The Tribe on the WARpath

No, I'm not referring to the Wahoos' recent thrilling six-game post-All-Star win streak; I'm talking about their WAR (Wins Above Replacement) thus far this season. I typically use OPS+ and ERA+ to quickly describe a player's performance (100 is league average and higher is better), and I think they're both excellent metrics. WAR is nice too because it provides a convenient benchmark: a value of 0 means that a given player is no better than a "replacement player," i.e. some Average Joe just called up from AAA. WAR then measures how many team victories this player is better than worse than Average Joe (Average Crowe?). Sabermetrics Library has this to say about WAR:

If you had to pick one statistic – and only one statistic – to use in evaluating players’ value to their teams, Wins Above Replacement (WAR) should be it, end of story.

Well, alright then. What I was wondering the other day is: how well does this statistic correlate to salary? Since I'm a Tribe fan, I wondered this specifically for the Indians, because it seems that everyone on the team is either wildly overpaid or wildly underpaid. You can figure out who is who. So, I went ahead and charted WAR vs 2010 salary for those who've donned the Chief thus far this season, using a generic salary figure of $200K for minor-league callups and players who haven't been with the club all year. I took all other salary and WAR data from Baseball-Reference.com; I understand Fangraphs uses considerably different calculations for WAR, but that's OK.

If players are actually being paid properly for their talents, then there should be a strong correlation between WAR and salary; in other words, the points on these graphs should trend up and to the right. If there's no correlation, it should be a random scatter plot, and if there's an inverse correlation, points should trend up and to the left (lower salary brings greater performance). Clearly, the most value to the team is gotten from players who appear at the top left of the chart; cheap and good, the way we like them. Let's see the hitters first, with selected players noted with text. Click the graph to enlarge.


I find this graph utterly fascinating. There is almost zero relationship on the 2010 Cleveland Indians between pay and performance. In fact, if you take off the Hafner data point (which skews the set because of how damn much he makes), the R² becomes 0.002, which is for all intents and purposes completely random.

While I'm on the topic of Hafner, have you ever noticed how much noise players, especially NFL ones, make about wanting "fair" deals when they signed a long-term deal that they want re-done because they've exceeded expectations? "Fair" is athlete code for "as much as possible." Funny, they never say a word about wanting a "fair" deal when they're underperforming their contracts.

Shin-Soo Choo and Carlos Santana are by a far sight the most valuable members of the club on the field, and this graph adds another dimension to their performance; they're dirt cheap as well! WAR is a cumulative stat, so don't be surprised at all to see Santana end the year in the 4-5 range, on a league minimum deal. For all the bellyaching people have done about Rusty, note that he was worth more on a $1.5 million deal than Travis Hafner ($11.5M) and Grady Sizemore ($5.77M), and just about as much as Jhonny Peralta ($4.85M). Eat it, Rusty-haters.

Austin Kearns was a nice value signing at $400K and a WAR of 1.2, while Mike Redmond (-0.6) and Mike Brantley (-1.0) have thus far proven themselves quite unworthy of Major League roster spots. Perhaps we should stop allowing Mikes on this team. Beyond those players I've noted, you can see the hallmarks of a rebuilding and slightly overachieving team, as most players have posted WAR numbers above or near 0 on slim contracts. Let's just say that the blue dots on the Yankee graph would be shifted considerably to the right.

To me, this graph says quite a lot about baseball economics. Essentially, young kids who can play a bit are terrific values, and big contracts absolutely kill you. Every time I see Hafner hit for the rest of the year I'm going to think of his little blue dot way the hell out to the right but below five of the Tribe's regulars, all of whom make less than half what he does and four of whom make less than a seventh. Consider the combination of Hafner and Sizemore, our two most highly-compensated position players. Their WAR combined is: 0.1. That's not a typo. We're shelling out over $18 million (over a quarter of the total payroll) for one-tenth of a win more than two average bench players provide. I'm speechless.

Let's take a look at the chuckers:


Same story here, too. The R² is a miniscule 0.007, dropping to 0 if you take away Jake Westbrook. Once again, we're getting staggeringly little value from our two highest-paid players. Add Jake and Kerry Wood together, and you've paid $23 million (about 38% of the team payroll) for a grand total Wins Above Replacement of:

Zero. 0.0.

This is literally saying that you could have had, for about $800K, what the Indians have received for $23M this season. Now, I understand that there's always risk in signing a player to a big deal, but this shows just how tough of a proposition paying veterans competitive salaries is for MLB ownership.

You can argue that Fausto Carmona has earned his money; he's paid a lot, but at least he leads the staff in WAR. Mitch Talbot and Chris Perez are both very good values, pulling in about $400K each and posting solid WAR numbers. Way down at the bottom of the chart, one can see that it might have been more than just his fondness for Twitter that kept David Huff in AAA last weekend.

After looking at these numbers, I think maybe I understand the Pittsburgh Pirates' ownership model. I also understand why the Indians have chosen to devote an increasing share of their resources to drafting and player development. Why pay $10M for a guy to nab you two wins when you can pay $400K? These data should also come as a rebuke to those who chastise Indians ownership for not shelling out more for free agents. If I saw these graphs, I wouldn't either.