Essentially, they're valuing Kentucky's additional three tournament appearances and additional two sweet sixteen's over UConn's additional two championships.
What they are doing is valuing Kentucky's one-and-dones more than our Kromahs and Giffeys. Simple as that.
The results are dependent on the algorithm they are using.
If you define "program success" as consisting substantially of "guys in the NBA," then you get one result. If you define "program success" as consisting substantially of "winning when it matters," then you get a different result.
The use of the algorithm is, really, self serving. They figured out which programs they thought were the best, and they went backwards from there to derive the algorithm they should use.
Frankly, my definition of "best program" would be heavily weighted toward, "what did you do with the talent you had?" and Kentucky wouldn't be top 5, and neither would Kansas or Sadexcuse. But, hey, I guess having more and doing less with it is high on some people's list.