2010 Big Ten Conference Recap: Adjusted StandingsPosted: August 17, 2010
To kick off my first analytical piece of the Big Ten as a whole, I’m going to present the Big Ten standings in a bit different fashion. One way to adjust the standings is to find a Pythagorean Record. Basically, you square a teams runs scored and runs allowed. You then take the squared runs scored and divide by the sum of the squared runs scored and runs allowed. I don’t want to say that this has become a “faux pas” among Sabermetrics recently, but it’s not the ideal method.
What I’m doing is using a Linear Weights-based BaseRuns method to estimate a runs scored and runs allowed total. In honesty, I’m stealing this from Patriot, a well-known Sabermetric writer.
Below, you will find the 2010 Big Ten standings based on a win percentage derived from my runs created/allowed methods:
The teams are sorted, obviously, by EW% which I’m calling a teams Expected Winning Percentage. This is based on the estimated runs created and allowed. PW% is the same method, only with actual runs scored/allowed inserted into the formula and Act. W% is the teams actual winning percentage.
Here, we see that the best team based on actual winning percentage — the Spartans — drop to second overall despite the best pythagorean win percentage. Purdue soars from third-to-first and the Wolverines drop from second to fourth.
Judging from the table, Michigan State looks to have played how we expected based on their numbers, while Purdue under-achieved a little bit. I hesitate to call teams “lucky” or “unlucky” but if you prefer those terms, by all means, use them.
This metric was not a fan of the Hawkeyes, which makes Iowa’s run to the Big Ten Tournament title game all the more interesting. They were ousted 15-5 by Minnesota in Iowa’s second game of the day.
But why is the EW% down on some teams, and not others? Well, here’s a table of each teams actual runs scored and allowed, as well as their runs created and runs created allowed.
|TEAM||RS||eRS||RA||eRA||RS Diff||RA Diff|
Here, I’ve got actual runs scored (RS), estimated runs scored (eRS), actual runs allowed (RA), estimated runs allowed (eRA), differential between actual and estimated runs scored (RS Diff) and the differential between actual and estimated runs allowed (RA Diff).
What we see here is that Michigan State actually scored 22 more runs than their underlying statistics would indicate, while giving up two less runs than estimated. Purdue rate so highly because the run estimations though they scored what they should’ve, but think the Boilermakers should’ve given up fewer runs.
On the whole, the Hoosiers led the conference in estimated runs scored and in estimated runs allowed — not a good combination. The model estimated Illinois should’ve scored 311 runs which is the lowest total in the conference.
As far as the differentials are concerned, Michigan State had a very high estimated runs total, but they still outscored it by 22 runs with Michigan being the only other team to best their estimated runs scored by more than 14 runs (and they were at 20).
Meanwhile, Northwestern’s putrid offense was estimated to be much better. They were estimated to score a full 47 runs more than they did. Ohio State and Minnesota both scored 28 runs less than estimated.
On the pitching/defensive side of things, the Hawkeyes gave up 25 more runs than estimated — a full ten more runs than the second biggest differential; Indiana gave up 15 more runs than estimated.
In terms of over-performing, the Wolverines gave up a robust 36 runs less than they were estimated to allow. The Buckeyes were second, giving up 21 runs less than estimated.
So it’s pretty easy to see why the Buckeyes come in in the middle of the pack of the estimated win percentage: they scored more runs than they were estimated, and gave up fewer runs than estimated.
Still, you have to feel for the Wildcats. Scoring 47 runs fewer than you were expected to, while giving up three more than estimated is a rough, rough season.