Doesn’t Kenpom factor in that we beat the snot out of them though? I don’t see why the system would care that much about beating lindenwood by 40 or a different cupcake ranked around 200 by 20.
Believe it or not, this is a more complicated question than it looks based on how Kenpom's engine works as there are a couple things to take into account including how would a team overperform against a team rated so lowly, and how he deals with blowouts which is not exactly 1:1. I will oversimplify this, just to show what I mean, so this isn't exactly how the system works, but it's somewhat ballpark. Let's say there's 3 major factors in determining how "good" a win is against a given team:
1. Actual Efficency Margin vs. Predicted Efficiency Margin
2. Was the Margin of Victory 30pts or less?
3. Location of game
Okay, let's start with margin of victory. In Kenpom's engine just know that when you beat a team by more than 30pts, it's considered a "blowout" and you get greatly diminishing return for each point over this margin to the point of it being almost zeroed out. Why is this? Well, most teams when winning by more than 30 in the 2nd half will start putting in their bench. So how much should teams be rewarded for say winning by 60 vs say winning by 35? The short answer in kenpom's system is not all that much. So any margin over this effectively gets muted. Now again, I am oversimplifying it, but let's put the "cap" at around 30pts and "absolute cap" at 35pts as how much beating a team by affects your adjusted efficiency margin in the metrics.
Now, let's consider Location of game. Kenpom adjusted metrics are based on neutral court matchups and know that playing at home gives you around 3 points (not exact but again, ballpark).
And now let's talk about predicted efficiency margin which again I'm oversimplifying it but ballpark we'll say it's your adjusted efficiency margin minus your opponent's adjusted efficiency margin (note in reality that tempo plays a part in this, but I am omitting it as it will just make things even more skewed).
Alright, so let's now look at Illinois' 300+ ranked opponents and see what Illinois would be predicted to beat them by (pretending tempo is 60 poss per game):
vEIU: [+22.58] (ILL adjEM) - [-21.53] (EIU ajEM) + [3.0] (Home game) =
+47.1pt predicted margin of victory
vMonmouth: 22.58 + 12.52 + 3.0 =
+38.1pt
vLindenwood: 22.58 + 16.94 + 3.0 =
+42.5pt
vAlabama A&M: 22.58 + 16.94 + 3.0 =
+42.5pt
vBethune Cookman: 22.58 + 12.98 + 3.0 =
+38.6pt
Now remember what I told you earlier, that your "absolute cap" for how much your margin of victory will affect your adjusted Efficiency Margin in any game is somewhere in the ballpark of about 35pts. And note how in every single one of those games for a "good win" you'd need to beat those teams by over 35pts, which from an adjusted efficiency standpoint can't happen no matter how much you beat them by. So, in other words, every single game you play against a 300+ ranked team in Kenpom's system is for all intents and purposes a win that is worse than you are expected to beat them by, hence why your metrics are artifically deflated for playing teams of this caliber. Now compare this to say a team ranked 250 who you should beat just as easily and your predicted margin of victory is +32.1pt. In other words, you can still get a better than predicted win against them.
One other thing to note is that when it comes to SOS and non-conf SOS most system metrics weight all teams equally, which when you consider the average B10 team you're playing is ranked #32 and the median non-conf opponent is #181, games against #300+ teams skews your average hard. Hence while we've played 3 Kenpom Top 10 team on neutral courts, our non-conference SOS is still only 217th in the country.
I apologize to everyone for the long post, but that is about as simple and efficient a summary I can give for why in Kenpom's metrics, playing 300+ ranked teams will always negatively affect your adjusted EM and SOS metrics regardless of how much you beat them by. Basically, you automatically lose just by scheduling them. Also know that this is a known "issue" (it's not really an issue or bug, it's just a limitation when it comes to how your system deals with blowouts) with Kenpom and most other metric based systems and when you schedule opponents coaching staffs and administrations in power conferences are well aware that you want to schedule cupcakes to get wins, but you don't want to schedule cupcakes that are too fatty. Hope that explains things.