P(t) = F ● v

Time to churn out some ratings. I'm still not sure if these are fair to teams that don't play, and even though RSL did pretty good this week they still dropped in rating a little, which on the surface seems odd, but then again some other teams are playing pretty good as well, so I'll stick with it. Here's how the ratings look after 8 weeks.

I suppose I should also note that Chicago, Colorado, and Houston are all tied, so place them as you will.

Team	Rank	This Week / Last Week	Rating	This Week / Last Week
Kansas City	1	0	3.78	-0.90
San Jose	2	0	4.33	+0.42
Seattle	3	0	6.44	+0.43
Salt Lake City	4	0	7.33	-0.21
New York	5	0	8.33	+0.17
Vancouver	6	0	9.00	+0.38
DC	7	+2	9.56	+0.94
Chicago	8	0	10.56	-0.56
Colorado	9	+1	10.56	+0.32
Houston	10	-3	10.56	-0.68
Dallas	11	0	11.33	-0.08
LA	12	+1	11.78	+0.22
Columbus	13	-1	12.44	-0.94
Chivas	14	-1	12.78	-0.78
Philadelphia	15	0	13.56	-0.06
Portland	16	0	14.44	-0.44
New England	17	0	15.00	-0.25
Montreal	18	0	15.78	+1.35
Toronto	19	0	19.56	-0.18

...post-dictions?

So I dropped the ball this week. Or, more accurately, worked and vacationed and opted not to bring a laptop, as relaxing was nice for a change. But one of the nice things about using data to make predictions is that you can go back and see how you would have done even after the results are out. So, parsing the numbers for last week, the predictions owuld have been as such...

Salt Lake City and Dallas draw
Montreal and Portland draw
New York beats New England by two goals
Philadelphia and San Jose draw
DC beats Houston by one goal
Vancouver and Columbus draw
Chicago and Seattle draw
Salt Lake City beats Toronto by two goals
Chivas and Colorado draw
LA beats Dallas by two goals

And then we can even do a quick look back and compare. The predictions went 4/9, with nothing completely wrong this week. On the season, the model is 35/71, so 49% accurate. Again, not great, but about on par with what MLS "experts" are doing. So I'm still pretty content with it.

Updated power rankings and such to come later.