Fangraphs and Baseball-Reference Agree to Unify Baseline for Replacement Level

[RVAbird]

Sean Foreman of Baseball-Reference and David Appelman and Dave Cameron of Fangraphs have agreed to use the same baseline for Replacement Level, which will alleviate some of the more drastic discrepancies in career WAR between the two sites.

The answer, quite simply, lies with replacement level. Our model used a lower baseline than Baseball-Reference did, so the same performance would result in a higher WAR in our model than in theirs. Over very long careers ? like Morris?, for instance, or many of the old time pitchers who threw forever ? this could really begin to add up, and give the appearance of large disagreements when the two systems didn?t actually see things all that differently. In the case of guys with substantial careers, many of the large discrepancies were simply driven by the fact that the two sites had a different definition of replacement level.

As David noted a few minutes ago, this new unified replacement level is now set at 1,000 WAR per 2,430 Major League games, which is the number of wins available in a 162 game season played by 30 teams. Or, an easier way to put it is that our new replacement level is now equal to a .294 winning percentage, which works out to 47.7 wins over a full season. Conveniently, this number is almost exactly halfway in between our previous replacement level (.265) and Baseball-Reference?s previous replacement level (.320), though the number wasn?t chosen solely as an equal compromise.

That paragraph is a little clumsy. 2,430 is HALF the number of games played by 30 teams over a 162 game season. So a record of 2,430-2,430 would yield a .500 winning percentage. 1,000 wins would yield a .294 winning percentage, or about 48 wins for a single team over the course of a season. This means that a roster filled with replacement players would now be expected to win about 48 games.

Note that this change does not affect the more intentional discrepancies between fWAR and bWAR, most noticeably the use of ERA vs. FIP to value pitchers and the difference in defensive metrics. fWAR and bWAR will still not be identical and interchangeable. The difference is that the two systems are now operating on the same scale, as the replacement threshold is the same for both.

This will make little difference for single-season WAR values. But, as far as fWAR is concerned, players who were rewarded for lengthy careers of solid performance will see their WAR fall more relative to players with shorter careers but more outstanding performance.

LINK: http://www.fangraphs.com/blogs/index.php/unifying-replacement-level/

[isestrex]

Dave Cameron talked on his podcast about him, David and Sean sitting down at a restaurant to talk about this possibility (and eventuality) of their WAR's unifying. I'm so grateful that this is going to be congruent at long last.

[Loophole]

Dave Cameron talked on his podcast about him, David and Sean sitting down at a restaurant to talk about this possibility (and eventuality) of their WAR's unifying. I'm so grateful that this is going to be congruent at long last.

Don't they use different input for pitcher WAR though? I thought FG used FIP and BR used ERA.

Another thing not mentioned in this thread is that FG is going to incorporate infield fly balls into FIP, counting them similarly to strikeouts.

[SteveA]

So is this just moving forward, or will all existing stats from previous years all be recalculated and republished based on these values?

Think of all the studies out there that have looked at total WAR derived from a particular round draft pick, or compared players' WARs to make a Hall of Fame or MVP arguments, or evaluated trades based on WAR each team got. If all the existing stats change then all those previous studies/articles/opinions will suffer some degree of loss of validity?

[skanar]

So is this just moving forward, or will all existing stats from previous years all be recalculated and republished based on these values?

Think of all the studies out there that have looked at total WAR derived from a particular round draft pick, or compared players' WARs to make a Hall of Fame or MVP arguments, or evaluated trades based on WAR each team got. If all the existing stats change then all those previous studies/articles/opinions will suffer some degree of loss of validity?

Not really, as long as they compared fWARs with fWARs or rWARs with rWARs. The numbers will change a bit, but they should be similar to each other, relatively speaking. All fWARs are going down a bit, all rWARs are going up a bit.

The biggest differences are for players with very long careers. Players with many valuable seasons but lower peaks (think Brooks) will drop a bit on aggregate relative to players with high peaks (think Cal). But even there the differences shouldn't be enormous.

Brooks: 92.6 --> 80.8

Cal: ~98-100 (this is from my own memory) --> 92.5

Both went down quite a bit, as you'd expect for any player with a long career. Brooks went down a bit more, because his value was more spread out over good seasons while Cal's was more concentrated in great ones.

[RVAbird]

Don't they use different input for pitcher WAR though? I thought FG used FIP and BR used ERA.

Another thing not mentioned in this thread is that FG is going to incorporate infield fly balls into FIP, counting them similarly to strikeouts.

From the OP:

Note that this change does not affect the more intentional discrepancies between fWAR and bWAR, most noticeably the use of ERA vs. FIP to value pitchers and the difference in defensive metrics. fWAR and bWAR will still not be identical and interchangeable. The difference is that the two systems are now operating on the same scale, as the replacement threshold is the same for both.

This is about making the scale the same, not the metrics themselves. There will always be subtle differences in their calculations.