The Frick League by Base Runs

[RonCo]

Another fun little chart on Stats+ is the use of Base Runs to calculate expected wins. Base runs is similar to the Pythagorean estimation in that it appears to be driven by run differential—but instead of using actual runs, it uses underlying stats created by a team to calculate the number of runs that team should have scored or allowed, and then estimates wins off those numbers.

I got interested in looking at these numbers, because there are some interesting swings. For example, tied for the largest “positive swing” (positive meaning that the team has won more games that Base Runs suggests they should have) is in the Pacific Division, where the Base Runs standings would look like this (with the variance between the projection to actual wins in parenthesis):

California 79-41(-2)
Long Beach 65-53 (-2)
San Fernando 63-56 (+8)
Valencia 54-66 (0)
Hawaii 50-66 (-1)

In other words, California has won 77 games when the projection says they should have won 79. So the Crusaders are -2 against their estimation.

I’m sure you can see the outlier.

Given the underlying stats they are creating, San Fernando has won 8 more game than the system projects they should have—which is a pretty amazing number, especially given that the season is only 115-120 games in.

It’s weird season, it seems. Online leagues can almost certainly have broader results than real teams (of one of many perfectly good reasons, I suppose). And if you scan out history you’ll see teams occasionally getting an 8-10 game swing in pyth. So maybe this isn’t too odd. And if you pick through some of the base run years you can find a few bigish numbers. But how many?

Looking at the Frontier we see even more warping:

Seattle 64-52 (1)
Calgary 54-64 (-6)
Edmonton 54-66 (-11)
Boise 47-70 (8)
Vancouver 40-72 (2)

I mean, holey struggles, Batman, but look at Edmonton! They’ve won only 43 games whereas their underlying numbers say they should be at 54—which may or may not be fantastic, but at least might bring some fans to the seats. Boise goes the other way, matching San Fernando by winning 8 more games than they “ should have.” Vancouver is, well, it’s a project.

And in the Heartland?

Yellow Springs 72-46 (0)
Omaha 66-52 (7)
Twin Cities 63-57 (-7)
Madison 54-64 (2)
Des Moines 52-68 (2)

Us folks in Yellow Springs like the way this one looks, of course. Instead of being in a death match with Omaha, we’d be fairly comfortably in the lead. But the Hawks are seven games better than their estimations and Twin Cities is seven games worse.

So, what does it mean?

I dunno.

But I’m finding this one interesting. And I’m finding certain comparisons interesting (for example, the River Monsters were +11 last season, but -7 this season…makes you want to dig into some spreadsheets, eh?).

Anyway, play around with the stuff yourself.

Let me know if you find anything interesting. I’m all ears.

[Gareth]

YS stats here is interesting in a way. Is it good that the predictor is right on and your squad is performing as it should or would you like to see a bump in the differential? Seeing as your record is pretty decent, I’d be happy with the result.

[RonCo]

Good questions. Like most of these things, I figure they're what you want them to be.

Personally I read them as an indicator of luck or some odd nuance the measure doesn't cover well. In that sense, I guess I like that we're right on target...except I'd like to be both lucky AND good.