2017 UZR Updates!
For the 2017 season, Mitchel Lichtman has made some improvements to the UZR methodology!
– UZR now uses hit timer data (hang time) rather than hit type designations, which is an improvement on the methodology and thus the results.
– The methodology has changed a little that allows UZR to account for some of the noise associated with imperfect data. The net result of this change is that extreme UZR’s, which were likely caused by, to some extent at least, noise in the data, rather than extreme performance, will be slightly ‘dampened.’ We think that these new values, while very close to the old ones in most cases, more accurately reflect the actual performance of the players in question.
These changes in UZR are currently active for 2017, and will also be rolled out for 2012 – 2016 data in the near future.
David Appelman is the creator of FanGraphs.
Is the methodology change simply the regression to the mean that’s been speculated on for some time, or is it more complicated than that?
I see Matt kemp is off to a robust defensive start like usual
Also the A’s are still last!
Interesting, I didn’t know you had hit timer data going all the way back to 2012. I certainly will be interested in how this affects certain extreme seasons. (Looking pretty closely at you 2013 Andrelton Simmons)
“These changes in UZR are currently active for 2017, and will also be rolled out for 2012 – 2016 data in the near future.”
What sort of difference are we talking? Are there players that will get 10+ runs better or worse in a season with the new system?
My guess is that it is possible, but unlikely. There aren’t that many people that either have 10 +/- difference between UZR and DRS. I’m thinking the extremes move in some, but wouldn’t expect more than a 2-5 run difference in even the extreme cases.
We can easily find out when they roll it out. Go and export the leaderboards from 2012-2016 then check them against the new updates.
I’m expecting not a lot of difference but will be curious who is hurt and who is helped.
The differences will be pretty minor for almost all players, but some players on the very high / low ends will be regressed a bit.
How often will uzr numbers update, every 2 weeks?
“We changed the way we fundamentally evaluate players no we won’t tell you what we did” great work as always
Cool — will you have some posts that scatterplot new vs. old data for 2016 and earlier years?
And is there enough historical hang time data to see whether the new method increases year-to-year correlation or prediction accuracy? Correlating New 2015 with New 2016, and Old 2015 with Old 2016, which is tighter?
Small sample UZR oddities from 2017:
Manny Machado is on -8 UZR/150 rate right now. Andrelton Simmons is even worse at -10 UZR/150 rate so far. I would expect some heavy regression back to their previously established talent levels… but what if it’s for real?!
On other other end, Joey Votto is at +20 UZR/150 rate, so I guess that “getting more athletic” and “working on my fielding” thing is working out? Nick Markakis is at +19 UZR/150 right now. See, the industry opinion on his defense was right all along!
This has an episode of Fun with Small Sample Defensive Metrics.