Re: Season 95, Vol 2: Michigan Tech '15-'16
Have no idea how you'd do it (or if it's worth the time), but do you or does KRACH take into account momentum? Take a Northeastern this year... My guess is their probability of winning this weekend is much higher than KRACH or their PWR suggests. Terrible start, but maybe one of the best teams out there in the second half.
Maybe some kind of rolling average on win % and SOS? Again, not sure it's worth the headache.
KRACH is reflexive: it only tells you who has done what, and the ideas of relative strength are, per the originator, not supposed to be predictive.
That said, you can view BELOW in a momentum-like manner. Through most of the second half, I would add last-10 or similar columns to the tables for my weekly rankings posts. For example, there's the penultimate one:
http://blog.wchaplayoffs.com/2016/02/28/2015-16-week-22-below/. If you took the 2016 results for each of those teams, you find the following:
Tech (+82) was rising hotter than NMU was (+21).
Ferris (+2) and Lake (-5) were playing pretty much at their potential over the last few weeks.
BGSU (+8) and UAH (+3) are who we thought they were, which makes that split look very strange.
BSU (+73) was rising up, while MSU was flagging (-21).
UAA (-125) was in freefall, while UAF was just merely down after their early precipitous fall (-18).
Mind you, what happened that week was largely predictable, but if I'd said, "Let's let some concept of momentum help drive our decision," I would probably have gone with four of those five results (I had BGSU creaming UAH while trying to get to the MacNaughton).
It's something that I'll have to evaluate over the summer as I tweak a predictive model.
I've done high/low; I can do week-by-week graphs if people are interested.
http://blog.wchaplayoffs.com/2016/03/06/2015-16-below-beginning-to-end/
GFM