fWAR Getting an Update to the Defensive Component

pokey_reese

Member
SoSH Member
Jun 25, 2008
16,253
Boston, MA
article here

I thought that this was worth a thread given how often WAR is used around here, and how often it leads to disagreements within threads, in particular because of how it handles defense. In particular, this will help address one of the main complaints about it, which is how credit assigned/taken away on a given play handles the shift. Overall, using the Statcast positional data is a pretty big change for one of the 'premiere' advanced metrics out there, and I will be interested to see whether it leads to more or less agreement with bWAR, as well as discussion/perception of Red Sox players like X and Devers who have received poor grades in the past, while being highly impacted by shift adjustments.
 

pokey_reese

Member
SoSH Member
Jun 25, 2008
16,253
Boston, MA
Interesting to note, JBJ and X are both among the players with the biggest change over the 2016-2021 time period they adjusted for. X grades out as about 3 wins worse over that time, and JBJ got an increase of 3 wins.
 

nvalvo

Member
SoSH Member
Jul 16, 2005
21,510
Rogers Park
I was juuuuust coming here to post about this. This is pretty huge.

Bogaerts is one of the biggest changes, and he loses 3.2 fWAR over the 2016-2021 period. To keep the scale in view, that's ~30 runs of difference over 6 seasons — this change is only to one element of the fielding component of fWAR.

Bogaerts goes from 26.8 fWAR over that period to 23.6. Baseball-Reference WAR has him at 23.7. In other words, the Great Fangraphs/B-R Bogaerts discrepancy is over, and it has been decided more or less in B-R's favor.

(JBJ gains about 3 WAR over that period, FWIW, going from 12ish to 15ish. Basically, Fangraphs just decided that Bradley is a 20% better player than they had previously held.)

edit: or, uhhh, what p-r said...
 

Rovin Romine

Johnny Rico
Lifetime Member
SoSH Member
Jul 14, 2005
23,721
Miami (oh, Miami!)
I was juuuuust coming here to post about this. This is pretty huge.

Bogaerts is one of the biggest changes, and he loses 3.2 fWAR over the 2016-2021 period. To keep the scale in view, that's ~30 runs of difference over 6 seasons — this change is only to one element of the fielding component of fWAR.

Bogaerts goes from 26.8 fWAR over that period to 23.6. Baseball-Reference WAR has him at 23.7. In other words, the Great Fangraphs/B-R Bogaerts discrepancy is over, and it has been decided more or less in B-R's favor.

(JBJ gains about 3 WAR over that period, FWIW, going from 12ish to 15ish. Basically, Fangraphs just decided that Bradley is a 20% better player than they had previously held.)

edit: or, uhhh, what p-r said...
Career Bradley:
bWAR: 18
fWAR: 17.6*
old fWAR: 14.4

*does not apply new metric for 1,800 defensive innings between 2013-2015.
 

Max Power

thai good. you like shirt?
SoSH Member
Jul 20, 2005
7,879
Boston, MA
This just goes to show how imprecise a measure of performance WAR is. It looks good with a single number that is calculated down to a decimal point, but Fangraphs literally just made a change where "The vast majority of players’ new WAR calculations fall within a +/-1 WAR range of their previous WAR figures." So two guys who both had 4 win seasons yesterday may have actually been a 5 and 3. Or maybe not. And some were even more off.

I know the amount of stats we have available can be overwhelming, but simply ignoring all of them and taking the single WAR number and treating it as absolute truth is foolish. At best it's a ballpark estimate of how a player performed with no consideration given to game context.
 

Toe Nash

Member
SoSH Member
Jul 28, 2005
5,600
02130
There was talk in the CBA about using WAR to determine player pay during arbitration, and I immediately thought "which WAR"? Was that ever hammered out?

The bigger issue with WAR to me is pitching, where fWAR and bWAR not only use different metrics but approach the issue philosophically differently by using defense-independent inputs, or not. At least with the position players' WAR, the goal was the same and the two sites are just using a different measure of defense. With pitching, it's a different goal.

Reasonable people can disagree on what's best (I agree with bWAR personally) but it really needs to be shouted from the rooftops that they're totally different when it comes to pitchers.
 

Niastri

Member
SoSH Member
Interesting to note, JBJ and X are both among the players with the biggest change over the 2016-2021 time period they adjusted for. X grades out as about 3 wins worse over that time, and JBJ got an increase of 3 wins.
Goes to show that the Red Sox internal measurements are probably more accurate than what we have available as the public.

Their decision making the last 12 months seems to reflect they already knew JBJ is more valuable and Bogaerts less valuable than popular perception.
 

ArttyG12

Member
SoSH Member
Jul 15, 2005
100
Goes to show that the Red Sox internal measurements are probably more accurate than what we have available as the public.

Their decision making the last 12 months seems to reflect they already knew JBJ is more valuable and Bogaerts less valuable than popular perception.
I think this somewhat overstates it. It's half a win per season. It still has JBJ as below replacement level last year.
 

Niastri

Member
SoSH Member
It's only a defensive update, it won't change the fact that he can't hit.
My point isn't that JBJ just became a great player, but that the Red Sox internal data probably thinks he's more valuable than we think he is.Hence trading for him. The update to fWAR just made his value a little closer to reality over the last few years.

I personally don't like it that JBJ is a starter, I would have preferred a better bat, but so far... in Bloom I Trust.
 

mauf

Anderson Cooper × Mr. Rogers
Moderator
SoSH Member
I know the amount of stats we have available can be overwhelming, but simply ignoring all of them and taking the single WAR number and treating it as absolute truth is foolish.
It’s better than concluding one player is better than another because solely he has a higher batting average, or more pitching wins, or even a higher OPS. Any attempt to measure value using a single metric will be deeply flawed. The difference is that people use WAR in a conclusory way and think they’re doing actual analysis. Which, of course, they aren’t.