Usual caveat: Rob Neyer is great. Really -- I'm among the many people out there who was first exposed to sabermetrics through his ESPN columns and "Rob and Rany on the Royals." He generally does a good job of keeping up with the latest stuff and getting it across in an easily-digested  manner for his readers.

Moreover, MVP races are what they are -- a good idea for an award that often ruined by voters who seemingly don't have the faintest idea of what makes a baseball player valuable.

Still, I was a bit put off by Neyer's recent blog posting responding to a reader's claim that Todd Helton, not Jeff Kent, should have won the 2000 NL MVP (Helton placed seventh).  Much of the debate, of course, is about the way Coors Field inflates hitters' numbers. The reader points out that Helton also had great numbers on the road. Let's skip to Neyer's conclusion:

Edmonds just edged Helton for fourth place. Maybe he shouldn't have. I certainly should not have dismissed Helton's road production, which was obviously considerable. But do you really want to argue -- with his road stats, anyway -- that Helton was better than Kent, Bonds, and Piazza? After including defense and baserunning?

Helton finished that season with 29 Win Shares. Kent finished with 37. The right man won the award.

Win Shares? Look, I'm not here to rip on Win Shares. Like Range Factor, VORP, skinny ties, and a geocentric cosmology, Win Shares had its time and place. But shouldn't a worldly guy like Rob Neyer (who, let's be honest, goes out of his way to flaunt his "progressiveness" every chance he gets) be a bit more circumspect when it comes to stuff like this? Who really should have won the 2000 NL MVP? Let's check in with Rally...

I'm not going to get into the ins-and-outs of Win Shares. Heck, I barely understand it -- and that's most of the problem. I think it's a great idea give the time in and purpose for which was conceived. If you want good criticisms of Win Shares, click here. But again, this isn't about Win Shares, it's about the 2000 NL MVP: How should we determine who "really" deserved it? Neyer's discussion (spurred by this reader's point that Helton wasn't just a Coors creation) focuses on the road stats of give players who finished high in that years MVP balloting, and so Neyer gives us this chart of road performances:

Again, seriously? We're going to go by by three-slashes (great for description, not so much for evaluation), runs scored and RBI? This is even worse than Win Shares. If only there was some "linear" fashion of "weighing" the relative offensive contributions of players that could include "adjustments" for "parks." I think we'd call it "location-sensitive relative massification" or something...

Granted, Neyer is writing for an audience with varyhing levels of knowledge of more recent (I hate the phrase "advanced metrics") stats and measures of player evaluation, so I understand this approach. But why bring up Win Shares? Outside of all the other objective criticisms, it's just so unnecessarily complicated.

So what would be a good way to decide which player was most valuable? What measure would account not just for offensive and defensive production, but also for position, and also the reality that runs at Coors are less valuable than they are at other parks... hmm... what could it be?

If you've been a reader here before, you know that the answer is Wins Above Replacement. It's been around in different  forms for a while now. There are currenlty  very good, freely accessible versions available at both FanGraphs (for 2002-2008, the seasons for which Ultimate Zone Rating is available using data from Baseball Info Solutions) and for all of baseball history at Rally's (AKA Sean "CHONE" Smith) incredible WAR database at Baseball Projection. You can read about the details behind the numbers here. Suffice it to say that Rally's database provides all the information we need: numbers for hitting (adjusting both for park and team context, so players don't get credit for more runs than their team actually scored), defense position, baserunning (not just steals, but also advancement), reaching on error, grounding into double plays, and playing time (replacement level adjusted for league and era). Sounds like exactly like the sort of "Uberstat" we're looking for. So what do Rally's numbers tell us about the five players mentioned in Neyer's chart duplicated above? The most important thing Rally's WAR tells us is that the list should be sorted differently.

I didn't do a comprehensive search, I just took the five players Neyer listed. I think the chart speaks for itself.

The great thing about WAR is that, properly implemented,  it takes all the "difficulty" and mystery out of comparing players in different run environments (parks/leagues), at different positions, of differing defensive skills, etc. It not only  gets away from Runs and RBI (while still using replacement level as a playing time guide to avoid the limitations of pure rate stats), but also uses linear weights to evaluate offense in an objective way that the three slash simply can't do on its own (i.e., trying to figure out how to weight each of the components [1.75*OBP+SLG is a good start, by the way]).

Jeff Kent was probably less liked than Barry Bonds, but was still an outstanding player (and borderline Hall-of-Famer, but that's another discussion) who helped his teams a great deal. But he wasn't even the best player on his team that season, despite actually being okay on defense. WAR isn't terribly precise, so one could make a case for either Bonds or Helton, it seems to me. But getting back to context -- these guys are all in the same league in the same year, so those considerations are irrelevant. It all comes down to a guy as smart as Rob Neyer seemingly "overadjusting" for Coors Field so much that he focuses on "road stats" for some reason. Why not simply reference park-adjusted stats?

Neyer writes:

But do you really want to argue -- with his road stats, anyway -- that Helton was better than Kent, Bonds, and Piazza? After including defense and baserunning?

Defense: Let's see, even though he played the easiest/least valuable position of the five, after you account for defense+position, the only player close is Edmonds, whom Rob already admits wasn't as good. 18 runs is 18 runs even at first base. Baserunning: Bonds was the better baserunner, but Helton wasn't bad, and clearly better than Kent and Piazza.

This is clearly a two-horse race between Helton and Bonds. One could make either argument. But to imply that it's sort of crazy to think Helton should be in the race is way off, especially in comparison to Piazza and Kent (the actual winner, whom Neyer tells us is the "right man.")

My main point isn't about the 2000 MVP Race, which is undoubtedly forgotten by most of us other than Kent, Bonds (who wasn't very happy about the result), Helton, and Neyer's correspondant. The main issue is that Neyer, who continues to be an excellent writer and analyst, obfuscates the issue with. It isn't just the "road stats" red herring, which is at least understandable on the level of "this is something my readers might get." But if you're going to bring up a total value "Uberstat" like Win Shares, why not go with one that's simpler, more intuitive, and more publicly available like WAR? C'mon, Rob!

It's not a big deal, but  hey, this is the blogosphere, where Little Things are a Big Deal.

NB: For another approach to MVP/Hall of Fame issues, see vivaelpujols' posts on Pennant and Playoff Probability Added.