One of the main criticisms of WAR by traditionalists is that no one really knows how to calculate it. I used to think this was true, until I discovered the world of wonder that is the FanGraphs Glossary. There, the entire process is explained, with links to other articles that go into greater detail. The equation for WAR (which can be found here) for position players is as follows:
WAR = wRAA + UZR* + BsR + PosA + RepL
*Use rSB + RPP for catchers.
There are a few complicating factors -- the aforementioned calculations for catchers, the park adjustments for wRAA, and the reasoning for the positional and replacement-level adjustments -- but as a whole, this seems pretty simple, right? Well, it is, but it hasn't always been.
See, 2002 was a big year for sabermetrics. That was the first season for which batted-ball data were available, which meant a lot of new stats could be created based off of that. Two of these were UBR -- which measures non-base-stealing elements of baserunning -- and UZR -- which measures fielding. For all player seasons from 2002 until the present, these have been the components of WAR.
But what was used for all years prior, you ask? Well, in lieu of UZR, TZ was implemented; because it is calculated retroactively from box scores and without batted ball data, TZ is less definitive than its replacement. There is not, in fact, a substitute for UBR in the previous years' equations.
So the main takeaway from this is that WAR from 2002 to 2013 is going to be a more accurate reflection of a player's ability than WAR from 1871 to 2001. What the question becomes (or what it became for me, at least) is this: How is this fair? Why should someone that plays in this era be judged differently than a player from the days of yore?
Being the justice-seeker (and, y'know, general loser) that I am, I decided to set things right. I looked at players, of whom there were 406, with at least 2000 plate appearances in the Batted Ball Era, and totaled up their career UBRs and UZRs*, along with their career RAR (Runs Above Replacement -- WAR in run form) and their TZ from 2002 to 2013. I then calculated the career RAR for each player with TZ instead of UZR* and without UBR -- in other words, the Era-Neutral RAR, or ENRAR.
*Again, using rSB + RPP for catchers.
What were the results? For most of the players, there wasn't a whole lot of change. The average RAR of the group was 208.55, and the average ENRAR was 209.59; the R² was .976, which means that there was a very strong correlation between the two.
This is not to say, however, that there were no outliers. Of the 406 players, 35 saw their career RAR increase or decrease by 50 runs or more; these players are reprinted below, that you may gaze upon them in wonder:
Some of the discrepancies (like those for Pierre, Utley, Ortiz, and Konerko) were primarily due to their UBR, while others (like those for Phillips, Zobrist, Cano, and Dunn) can be attributed to differences between UZR and TZ.
But this doesn't really tell us what we want to know, which is how much the players' WAR will change from its old figure (i.e. a percentage increase or decrease). To determine this, I divided their ENWARs by their WARs and subtracted 1, which gave me the percentage change. Again, most players only saw marginal changes, but some (mainly those with very little playing time) were greatly affected. Below, you'll find a table of all players that saw a change of 50% in their career RAR when adjusting for era:
(Note: players with negative career RAR were not included.)
Similar trends appeared, with players seeing baserunning (Millar, Bonifacio) or fielding (Atkins, Wigginton) affect their scores. The largest change, though, was by Teahen, as the title of this post might've suggested; his defense was valued so poorly by TZ (relative to UZR) that he went from a two-win player to a negative two-win player.
So, to reiterate, the change from TZ to UZR and UBR is definitely a good one -- it gives us a much more valid view of a player's performance -- but, when doing historical comparisons, the standards should be the same for everyone, even if it isn't beneficial. (Again, sorry, Teahen!)
. . .