Has Mookie peaked?

absintheofmalaise

too many flowers
Dope
SoSH Member
Mar 16, 2005
13,001
The gran facenda
By ESPN's WAR calculation (they use b-ref's WAR formula, I think), Betts is still ahead of Bogaerts, despite what seems like Bogaerts having a much better season at a premier position so far.

Betts: 13 hr, .268/.388/.458/.847, 3.3 bWAR
Bogaerts: 17 hr, .295/.385/.535/.920, 2.7 bWAR

But then Fangraphs has it different:

Betts: 2.6 fWAR
Bogaerts: 3.7 fWAR

Which seems to be more accurate. How does b-ref calculate WAR different than Fangraphs? I'm sure that's been asked here before but it seems like those are quite disparate numbers.
Fangraphs has a good explanation on their site. And yes, it's been discussed here many times.
 

Savin Hillbilly

loves the secret sauce
SoSH Member
Jul 10, 2007
18,426
The wrong side of the bridge....
By ESPN's WAR calculation (they use b-ref's WAR formula, I think), Betts is still ahead of Bogaerts, despite what seems like Bogaerts having a much better season at a premier position so far.

Betts: 13 hr, .268/.388/.458/.847, 3.3 bWAR
Bogaerts: 17 hr, .295/.385/.535/.920, 2.7 bWAR

But then Fangraphs has it different:

Betts: 2.6 fWAR
Bogaerts: 3.7 fWAR

Which seems to be more accurate. How does b-ref calculate WAR different than Fangraphs? I'm sure that's been asked here before but it seems like those are quite disparate numbers.
They use different defensive metrics. BBref uses DRS, while FG uses UZR. Both metrics love Mookie (UZR perhaps a shade less), but DRS loathes Xander with the heat of a thousand suns, while UZR thinks he's fringe-average. There are other differences, but that's probably the big one in this case.
 

nvalvo

Member
SoSH Member
Jul 16, 2005
16,746
Rogers Park
They use different defensive metrics. BBref uses DRS, while FG uses UZR. Both metrics love Mookie (UZR perhaps a shade less), but DRS loathes Xander with the heat of a thousand suns, while UZR thinks he's fringe-average. There are other differences, but that's probably the big one in this case.
Adding to this: Bogaerts is routinely one of the players with the largest discrepancy between the systems on his defensive performance and thus overall value.
 

Sandy Leon Trotsky

Member
SoSH Member
Mar 11, 2007
2,475
Adding to this: Bogaerts is routinely one of the players with the largest discrepancy between the systems on his defensive performance and thus overall value.
Isn't this yet again, reason to seriously question the legitimacy of advanced defensive statistics? Most of the systems seem far too subjectively based still and continue to look like they're making statistics for the sake of making statistics when the system itself prohibits the application of a statistic based approach. I'll continue to look at DRS, UZR, (there's a few others that I can't recall) with a pile of skepticism. I get that even a pretty intuitive statistic like OPS is still context driven and should be looked at with some level of skepticism..... but the defensive ones are almost comical
 

Savin Hillbilly

loves the secret sauce
SoSH Member
Jul 10, 2007
18,426
The wrong side of the bridge....
Isn't this yet again, reason to seriously question the legitimacy of advanced defensive statistics? Most of the systems seem far too subjectively based still and continue to look like they're making statistics for the sake of making statistics when the system itself prohibits the application of a statistic based approach. I'll continue to look at DRS, UZR, (there's a few others that I can't recall) with a pile of skepticism. I get that even a pretty intuitive statistic like OPS is still context driven and should be looked at with some level of skepticism..... but the defensive ones are almost comical
What's fascinating is that the gap, in Xander's case, continues to grow. It was a modest 3.9 runs in 2015, then 7.6 in 2016 and 10.3 the following year. Last year it hit a thoroughly absurd 20.1 runs -- two wins -- and I thought surely that's got to be the limit; it'll narrow in 2019. But no; if Xander plays 1310 defensive innings this year -- his average over the past four seasons -- and the metrics continue at their current clip, the gap will hit 22 runs this year.
 

Cesar Crespo

79
SoSH Member
Dec 22, 2002
10,564
Isn't this yet again, reason to seriously question the legitimacy of advanced defensive statistics?
There's a lot of reason to question them, including the 3 year sample size needed, and the fact that once you collect that 3 year sample size, it's obsolete because the player is now older and probably worse.
 

brandonchristensen

mad photochops
SoSH Member
Feb 4, 2012
25,133
3 for 5 tonight with a double and a triple. 3 for 4 yesterday with a double. Maybe he’s seeing the ball a bit better. Went 3 for 10 in the Blue Jays series.
 

BaseballJones

goalpost mover
SoSH Member
Oct 1, 2015
6,641
3 for 5 tonight with a double and a triple. 3 for 4 yesterday with a double. Maybe he’s seeing the ball a bit better. Went 3 for 10 in the Blue Jays series.
Yep, last 5 games: 9-19 (.474), 3 doubles, 1 triple, 10 runs scored, 3 rbi, 5 walks, 1 strikeout, and bumped his batting average up from .261 to .271. Nice little stretch for Mookie.
 

SoFloSoxFan

Member
SoSH Member
Jul 14, 2005
1,920
Yep, last 5 games: 9-19 (.474), 3 doubles, 1 triple, 10 runs scored, 3 rbi, 5 walks, 1 strikeout, and bumped his batting average up from .261 to .271. Nice little stretch for Mookie.
So what I am hearing is its time to trade him when his value is high?
 

Pitt the Elder

Member
SoSH Member
Sep 7, 2013
627
You can really see Mookie's quality of contact really picking up over this past week. Batted balls per Baseball Savant:

DatePitcherEV (MPH)LA (DEG)Dist (ft)xBAOutcome
7/2/2019RHP84.4212840.480Single
7/2/2019RHP102.612470.470Single
7/2/2019RHP84.4213490.110Lineout
7/3/2019RHP94182970.420Double
7/3/2019RHP70.3-18990.030Fielders Choice
7/4/2019RHP103.961550.660Groundout
7/4/2019LHP70.9161710.630Lineout
7/4/2019RHP100.7203290.520Lineout
7/4/2019RHP101.681540.730Lineout
7/5/2019LHP100.0162610.510Double
7/5/2019LHP103.331810.580Single
7/5/2019LHP93.8-61170.300Forceout
7/5/2019RHP100.8173420.470Sac Fly
7/5/2019RHP101.4173190.470Single
7/6/2019RHP104.7244220.940Triple
7/6/2019RHP107.3133130.820Double
7/6/2019RHP101.981180.730Lineout
7/6/2019LHP66.5-4040.100Groundout
7/6/2019RHP91.7-11950.330Single

Lots of batted balls over 100 mph, especially the last two games. Also, in his 5 at bats against LHP during this time, Mookie has hit the ball over 100 MPH twice and once more over 90, so that's also encouraging.
 

The Needler

lurker
Dec 7, 2016
1,750
Season line now .273/.393/.469, and he's at .292/.414/.495 since April 19. That ghastly first few weeks has taken a while to get over.
Come on. His OPS was at .922 on May 2. Let’s not pretend he’s just overcoming a bad start.

From May 11 to June 15, however, he was .227/.349/.383.
 

Hawk68

lurker
Feb 29, 2008
172
Massachusetts
Yaz had a .329 to .254 in '70 - '71. He changed his stance a few times and mostly righted the ship. The .254 was in his age 31 year, so maybe he was starting the decline anyway.
As I recall the rise and decline of Yaz, his drop from the peak was driven by wrist injury that dramatically cut his power production. As late as May 29, 1971 his OPS = 1.002. From there steady decline through end of 1971. More of the same reduced production in 1972, excepting June and Sept. After that, he regained reduced but good productivity, with steady age induced dealing thereafter.
 

BaseballJones

goalpost mover
SoSH Member
Oct 1, 2015
6,641
Come on. His OPS was at .922 on May 2. Let’s not pretend he’s just overcoming a bad start.

From May 11 to June 15, however, he was .227/.349/.383.
Betts' Season in Four Segments So Far:

Mar 28-Apr 17: 19 g, 82 PA, 3 hr, 7 rbi, .200/.305/.371/.676, .208 babip
Apr 18-May 2: 13 g, 60 PA, 3 hr, 9 rbi, .449/.533/.735/1.268, .442 babip
May 3-June 15: 38 g, 183 PA, 5 hr, 15 rbi, .227/.366/.387/.753, .246 babip
June 16-July 6: 17 g, 88 PA, 2 hr, 9 rbi, .319/.432/.556/.987, .362 babip (not counting today's game, obviously)

Yes, I cherry-picked the cut off points.
 

The Needler

lurker
Dec 7, 2016
1,750
Betts' Season in Four Segments So Far:

Mar 28-Apr 17: 19 g, 82 PA, 3 hr, 7 rbi, .200/.305/.371/.676, .208 babip
Apr 18-May 2: 13 g, 60 PA, 3 hr, 9 rbi, .449/.533/.735/1.268, .442 babip
May 3-June 15: 38 g, 183 PA, 5 hr, 15 rbi, .227/.366/.387/.753, .246 babip
June 16-July 6: 17 g, 88 PA, 2 hr, 9 rbi, .319/.432/.556/.987, .362 babip (not counting today's game, obviously)

Yes, I cherry-picked the cut off points.
Yes, you did. I think you may be agreeing that he has been inconsistent. I can't imagine you're attempting to prove anything with those selections, but if we're playing fun with cherry-picking, I would submit the hot and cold has been even more frequent - closer to 10 to 15-game chunks:

Mar 28-Apr 5: 9 g, 43 PA, 2 HR, .289/.372/.526/.898, .310 babip
Apr 6-Apr 17: 10 g, 39 PA, 1 HR, .094/.231/.188/.418, .083 babip
Apr 19-May 2: 13 g, 60 PA, 3 HR, .449/.553/.735/1.268, /442 babip
May 3-May 14: 9 g, 46 PA, 1 HR, .200/.391/.343/.734, .240 babip
May 15-May 24: 9 g, 43 PA, 1 HR, .351/.442/.541/.982, .387 babip
May 25-June 8 (1): 13 g, 61 PA, 2 HR, .154/.279/.288/.567, .140 babip
June 8 (2)-June 24: 15 g, 76 PA, 3 HR, .274/.408/.548/.956, .304 babip
June 25-July 4: 7 g, 34 PA, 0 HR, .222/.382/.259/.642, .286 babip
July 5-July 6: 2 g, 11 PA, 0 HR, .667/.636/1.111/1.747, .600 babip
 

Pitt the Elder

Member
SoSH Member
Sep 7, 2013
627
It's still a pretty small sample, but, so far in July, Mookie appears to have regained his stroke against LHP:

PitcherBBEEV (MPH)LA (DEG)Hard HitBarrel %
Total3691.8220.4253%8.3%
RPH2591.7722.2848%4.0%
LHP1191.9516.1864%18.2%

He's been consistently hitting the ball 95+ MPH this month, at an even higher clip against LHP than RHP. Interestingly, his barrel % is down across the board, even though his EV and LA both suggest that he should be getting more of those. Anecdotally, he's been hitting a lot of line drives and fewer bombs. If this gets him back to his usual self, I'm fine with this. Maybe we'll start seeing him go bridge a bit more.

BTW, it looks like Mookie has started crushing off-speed stuff this month. His exit velocity on off-speed this season by month:
  • Apr: 83.9
  • May: 86.5
  • June: 92.1
  • July: 99.0
His K% is also down to 10.4% this month while his BB% rate is still a healthy 14.6%. I very much like what I see.

EDIT: apparently I missed a barreled ball in my analysis...Mookie actually at 18% vs LHP in July and 8.3% overall
 
Last edited:

Pitt the Elder

Member
SoSH Member
Sep 7, 2013
627
A lot less chatter in the Mookie thread now that he's raking again. For kicks, here's Mookie's slash the day this thread started versus where it is today (and what he hit in between):
  • Jun 30: .261/.381/.453
  • July 18: .284/.399/.480
  • Interim: .431/.508/.647
That's MVP Mookie, and I think SoSH should take full credit for this turn-around. Oh, and Mookie hit two more 100+ MPH balls against LHP today, including a barreled flyout, as well as a 104 MPH dinger of a RHP that went 418 ft.

With Devers and Bogie doing what they're doing, MVP Mookie makes this a very dangerous team. If the pitching rounds into form, the Sox still have time to make a run. With the Yanks and Rays coming up for 14 straight games, it's a great opportunity to do some damage.
 

Van Everyman

Member
SoSH Member
Apr 30, 2009
18,307
Newton
Tonight when Guerin Austin asked Mookie why he was hitting so well, his reply was “I’ve put in a lot of work.”

You could say that. The day after the All Star Game, my daughter was at a local batting cage with a friend and they looked over at the last tunnel and saw Mookie working on his swing. Here was the reigning MVP—mired in a slump and a late inning replacement in said ASG—flying directly home the day after and hitting the cages at some dingy humid place open to the public. Another girl saw him there again the following day.

Only thing missing:
25399
 

teddywingman

Looks like Zach Galifianakis
SoSH Member
Jul 31, 2009
8,262
a basement on the hill
That is so cool. My love of Mookie knows no bounds, and if the Sox fail to reach an agreement with him, there's a good chance it could affect my rooting interests.

Don't get me wrong, I will always root for for the olde town team, but I'm also going to keep rooting for Mookie.
 

Savin Hillbilly

loves the secret sauce
SoSH Member
Jul 10, 2007
18,426
The wrong side of the bridge....
After last night's outburst, Mookie has now climbed into the top 10 in MLB in fWAR, which means the Red Sox have three players in that group (Xander #4, Devers #6). If you're looking at just the AL, they are 2, 3 and 5.
 

BaseballJones

goalpost mover
SoSH Member
Oct 1, 2015
6,641
Mookie's last 33 games: 164 pa, 7 hr, 6 sb, .345/.433/.626/1.059

Since April 19: 84 g, 407 pa, 15 hr, 11 sb, .308/.418/.536/.953

So after that rough start of the season, he's been really really really good. For comparison's sake, we are lauding (rightly so) Rafael Devers' breakout season that, if not for Mike Trout, would be an MVP-caliber season. But look at what Mookie has (pretty quietly) done since April 19 compared to Devers' season:

Mookie (since Apr 19): .308/.418/.536/.953
Devers (whole season): .326/.378/.574/.952

Virtually identical OPS numbers. And this is a "down" year for Betts. He's that good.
 

Sandy Leon Trotsky

Member
SoSH Member
Mar 11, 2007
2,475
Mookie's last 33 games: 164 pa, 7 hr, 6 sb, .345/.433/.626/1.059

Since April 19: 84 g, 407 pa, 15 hr, 11 sb, .308/.418/.536/.953

So after that rough start of the season, he's been really really really good. For comparison's sake, we are lauding (rightly so) Rafael Devers' breakout season that, if not for Mike Trout, would be an MVP-caliber season. But look at what Mookie has (pretty quietly) done since April 19 compared to Devers' season:

Mookie (since Apr 19): .308/.418/.536/.953
Devers (whole season): .326/.378/.574/.952

Virtually identical OPS numbers. And this is a "down" year for Betts. He's that good.
Devers also had a slow start to the season. He was hitting and getting on base, but it took him32 games before hitting his first HR. If we're going to cut "struggles" out of the season, and use Devers first game cranking a HR on the season, he looks like this:

Devers (since May 2)- .334/.374/..643/ 1.017

Or just since the same time frame you picked for Mookie:

Devers (since Apr 19)- 338/383/624/ 1.008
 

Pitt the Elder

Member
SoSH Member
Sep 7, 2013
627
Mookie's last 33 games: 164 pa, 7 hr, 6 sb, .345/.433/.626/1.059

Since April 19: 84 g, 407 pa, 15 hr, 11 sb, .308/.418/.536/.953

So after that rough start of the season, he's been really really really good. For comparison's sake, we are lauding (rightly so) Rafael Devers' breakout season that, if not for Mike Trout, would be an MVP-caliber season. But look at what Mookie has (pretty quietly) done since April 19 compared to Devers' season:

Mookie (since Apr 19): .308/.418/.536/.953
Devers (whole season): .326/.378/.574/.952

Virtually identical OPS numbers. And this is a "down" year for Betts. He's that good.
To be fair, Devers didn't really get going until May (hit his first HR May 2) so I don't think that's a fair comparison. Also, Mookie had a really good April followed by a poor May and June, so I'm not sure if your April 19 cut off point really tells the full story. I agree, though, that he's been MVP Mookie for a month now.
 

BaseballJones

goalpost mover
SoSH Member
Oct 1, 2015
6,641
Devers also had a slow start to the season. He was hitting and getting on base, but it took him32 games before hitting his first HR. If we're going to cut "struggles" out of the season, and use Devers first game cranking a HR on the season, he looks like this:

Devers (since May 2)- .334/.374/..643/ 1.017

Or just since the same time frame you picked for Mookie:

Devers (since Apr 19)- 338/383/624/ 1.008
To be fair, Devers didn't really get going until May (hit his first HR May 2) so I don't think that's a fair comparison. Also, Mookie had a really good April followed by a poor May and June, so I'm not sure if your April 19 cut off point really tells the full story. I agree, though, that he's been MVP Mookie for a month now.
I don't disagree with you guys. Devers has been great. It's just that Mookie quietly has been a LOT better than it feels like he's been.

BTW: The Sox lead the majors in runs scored, hits, doubles, RBI, avg, and obp. Not bad.
 

Al Zarilla

Member
SoSH Member
Dec 8, 2005
50,907
San Andreas Fault
Last night, Mookie added 7.25% to his total base total for the year, in his 103rd game. His 5 RBI made the difference. Without those, we may still be playing. Well, except for bullpens.
 

Savin Hillbilly

loves the secret sauce
SoSH Member
Jul 10, 2007
18,426
The wrong side of the bridge....
If only we led the league in wins. Really points out how terrible the pitching has been.
Has it been terrible, really? Or only snakebit? The Sox rank 6th in MLB in SIERA, 7th in FIP--and 19th in ERA. Their ERA-minus-FIP of 0.46 is second only to the Rockies. They are 26th in LOB%. Maybe you could argue that that's more un-clutch than unlucky, but whatever the reason, the performance is translating to results much less efficiently than normal.
 
Jul 5, 2018
223
Tonight when Guerin Austin asked Mookie why he was hitting so well, his reply was “I’ve put in a lot of work.”

You could say that. The day after the All Star Game, my daughter was at a local batting cage with a friend and they looked over at the last tunnel and saw Mookie working on his swing. Here was the reigning MVP—mired in a slump and a late inning replacement in said ASG—flying directly home the day after and hitting the cages at some dingy humid place open to the public. Another girl saw him there again the following day.

Only thing missing:
View attachment 25399
If all it took was hard work, there would be no batting slumps in MLB.

Last year Mookie met with Doug Latta, "the swing whisperer", and immediately went on a batting tear. But then he struggled in September and the post season. Shouldn't Mookie had met with him again?

Some people believe that JDM has discovered an innovative approach to hitting and has helped other Sox players, but it's not working as well for him this year. My understanding is that he is chasing more bad pitches this year which would indicate he's thinking too much at the plate.

I've read that Cora (career 72 OPS+) wanted Betts to start hacking at the first pitch, but Betts' 3 homers last night were on 3-2, 2-2 and 3-1 counts.

Why batters have hot streaks and slumps is a riddle that will never be solved.
 

The Needler

lurker
Dec 7, 2016
1,750
Has it been terrible, really? Or only snakebit? The Sox rank 6th in MLB in SIERA, 7th in FIP--and 19th in ERA. Their ERA-minus-FIP of 0.46 is second only to the Rockies. They are 26th in LOB%. Maybe you could argue that that's more un-clutch than unlucky, but whatever the reason, the performance is translating to results much less efficiently than normal.
I think it's a little bit of both.

Of the three general reasons for differences between FIP and ERA (defense, luck, and timing/sequencing), the "luck" aspect of the FIP/ERA differential is usually attributed to babip (and the Sox have the third highest in MLB). The highlighted I think refers more to sequencing, which si similar to luck in that it usually is outside of a pitcher's control. As you note, the Sox are 26th in LOB%, largely because they've pitched much worse than last year with men on and in scoring position. And it is true that they have a significantly higher babip with men on than with the bases empty, nearly a 50 point difference. But it's also true that the things they can control (K/BB) have been significantly worse than in those situations last year and in bases empty situations, and with a much larger differential, and those they can somewhat control (HR) are also significantly worse than last year.

2018:
Bases empty: .687 OPS, .289 babip, 3.04 K/BB, 33.46 PA/HR
Runners on: .715 OPS, .302 babip, 3.05 K/BB, 37.41 PA/HR
Runners ISP: .711 OPS, .283 babip, 2.70 K/BB, 36.49 PA/HR

2019:
Bases empty: .736 OPS, .301 babip, 3.13 K/BB, 28.40 PA/HR
Runners on: .782 OPS, .331 babip, 2.73 K/BB, 32.04 PA/HR
Runners ISP: .829 OPS, .347 babip, 2.22 K/BB, 32.24 PA/HR

There's also the matter of defense, of course. The Sox are 19th in MLB according to UZR (.9) this season; they were 7th (25.3) last year.
 

Average Reds

Dope
Staff member
Dope
V&N Mod
SoSH Member
Sep 24, 2007
25,818
Southwestern CT
There's also the matter of defense, of course. The Sox are 19th in MLB according to UZR (.9) this season; they were 7th (25.3) last year.
I am not one of those who hate UZR, but even proponents of the tool believe that a single season of data is far too small to draw conclusions with. The idea that a half season+ is determinative seems a bit much.
 

The Needler

lurker
Dec 7, 2016
1,750
I am not one of those who hate UZR, but even proponents of the tool believe that a single season of data is far too small to draw conclusions with. The idea that a half season+ is determinative seems a bit much.
A bit much of what? FIP literally “fielding independent.” Attributing the delta between FIP and ERA solely to luck is serious head in the sand stuff.
 

Devizier

Member
SoSH Member
Jul 3, 2000
10,526
Somewhere
A bit much of what? FIP literally “fielding independent.” Attributing the delta between FIP and ERA solely to luck is serious head in the sand stuff.
I think the problem is that we can’t confidently attribute the difference to much at this point. I agree that defense is part of the equation, but how much? Ergo the comments about low sample UZR.
 

The Needler

lurker
Dec 7, 2016
1,750
I think the problem is that we can’t confidently attribute the difference to much at this point. I agree that defense is part of the equation, but how much? Ergo the comments about low sample UZR.
I’ve never seen criticisms of team (not player) UZR based on only 100 games, but maybe I’ve missed it. In any case, the research I’ve seen attributed about 40-50% of the delta to defense, and showed a correlation between UZR and DRS with FIP. As I said in my initial post, it’s likely a combination of worse pitching and bad luck as well as defense.
 

shaggydog2000

Member
SoSH Member
Apr 5, 2007
6,346
I’ve never seen criticisms of team (not player) UZR based on only 100 games, but maybe I’ve missed it. In any case, the research I’ve seen attributed about 40-50% of the delta to defense, and showed a correlation between UZR and DRS with FIP. As I said in my initial post, it’s likely a combination of worse pitching and bad luck as well as defense.
UZR probably is not accurate over a player half-season, but a team half season is a much larger number of samples, by almost an order of magnitude. 100 games for a team is 5 player seasons. That much data over a relatively short period of time should get us to a predictive level of information, considering player data is assumed to be predictive after 2-3 seasons.
 

Savin Hillbilly

loves the secret sauce
SoSH Member
Jul 10, 2007
18,426
The wrong side of the bridge....
UZR probably is not accurate over a player half-season, but a team half season is a much larger number of samples, by almost an order of magnitude. 100 games for a team is 5 player seasons. That much data over a relatively short period of time should get us to a predictive level of information, considering player data is assumed to be predictive after 2-3 seasons.
Yes, but isn't UZR calculated for each player relative to his peers at that position? If so, and unless I'm confused, "team UZR" is a bit of a legal fiction, because the denominator for each player (and even for each player/position) is different, so all you're really doing is summing a bunch of noisy numbers. That doesn't reduce the noise, does it? Or does it? Happy to bow to my statistical overlords here.
 

Devizier

Member
SoSH Member
Jul 3, 2000
10,526
Somewhere
Yes, but isn't UZR calculated for each player relative to his peers at that position? If so, and unless I'm confused, "team UZR" is a bit of a legal fiction, because the denominator for each player (and even for each player/position) is different, so all you're really doing is summing a bunch of noisy numbers. That doesn't reduce the noise, does it?
You are correct. If you combine two measurements the result will have more error than the original measurements. How much more depends on how you combine them.

And that makes sense because if you have some uncertainty about one thing (x) and uncertainty about another (y), are you going to be more or less certain about the result of x+y?
 

shaggydog2000

Member
SoSH Member
Apr 5, 2007
6,346
You are correct. If you combine two measurements the result will have more error than the original measurements. How much more depends on how you combine them.

And that makes sense because if you have some uncertainty about one thing (x) and uncertainty about another (y), are you going to be more or less certain about the result of x+y?
Depends on whether the error is correlated or non-correlated. I would assume that the error in the measurement of separate player defensive ratings would be uncorrelated. But maybe there is some systemic error that is correlated between the different measurements. The best way to combine the separate errors would be to do a weighted root sum of squares. That actually has the effect of lessening the total error, it sort of averages them across the different measurements because some errors are positive and some are negative.
 

The Needler

lurker
Dec 7, 2016
1,750
My understanding is that individual UZR ratings are reached through a division of of the team's responsibility (fault or credit) for each ball hit, rather than the team UZR being calculated by addition of each individual player rating. Because an error in such allocation of responsibility (which is often split among two or more players on a given play) can have a great impact on an individual player's UZR when he has fewer opportunities, sample size is a legitimate concern. But since team UZR is based on every single ball hit, that concern is not present.

EDIT: I hope what I'm trying to express is clear. What I'm saying is, all UZR does is compare what the outcome of a similarly hit ball over the course of several years of data to the result achieved on a given play. So while it might be arguable to say a given player is being rated unfairly because of the way he or his coaching staff positioned him, or because of the allocation of fault applied to him, those concerns don't really apply to team UZR, which is just considering whether the play should have been made by someone on the team. It doesn't really matter whether it's a positioning thing, or a range thing; other teams make the play more frequently, so it's on the defense.
 
Last edited: