I know this will be a TLDR post for many who prefer to creative narratives or bash numbers, rather than try to understand what something is doing and its strengths and weaknesses.
I respect the system and what it does (but I understand its limitations) and I have no agenda except understanding numbers.. but at the same time, myself like KP, will also say KP should not be anywhere close to a primary basis to select NCAA teams, and at best a very low secondary basis for selecting item. And it's not a primary factor or even really a secondary factor for choosing NCAA teams. Which I agree with!
And the W by Syracuse today is much more important than whatever our KP movement is from one isolated game in a sample.
Its a margin based system done to rank 362 teams in order to help people predict outcomes. Good luck trying to rank 362 teams on straight wins and losses or "Eye Test", when there are such large variances in schedule strength. You have to use margin to try to rank such a large amount of teams. You can try do those types of things (Who beat Who) when ranking the top 25, but you can't do that effectively when its a larger group
The change by Syracuse was totally as expected. I'll explain the math.
It was basically a pick em game, since Syracuse has the 3.5 point advantage being at home.
We won by 8 points, so 8 points above expected margin (before considering pace)
8 points / 13 games / .724 (our tempo of 72.4 possesions) = .84 improvement in EM
Our current Adjusted Efficiency Margin of 9.58 is #81
Before the game it would have been around 9.58 - 0.84 = 8.74, which would be #87 as of now.
Margin matters, not just straight W and L, for KP.
For example if we lose to Duke by 1 point, I ran the calculations and our EM would improve by about 1.4 and move our ranking up to about #71 or #72 And if we win by 1 point, it would be 69 or 70. Not much difference in either ranking because its only 2 points. But of course the difference between Winning and Losing by 1 point would be massive for our tournament lives -- as it should be when choosing teams.
So what do I like about KP - its a good way to get a general understanding of how good a team is, by looking at its "range" a 60-80 type team for example, which is quite helpful for mid-majors that we never see -- to me that is the best part of the system. It also predicts teams ultimate outcome better than the current AP Top 25 as a whole for example. Again KP is being using to rank 362 teams, and help us predict outcomes for 362 teams.
But I would never use the simple KP ranking to pick a P6 team for the tournament, and what P6 teams should get in. That should be based on the straight quality of your W's and L's.
Nor would I use it to prepare my current top 25, "AP Style".
That being said the top 25 in KP as of today, will very likely end up doing better in conference and end of year and the tournament, then the current AP top 25 who will have more teams tank a bit. That is because KP is a good predictive system and will predict where teams end up better than the current AP Top 25. Of course there will always be exceptions and anomalies,
