Early season sense and nonsense

Part of the fun of early season baseball is trying to figure out which among the hot teams and players are for real and which ones are flukes.

The most interesting early season standings are in both leagues' Central Divisions, where in the NL, the top 4 teams are essentially tied and in the AL where the Cleveland Indians, picked by many to finish last, are in first place, looking impressive, and accumulating a bandwagon among the pundits, if not their fans (though it may be 35 degree damp weather that lacks fans rather than the team itself). Hot in pursuit of the Indians are the surprising Kansas City Royals, followed not so closely by perennial contenders Chicago, Detroit and Minnesota.

So when I look at coolstandings.com, a site that updates playoff odds on a daily basis, why do they show clear favorites in both divisions? It's not because of the records, it's because of the runs. If you look closely at the NL Central, you can see that St. Louis has outscored its opponents 108-78. Feed this into what Bill James calls the Pythagorean Theorem (runs scored squared divided by sum of runs scored squared plus runs allowed squared equals winning percentage) of baseball and you get a .650 winning percentage, far above the team's actual .550. This implies that the Cardinals are underperforming and are likely to do better as the season progresses. Cincinnati and Milwaukee, on the other hand are only slightly underperforming and should do better, but not as well as St. Louis. The Cubbies have scored few runs than they've allowed and are projected to a .380 winning percentage, which translates to 61 wins for the season.

In most of the other divisions, teams are performing about how you'd expect. Atlanta is a clear underperformer. It's hard to be 9-12 when you've outscored your opponents overall. In the AL Central, the Indians have the highest run differential in baseball, 99 scored vs. 66 allowed. This suggests that their fast start is not a fluke, while the Royals have only a 104-94 differential, having won a great many close games. That's still good, but projects to a more .500 kind of record. The defending division champ Twins have the worst run differential in baseball, 57-89. You don't want to know what that projects to.

Obviously, this is not the only measure one uses in projecting team performance, but combine it with strength of schedule and how closely players' performance relates to their actual skills, and you can get a decent picture of how it lays out. Coolstandings includes a strength of schedule component, which is why their projected wins are not just their Pythagorean projection.