Monday, April 16, 2012

Monday morning fullbacking

You know, the original intent of this blog was just to do a comparison of predictions vs reality. It's sort of grown into more than that, which is a good thing, but I've got to remind myself to stick with the simple stuff and not get too far off into the weeds.

Anyway, another week is in the books which means another chance to parse some data. Here's what the projected table (taken from the BS WPL game) would look like, compared to the actual table.

Team Predicted Points Actual Points Difference
Kansas City 14 18 -4
Salt Lake City 13 15 -2
Seattle 13 10 3
LA 13 6 7
Dallas 11 10 1
New York 11 10 1
Portland 10 4 6
San Jose 9 13 -4
Vancouver 8 8 0
Colorado 7 9 -2
Chicago 7 5 2
Columbus 6 6 0
Toronto 5 0 5
DC 4 8 -4
Chivas 3 9 -6
Montreal 3 4 -1
Philadelphia 3 4 -1
Houston 2 7 -5
New England 1 6 -5

There's actually not a huge change form how things played out last week, except that LA got their win. Still, LA has been the most underpforming team this year. Portland is still misfiring as well, and while that may be a bit of a surprise to people, I'm really not that shocked that a team that's been in MLS as little as they have are struggling. And Toronto of course is still there, because who thought they would be this awful.

On the other hand, Chivas gets no respect, despite finding ways to win games. Gotta love the goats for that. Houston's road trip continues and they continue to get results. It's going to be scary to see what happens to them when their new stadium opens up. And New England is still there as well with a pair of wins, despite fans assurance that they're awful. And don't get me wrong, I don't think they're that good either, but they're level on points with LA so that's gotta count for something.

And, for those interested, prediction accuracy ticked up a bit this week, going to 28/53, for 53% correct. Total wrong guesses (i.e. predicted Team A beats team B when Team B actually beats Team A) dropped to 15%. The remaining 32% is made up of matches that were predicted as a draw where a team got a result, or where a result was predicted but the match ended in a draw. But conservatively, call it a 53% success rate. Not too bad, I don't think. 20% better than picking at random, at least....

No comments:

Post a Comment