WAR Changes and Updates!
A few weeks ago, you may have seen this tweet:
In today’s Crazy Big Nerd News, FanGraphs and Baseball-Reference.com have agreed to standardize Replacement Level. #SABRanalytics
— robneyer (@robneyer) March 8, 2013
Today, we’ve rolled out those changes, and some other ones!
Here are the details:
– Replacement level is now set at an even 1,000 WAR per 2,430 game season. Batters are given 57% of all total WAR with pitchers receiving the remaining 43%. This will cause players to drop in WAR by about 0.3 WAR per recent season.
– The leagues for batters are now zeroed out in terms of runs above average. Most of this adjustment is done in Batting, while the final league adjustment is done behind the scenes. We will soon have a “League Adjustment” stat which shows exactly what the league adjustment is for each player. On a seasonal level, AL position players will see about a 0.2 decrease, with NL position players receiving about a 0.4 decrease in WAR.
– Infield fly balls are now part of our pitcher WAR calculation. They are counted the same as a strikeout in the FIP calculation (though generic FIP on the site has not changed). You can see Dave Cameron’s original post to find out more about this change.
– UZR has been given a minor update. There was a small issue with calculating the foul line and this has now been corrected. Also, the 2012 park factors were updated. Players may have had their UZR values changed by a run or two at most.
Dave Cameron will have another post up in about 15 minutes going into more detail on the replacement level changes.
David Appelman is the creator of FanGraphs.
Dagnabbed gubmint meddlin’ in our baseball!
who cares?
most of the people reading.
Sprot your comment adds absolutely nothing to the discussion. Why even bother commenting?
You try to rob me of what little WAR I have. Everybody’s always trying to bring me down…
Not me Yuni!!!1 I think you’re a heckuva ballplayer.
Meh.
This is cool.
Could bbFIP, the one with IFFBs, be published too?
Yes, please.
It would be helpful to see the formula for FIP with IFFBs included and also “PU%” i.e. IFFB/PA, not IFFB/FB. I don’t care what it’s called, but it’d be nice to have these stats.
Will you show IFFIP on the dashboard now? Since it is a WAR component, it would help with quick diagnoses of players.
This is great news. I think it’s a non-trivial step for the quantitative approach to normalize among its advocates. Terrific work!
Bravo to Dave, Dave, Sean, Sean, and any other FG and BR staffers who were involved in this awesome step forward, but through a parenting oversight aren’t named Dave or Sean.
Seriously, though, bravo.
Funny you would call that a parenting oversight coming right after your cool names series.
It’s like how all the major tech/science guys in the 90s were named Steve. Steve Jobs, Steve Wozniak, Steve Hawking, Steve Pinker, Steve Weinberg, Steve Gould
I have never heard anyone refer to “Steve” Hawking.
Steve and I used to go for fish and chips and Fullers ESB at the Four Candles every Saturday.
That’s actually not true.
Plus the original host of Blue’s Clues was named Steve. How quickly we forget!
“Four candles – you know, handles for forks.”
I’m missing something fundamental here in regards to “Replacement level is now set at an even 1,000 WAR.” Isn’t replacement level definitionally worth 0 WAR?
A replacement level team would, but the league as a whole is Average. Thus, the 1000 WAR is the amount of wins an 81 win team has over a replacement team x 30 teams.
Thank you, you explained this much better than the post did (assuming you’re right).
So an all-replacement level team would now win 48 games?
Gotcha. Thanks.
Will Infield Flies be adjusted based on whether Sam Holbrook is calling them?
something so nerd-tastic about this.
Exactly. I’m really excited about this, but it would take a ridiculous amount of nerdy background to explain to anyone here at work.
Somewhat related… I just found myself trying to explain to a Royals fan how Shields is better than he looked in 2012 and not quite as good as he seemed in 2011… but probably somewhere in between. I had a screenshot of his player profile and said “f it” and deleted the whole e-mail.
his fangraphs profile that is
How does this adjust the meaning of “average” player? Is 2 WAR still a good estimate of “average”, with 5 and 8 WAR as benchmarks for All-star and MVP? Or should these be adjusted downward as well?
I would not re-adjust your scale of what “average”, “all-star”, or “MVP” is. On a seasonal basis, there was very little change in WAR. The big differences will show up in career WAR.
So by standardizing WAR they are going to use the same WAR calculations now? MEaning BBREF is going to calculate their WAR using FIP now, like FG?
Nope. Only replacement level is standardized. The inputs and calculations are still different
Can someone explain to me like I’m five years old why FIP with IFFB is being incorporated into WAR, but why ‘generic FIP’ will remain unchanged? If this is an improvement to FIP, why only include it in WAR? I’d rather see ‘generic FIP’ include IFFB, or at least have ifFIP available separately.
Because FIP is FIP, but Fangraphs can calculate WAR however they damn well please.
Which is I why I suggested they could also offer iffbFIP separately. The question is why not also offer the latest and greatest instead of just only rolling it into WAR?
Well your first question was why they aren’t changing generic FIP, then you asked that again with different phrasing, then you requested it again, and then you threw in “the question” at the end there as a less desirable alternative. But as to that minor point of your original comment, sure, they could offer it separately.
Odd that the consensus would be to net to zero the performance above average by league given that, with interleague play, the number of actual wins by league doesn’t net to zero.
This, or I’m misunderstanding what happened.
This change greatly hurts pitchers who dont necessarily induce ‘infield flies’ but a lot of high flyballs to the shallow outfield, because all those near-guarantee outs aren’t being considered like they should be. It would have been better to withhold this update until you can use fieldFX calculations to determine if a flyball is an automatic out–not an arbitrary definition that changes depending on the weather, game scorer, and whether or not the infield lines are painted on the field. A blooper that stays in the infield has to be dived after to be caught is anything but a sure out, yet it gets counted nonetheless.
The babip for what have been scored as infield flies over the past several years is tiny. A strikeout isn’t even a guaranteed out unless there is a runner on first with fewer than 2 outs.
But even then, it’s more fielding independent than Ks. I don’t think we need to wait.
Will the park factors be updated with 2012 information in the Guts! section? Thanks
Don’t forget to remove in field flies from UZR (if they were ever there), otherwise there is double credit going around in the WAR statistics. I assume this hasn’t been done because I don’t see it mentioned in the UZR changes. Please correct me if I’m mistaken.
“Batters are given 57% of all total WAR with pitchers receiving the remaining 43%. ”
Whats the reasoning for this particular split?