View Full Version : New statistics on espn page
07-03-2008, 11:08 AM
I just noticed that on the standings page on espn.com, they've added three new categories for the percent chance a team will win the division, the wild card, or make the playoffs. How do they come up with these numbers? Does anybody know if they are using some predictive formula, like Pythagorean winning percentage, to project each team's record, and then figure out the chance that they win their division or wild card? It seems to me that they must be doing something else using probabilities, because their results are expressed as probabilities.
As a side note, I recently looked at the Pythagorean winning percentage formula that everyone is used to, and came up with a slightly more accurate way to predict a team's record based on its runs scored and allowed. I was going to write it up and perhaps find a place to publish it until I found out that a math professor at Brown University had already published almost the exact same thing.
07-03-2008, 01:15 PM
I just noticed that on the standings page on espn.com, they've added three new categories for the percent chance a team will win the division, the wild card, or make the playoffs. How do they come up with these numbers?
07-03-2008, 01:19 PM
Just about every statistical analysis had the 2005 Orioles and the 2007 Mets in the playoffs.
07-03-2008, 03:14 PM
Actually, they're using CoolStandings to provide this. The guy that runs it is a cool guy.
Look, if you don't like the pythagorean stuff, the "dumb" setting is still very useful. It doesn't take that projection into account. It assumes a 50% chance of winning each game- flips a coin.
No, it does not predict the future. But it does actually give you a picture of how "in it" your team really is- especially at the end of the season.
To each their own. Its interesting to know that if the rest of the season were coin flips, the Sox would make the playoffs 68/100 times.
07-03-2008, 03:18 PM
Bah. Just about every statistical analysis had the 2005 Orioles and the 2007 Mets in the playoffs.
This is where you get it wrong. People who dig this stuff (me, for example) don't claim that, say, the White Sox *WILL* make the playoffs because of these calculations. Its nice to know they have a better chance- and what that chance really is.
Last year in August, the Rockies had a 7% chance of making it. I was not wrong for saying that in July '07. I was right. They did. And their miracle run got them in from "almost no chance." What's wrong with quantifying that small chance?
You say "they have barely a chance to make it"... and I say "right, they have a 7% chance of making it".
I don't get the hatred.
07-03-2008, 03:58 PM
Here is another interesting site that tries to figure the odds of making the playoffs for several sports.
07-03-2008, 11:26 PM
I like it. Definitely something interesting to follow as the season progresses.
It reminds me of a feature ESPN (or maybe it was somewhere else) had in their in-game updates years ago that told you what % chance each team had of winning the current game based on the situation. I'm a sucker for probability statistics.
07-03-2008, 11:41 PM
I don't get the hatred.
For the umpteen millionth time, nobdy hates statistics.
People hate the misuse and abuse of statistics.
vBulletin® v3.8.7, Copyright ©2000-2013, vBulletin Solutions, Inc.