Team WAR and team wins

So, the first time I did this, I made a little mistake. But I'm no quitter. I inadvertently only included postional player WAR the first time around, but now I've fixed it. But after fixing it, everything I said previously still pretty much stands. Keep reading:

----
Rob Neyer answered some questions today for Camden Chat. It's all good stuff, and I just happened to file something for the Hardball Times on the Orioles before I read this. But something in particular piqued my interest:

Camden Chat: Newer statistics, particularly Wins Above Replacement, cause a number of disagreements on our site. What are the cracks in the WAR metrics that would allow a team to substantially outperform (in the win column) their total WAR over the course of a season?

Rob Neyer: Do they? I've actually never seen team WAR compared to team wins. Just off the top of my head, WAR doesn't account for baserunning (except for steals). Nor does it account for luck. And that's the big one. The vast majority of fans simply don't understand the degree to which luck impacts wins and losses, even over the course of 162 games.

As luck would have it, I was doing the equivalent of artist doodling a few weeks ago, otherwise known as screwing around with baseball stats and excel. Well, I did something that relates exactly to Camden Chat's question: what are the cracks in the WAR metrics that would allow a team to substantially outperform (in the win column) their total WAR over the course of a season?

You can check out what I did below  for the 2010 season (I'll also add it to my box.net tool on the right hand side, so you can get your hands on the actual spreadsheet). Basically, I looked at win and loss records, run differentials, pythagorean wins and losses and ... WAR. I grabbed each teams WAR from FanGraphs, figured out what replacement was (basically 81 wins minus the average team WAR, which ends up being 45*), then added their team WAR plus replacement to find out how many wins their WAR totals would have predicted. I then calculated the delta between their WAR and actual number of wins -- this is essentially the question Camden Chat was asking.

*Note that the first time I did this a replacement level team would win 60.79 games. That's too many, and I knew this. But I forged on. Now that it's 45, I'm very comfortable I've done this right -- until someone tells me I'm wrong, of course.

If you take a look, you'll find that the Diamondbacks really underperformed in the win column versus their WAR total. That is no shocker to me, as their bullpen was truly atrocious and that will cost you a ton of games. Another team that posted an excellent WAR but turned in a rough season was the Brewers. I'm thinking their additions of Shaun Marcum and Zack Greinke may push them to the top of the pile in 2011. Finally, the Phillies were the team that were the "luckiest." I don't know what the implications are of that, but Rob mentioned base running and they've been excellent each season in that. Davey Lopes was one of the most efficient base runners around, and he's modeled the Phillies after himself. Unfortunately for them, he's gone next season.

Something I noticed after getting it right was the Rays. They are technically the second "luckiest" team. I think this jives with the fact that base running (non-stealing bases gained and lost) isn't accounted into the WAR caclulation. I know for a fact that Carl Crawford is a dream in that particular category, and I do believe the Rays were known for it in general. Oddly, it was this very thing that cost them big time against the Rangers when they were bumped in the first round.

I never really knew what to make of what I created, but here it is for your pleasure. I personally think it might give us some insight into which teams might be closer to contention. Also, I think it certainly shows which teams know there way around the bags. That said, draw your own conclusions.

Sources: FanGraphs for WAR