Obviously Self coaching KU is better than Norm Roberts.
But, the points I made in my original post stands. It's using analytics to calculate the standard deviation between a team's winning percentage and their expected record. High luck rating means you're probably over seeded. And you never see teams in the top 25 of the luck metric year-after-year the same way you'd see teams annually of the AP Top 25. Why? Because it's not a sustainable outcome. Superior teams don't need the final 1-2 minutes of a game to win. Self has a winning percentage of 81.5% at KU. Mostly because he wins games in the first 38 minutes. I don't need to look up the stats to know that he absolutely isn't winning 81.5% of the games that were decided by 5 points or less.
Look at KU's 2 National Title teams under Self, and for good measure, we'll throw in 2020. Their ratings in the luck metric were 138th in '08, 79th in '20, and 88th in '22. No doubt their rating went up after winning 6 games. So yes, Self's best teams weren't especially lucky, they were just good.