL) 495 1992 Jay Bell (PIT) 268 Cal Ripken+ (BAL) 287 1992 Jay Bell (PIT) 526 Gary Disarcina (CAL) 486 1991 Shawon Dunston (CHC) 261 Cal Ripken+ (BAL) 267 1991 Jay Bell (PIT) 491 Cal Ripken+ (BAL) 528 1990 Jay Bell (PIT) 260 Tony Fernandez (TOR) 297 1990 Barry Larkin+ (CIN) 469 Tony Fernandez (TOR) 480 1989 Kevin Elster (NYM) 235 Cal Ripken+ (BAL) 276 1989 Ozzie Smith+ (STL) 483 Cal Ripken+ (BAL) 531 Total Putouts 6209 6610 12819 Total Assists 12061 11955 Average Games 126 126 126 126 # of Players 27 25 27 25 Total Games 3402 3150 6552 3402 3150 Average Putout/Game 1.96 Average Assist/Game 3.80 5.75 [/tablegrid]

SoxJox · May 11, 2014

One thing that is kinda interesting given the implied question of how important a good defensive SS might be: the average number of games started = 127. That means you have someone else starting a good 22% of your games. Granted, they're probably "capable" infielders, but that's a pretty hefty chunk of games being played by someone other than your starter.

EricFeczko · May 11, 2014

teddywingman said:
Also facepalmer. Who gives a shit what we discuss in a game thread after the game has ended.

And no, a thirty game sample of UZR does not consistently give you an accurate assessment of what a player actually did over any 30 game stretch, because of the limitations of the UZR stat itself.

That's fine if you want one.

I'm assuming you're referring to this: http://www.hardballtimes.com/tht-live/how-reliable-is-uzr/

The study there uses a pearson's correlation coefficient on normally transformed seasonal UZR data to determine whether UZR is predictive from one year to the next*. All this tells us is whether UZR is reliable; does prior UZR predict future UZR?

We can do this for any statistic: FP, DRS, or hitter statistics like batting average.

That being said. These are still measuring how a player is fielding defensively, and therefore have value. An analogous example would be batting average. Batting average across 30 games is a valid measure of how well a player has hit in those 30 games. What it doesn't provide us is an expectation of what the player will hit in the next 30 games.

By the three measures we have, X has played below-average (FP,DRS) to average (UZR) defensively across 34 games. To my eye, this matches what I've seen across the 34 games: generally average play in the majority of the games and one or two poor plays in each of the remainder.

EDIT: As a side note, I should probably add that this method for measuring reliability likely overestimates the stability, because it doesn't properly separate the within-subject variance from the between-subject variance. Since I've never seen someone do a study using a statistic that does (e.g. chronbach's alpha), one should take all stability measures with a grain of salt.

SoxJox · May 11, 2014

I thought I'd have a bit more fun with this.

The record # of Putouts in a season is held by Donie Bush (Washington Senators) in 1914: 425. They did not make the WS.

The record # of Assists in a season is held by Ozzie Smith (St. Louis Cardinals) in 1980: 621. They did not make the WS.

If limiting to the same "last 25 years" in my analysis:

The highest # of Putouts in a season is held by Julio Lugo (Tampa Bay [then] Devil Rays) in 2005: 310. They did not make the WS.

The highest # of Assists in a season is held by Troy Tulowitzki (Colorado Rockies) in 2007: 561. They made the WS, but lost to...guess who?

EricFeczko · May 11, 2014

teddywingman said:
No. Batting average would not be an analogous example.

Why not? In what way are defensive statistics invalid?

Reverend · May 11, 2014

Teddy, he's not speaking about the true talent of the X-man, just what has been recorded to this point.

That's the disconnect, I think,

EricFeczko · May 11, 2014

teddywingman said:
From FanGraphs:

UZR uses Baseball Info Solutions (BIS) data in calculating its results. It’s important to note that this data is compiled by human scorers, which means that it likely includes some human error.

also:

Beware of sample sizes! If a player only spent 50 innings at a position last season, it’d be a good idea not to draw too many conclusions from their UZR score over that time. Like with any defensive statistic, you should always use three years of UZR data before trying to draw any conclusions on the true talent level of a fielder.

And this is why I'm facepalming. You don't seem to get the difference between what a player has done and the expectations of what a player will do in the future.
If you go back and read my posts, you'll see that I make no conclusions regarding Xander's true talent whatsoever. Someone (Ed Hillel, I think) said that X is not a major league shortstop, and I pointed out that the evidence does not support such a conclusion based on what Xander has done so far. Your eye may be completely correct and Xander will make more and more mistakes as he continues to play at SS, in which case the statistics will reflect his poor play. He may even end the year with the worst UZR at shortstop, and then have the best UZR at shortstop next year because what a player does defensively fluctuates greatly from year to year.

EricFeczko · May 11, 2014

Reverend said:
Teddy, he's not speaking about the true talent of the X-man, just what has been recorded to this point.

That's the disconnect, I think,

Yup.

EricFeczko · May 11, 2014

teddywingman said:
Watching Xander Bogaerts this year--I keep thinking about this nebulous idea--let me see if I can frame it with some clarity.

He's got talent crackling off his spit, so he probably has the ability to become an average defensive shortstop; but he's certainly not one yet.

The Red Sox seem to be all-in with this year's hand. Let Bogaerts "marinate" in the show, and see once and for all if WMB can lay off outside shit and smash 30.

After watching last year's team, and the noodle bat tandem of Iggy and Drew, I have become a believer that winning in baseball's current landscape hinges upon making every possible out, an out. Last year's team was incredible at this; the best defensive Sox team of my lifetime.

This year's team is an interesting contrast: as we've seen from Victorino's absence, the early Sizemore-in-center experiment, spotty fill-ins at 3rd, and anyone but Napoli at first, plus the X-man himself.

It's early and all: for the 2014 team and the young career of Xander Bogaerts. But does he really have to be a shortstop?

My oppinion--and I'm looking for numbers and sharper observations from folks who disagree--but a strong SS who manages to field 1 extra out per game might be worth as much or more than a poor fielding SS---even if they're raking like Teddy Ballgame.

Alternate title: Bogaerts: Lugo two, electric boogaloo

This is a very interesting thread. A few points:
1) When you write, "making every possible out", and "1 extra out per game", I assume you mean as a function of opportunities. In that case, using the number of assists and putouts is not a great measure to use. I suggest either using Bill James' range factor (for simplicity), or using zone-rating.

2) The fact that the teams with the largest number of outs aren't WS contenders may reflect several other factors: poor offense independent of the shortstop, low strikeout counts resulting in a larger number of opportunities. A better approach may be to examine the year-to-year variance in shortstop performance, and how it covaries with winning percentage.

3) Looking at the contribution of an extra out to RAR may be a simpler way of estimating how much offense is needed to compensate for a fielder who is capable of making one extra out.

4) Love the alternate title, but I feel this thread is covering a broader topic than just the Red Sox, maybe this should be moved to the MLB forum?

kieckeredinthehead · May 11, 2014

Reverend said:
Teddy, he's not speaking about the true talent of the X-man, just what has been recorded to this point.

That's the disconnect, I think,

EricFeczko said:
And this is why I'm facepalming. You don't seem to get the difference between what a player has done and the expectations of what a player will do in the future.
If you go back and read my posts, you'll see that I make no conclusions regarding Xander's true talent whatsoever. Someone (Ed Hillel, I think) said that X is not a major league shortstop, and I pointed out that the evidence does not support such a conclusion based on what Xander has done so far. Your eye may be completely correct and Xander will make more and more mistakes as he continues to play at SS, in which case the statistics will reflect his poor play. He may even end the year with the worst UZR at shortstop, and then have the best UZR at shortstop next year because what a player does defensively fluctuates greatly from year to year.

It's true that a month and a half of UZR data is not predictive, but for this particular metric, it's also not a good measure of what has already happened. The issue with UZR is that it makes best guesses as to what plays should be made. Measurements are made like "soft grounders 10 ft to the left of 2b by a slow runner with 2 outs and none on in Fenway have been handled 90% of the time by AL SS." If Bogaerts gets one of those plays, all of a sudden he's above average in that situation. It would only be comparable to batting average if hits were predicted based on average outcomes from pitch speed, location, and count.

Note that UZR is then scaled to an average player for this season, not the full six year sample, which means it's even worse as a measurement for comparing how Bogaerts has done over the past month and half versus the rest of the league, since measurements of other shortstops are also unreliable to this point.

UZR is not a measure of what happened. It's a measure of what probably happened. It's not saying "Bogaerts just threw six heads in a row. Even if the coin isn't weighted, that's a pretty good run." It's more akin to saying, "people who toss heads smile 95% of the time after they do, and Bogaerts has smiled six times in a row." DRS is measured differently but based on the same principle. Unlike batting average, there is significant measurement error involved; in other words, there's variance around what has already happened. Bogaerts may have been playing like a below-average shortstop, or as a terrible one.

derekson · May 11, 2014

The thing that's been missed in this discussion is that the standard of "an extra play/out per game" is a level of ability difference on defense that just doesn't actually exist. You'll see guys with a range factor that much higher in a given season, but that's more a function of opportunities than ability and range. A single is worth just under half a run. If you could get a SS that made 1 more play per game (in the same opportunities), you're looking at gaining something like 70 runs on defense in 150 games over the SS you're comparing him to. That kind of gulf just doesn't exist.

SoxJox · May 11, 2014

Playing out EricFeczko's suggestion:

[tablegrid= A Twist - Using RF/G and RF/9 as Lurker EricFeczko Suggests ]RF/9 RF/G RF/9 RF/G Year NL RF/9 AL RF/9 NL RF/G AL RF/G WS Winner RF/9 RF/G DiffNL DiffAL DiffNL DiffAL 2013 Andrelton Simmons (ATL) 4.92 Pedro Florimon (MIN) 5.29 Andrelton Simmons (ATL) 4.74 Pedro Florimon (MIN) 4.86 Stephen Drew 4.18 4.1 -0.74 -1.11 -0.64 -0.76 2012 Brandon Crawford (SFG) 4.81 Brian Dozier (MIN) 4.89 Starlin Castro (CHC) 4.51 J.J. Hardy (BAL) 4.89 Brandon Crawford 4.81 4.24 0.00 -0.08 -0.27 -0.65 2011 Troy Tulowitzki (COL) 5.05 J.J. Hardy (BAL) 4.88 Troy Tulowitzki (COL) 4.84 J.J. Hardy (BAL) 4.76 Ryan Theriot 4.42 4.08 -0.63 -0.46 -0.76 -0.68 2010 Troy Tulowitzki (COL) 5.06 Cliff Pennington (OAK) 4.93 Troy Tulowitzki (COL) 4.91 Alexei Ramirez (CHW) 4.79 Juan Uribe 3.82 3.56 -1.24 -1.11 -1.35 -1.23 2009 Brendan Ryan (STL) 5.41 Cesar Izturis (BAL) 4.89 Brendan Ryan (STL) 4.75 Elvis Andrus (TEX) 4.61 Derek Jeter 3.9 3.64 -1.51 -0.99 -1.11 -0.97 2008 Troy Tulowitzki (COL) 5.22 Erick Aybar (LAA) 4.77 Troy Tulowitzki (COL) 4.96 Orlando Cabrera (CHW) 4.43 Jimmy Rollins 4.52 4.44 -0.70 -0.25 -0.52 0.01 2007 Troy Tulowitzki (COL) 5.39 John McDonald (TOR) 4.98 Troy Tulowitzki (COL) 5.31 Jhonny Peralta (CLE) 4.61 Julio Lugo 4.21 3.96 -1.18 -0.77 -1.35 -0.65 2006 Craig Counsell (ARI) 5.17 Jhonny Peralta (CLE) 4.90 Rafael Furcal (LAD) 4.88 Michael Young (TEX) 4.73 David Eckstein 4.73 4.51 -0.44 -0.17 -0.37 -0.22 2005 Rafael Furcal (ATL) 5.23 Julio Lugo (TBD) 4.93 Rafael Furcal (ATL) 4.99 Julio Lugo (TBD) 4.71 Juan Uribe 4.67 4.6 -0.56 -0.26 -0.39 -0.11 2004 Jack Wilson (PIT) 4.82 Miguel Tejada (BAL) 4.99 Jack Wilson (PIT) 4.65 Bobby Crosby (OAK) 4.94 Pokey Reese 4.88 3.87 0.06 -0.11 -0.78 -1.07 2003 Adam Everett (HOU) 4.96 Ramon Santiago (DET) 4.84 Rafael Furcal (ATL) 4.63 Julio Lugo (TBD) 4.68 Alex Gonzalez 4.54 4.42 -0.42 -0.30 -0.21 -0.26 2002 Juan Uribe (COL) 5.23 Mike Bordick (BAL) 5.08 Juan Uribe (COL) 4.94 Mike Bordick (BAL) 4.86 David Eckstein 4.24 4.1 -0.99 -0.84 -0.84 -0.76 2001 Jack Wilson (PIT) 4.87 Rey Sanchez (KCR) 5.16 Neifi Perez (COL) 4.70 Alex Gonzalez (TOR) 4.92 Tony Womack 4.11 3.92 -0.76 -1.05 -0.78 -1.00 2000 Neifi Perez (COL) 5.20 Felix Martinez (TBD) 5.67 Neifi Perez (COL) 5.01 Felix Martinez (TBD) 5.27 Derek Jeter 4.12 3.96 -1.08 -1.55 -1.05 -1.31 1999 Mike Benjamin (PIT) 5.40 Rey Sanchez (KCR) 5.53 Neifi Perez (COL) 4.72 Rey Sanchez (KCR) 5.18 Derek Jeter 4.00 3.93 -1.40 -1.53 -0.79 -1.25 1998 Neifi Perez (COL) 5.12 Kevin Stocker (TBD) 4.99 Neifi Perez (COL) 4.86 Miguel Tejada (OAK) 4.81 Derek Jeter 4.25 4.16 -0.87 -0.74 -0.70 -0.65 1997 Walt Weiss (COL) 5.22 Benji Gil (TEX) 5.13 Walt Weiss (COL) 4.73 Pat Meares (MIN) 4.67 Edgar Renteria 4.45 4.29 -0.77 -0.68 -0.44 -0.38 1996 Rey Sanchez (CHC) 5.37 Alex Gonzalez (TOR) 5.09 Rey Sanchez (CHC) 4.98 Alex Gonzalez (TOR) 5.06 Derek Jeter 4.52 4.38 -0.85 -0.57 -0.60 -0.68 1995 Royce Clayton (SFG) 4.88 Benji Gil (TEX) 5.18 Royce Clayton (SFG) 4.66 Benji Gil (TEX) 4.88 Jeff Blauser 4.41 4.24 -0.47 -0.77 -0.42 -0.64 1993 Rey Sanchez (CHC) 5.60 Tony Fernandez (TOR) 4.99 Ozzie Smith+ (STL) 5.24 Tony Fernandez (TOR) 4.85 Tony Fernandez 5.02 4.85 -0.58 0.03 -0.39 0.00 1992 Ozzie Smith+ (STL) 5.07 Greg Gagne (MIN) 5.07 Jay Bell (PIT) 4.99 Andy Stankiewicz (NYY) 4.81 Manuel Lee 4.32 4.05 -0.75 -0.75 -0.94 -0.76 1991 Barry Larkin+ (CIN) 5.22 Omar Vizquel (SEA) 5.13 Barry Larkin+ (CIN) 5.03 Cal Ripken+ (BAL) 4.91 Greg Gagne 4.70 4.07 -0.52 -0.43 -0.96 -0.84 1990 Barry Larkin+ (CIN) 4.84 Alvaro Espinoza (NYY) 5.33 Barry Larkin+ (CIN) 4.63 Dick Schofield (CAL) 4.93 Barry Larkin 4.84 4.63 0.00 -0.49 0.00 -0.30 1989 Barry Larkin+ (CIN) 5.33 Tony Fernandez (TOR) 5.32 Barry Larkin+ (CIN) 4.99 Tony Fernandez (TOR) 5.25 Mike Gallego 5.02 4.33 -0.31 -0.30 -0.66 -0.92 [/tablegrid]

Seems to me the same conclusion can be reached. The numbers generally are all lower- and in most cases substantially lower than league leaders in RF/G and RF/9. In fact, there is not a SINGLE WS winner who led either league in RF/G or RF/9.

absintheofmalaise · May 11, 2014

I just moved a bunch of posts from the game thread into this thread. It would be a good idea to read through here again if you weren't following the discussion in the GT.

SoxJox · May 11, 2014

derekson said:
The thing that's been missed in this discussion is that the standard of "an extra play/out per game" is a level of ability difference on defense that just doesn't actually exist. You'll see guys with a range factor that much higher in a given season, but that's more a function of opportunities than ability and range. A single is worth just under half a run. If you could get a SS that made 1 more play per game (in the same opportunities), you're looking at gaining something like 70 runs on defense in 150 games over the SS you're comparing him to. That kind of gulf just doesn't exist.

In absolute numbers, if you're talking about ALL SSs, I would agree. But that was not the point I was making with my summary - which was limited to the context of shortstops winning WS, and for which you clearly can see that they DID have at least a chance of improving by an extra play/out per game. They averaged 4.2 outs/game. League leaders averaged 5.75.

Now, whether they had the ABILITY...different discussion, I admit.

Edit: "chance" = "mathematical probability"

Reverend · May 11, 2014

kieckeredinthehead said:
It's true that a month and a half of UZR data is not predictive, but for this particular metric, it's also not a good measure of what has already happened. The issue with UZR is that it makes best guesses as to what plays should be made. Measurements are made like "soft grounders 10 ft to the left of 2b by a slow runner with 2 outs and none on in Fenway have been handled 90% of the time by AL SS." If Bogaerts gets one of those plays, all of a sudden he's above average in that situation. It would only be comparable to batting average if hits were predicted based on average outcomes from pitch speed, location, and count.

Note that UZR is then scaled to an average player for this season, not the full six year sample, which means it's even worse as a measurement for comparing how Bogaerts has done over the past month and half versus the rest of the league, since measurements of other shortstops are also unreliable to this point.

UZR is not a measure of what happened. It's a measure of what probably happened. It's not saying "Bogaerts just threw six heads in a row. Even if the coin isn't weighted, that's a pretty good run." It's more akin to saying, "people who toss heads smile 95% of the time after they do, and Bogaerts has smiled six times in a row." DRS is measured differently but based on the same principle. Unlike batting average, there is significant measurement error involved; in other words, there's variance around what has already happened. Bogaerts may have been playing like a below-average shortstop, or as a terrible one.

Great post. I was going to post that I know more about the macro level stuff on fielding stats and the problems in using them for projecting especially in smaller samples but far less about how they work at the micro level and the variation involved in what they report, but you already addressed it. Good show.

To follow up, is there any sense of what the error bars would be at different samples, or does the methodology not allow for that?

Also: you should post more. Seriously. Some of us have literally been talking about this. But you might have a life or something, so, like, whatever.

The Shortstop Position: Weighing the value of defense.

Looks like Zach Galifianakis

New Member

Member

Member

Looks like Zach Galifianakis

will bailey

Looks like Zach Galifianakis

New Member

Member

Looks like Zach Galifianakis

Member

Wants to be startin somethin

Member

Looks like Zach Galifianakis

Wants to be startin somethin

gained membership the hard way - fluffing

Member

Looks like Zach Galifianakis

Member

Member

Looks like Zach Galifianakis

Member

Member

Member

too many flowers

Looks like Zach Galifianakis

Looks like Zach Galifianakis

Member

Member

Member

Looks like Zach Galifianakis

Looks like Zach Galifianakis

Member

Member

Member

Looks like Zach Galifianakis

Member

Looks like Zach Galifianakis

for king and country

Member

Member

Looks like Zach Galifianakis

Looks like Zach Galifianakis

Member

Member

Member

Member

too many flowers

Member

for king and country