ESPN.com – MLB – MLB Standings

The “expected won-loss” record of the entire National League is 244-242. Bit of a problem there.

Skip to content # Rounding, I presume

## 9 thoughts on “Rounding, I presume”

## Leave a Reply

ESPN.com – MLB – MLB Standings

The “expected won-loss” record of the entire National League is 244-242. Bit of a problem there.

I’m guess that it’s a rounding problem? They use the RS^2/(RS^2 + RA^2) method, right?

Yes. Since the runs scored and runs allowed are equal (before interleague play) the record should be .500. If it’s not, it has to be rounding.

Are you sure that games like Hotlanta’s 16-0 drubbing of the Astros don’t count for more than one win? I think they should. That was a 2-win game if ever there was one.

It should be something like RS^1.85/(RS^1.85 + RA^1.85).

To whatever power you take it, RS = RA, and RS^n = RA^n. So x = .5.

I don’t think I’ve ever felt more out of the loop with respect to baseball statistics. You guys are great fans of the game.

Don’t be. I just exhausted my algebra. Basically, what I mean is that if two quantities are equal, the ratio will always be 1:1 no matter how you manipulate them. If x = y then x times any number — 2, 10, 1,678,459,002 — is equal to y times that number.

I am too lazy to actually look up the link, but if the expected W/L is a sum of the individual team pythag W/Ls, rounding error could accumulate to a large enough magnitude to make this difference. If that’s not the case, then you’re right Mac, they’re using some funky math.

Clearly it ought to be rounding error. The Braves have scored 149 and scored 100. By the original Pythag, that should be a .689 winning percentage. Over 31 games, they “should” have won 21.37 games. Except in arguments with my wife, I can’t think of a way to win a partial contest.