There are 21 Coach K years with publicly available KenPom rankings (20 tournaments due to COVID). Here's a summary of their final KenPom rankings:
Top 10: 16 (76%)
Top 5: 10 (48%)
Top 3: 8 (38%), last in 2018
Top 2: 5 (24%), last in 2011
#1: 3 (14%), last in 2010
Their final KenPom rankings under Jon Scheyer (four years) look similar if not better:
Top 10: 3 (75%)
Top 5: 2 (50%)
Top 3: 2 (50%)
Top 2: 1 (25%)
#1: 1 (25%)
You could argue that the last couple years under Coach K were also below standard, but that was really just his second-to-last year (affected by COVID). Ignoring that year wouldn't change the percentages much.
But Duke did win two national championships during those Coach K years. A very unreliable AI guess is that they "should" have won ~2.45 during that time based on their pre-tournament odds (or estimated odds using KenPom ratings), so right around what they did.
Another way to look at it: in those Coach K years, they won 1 championship for every 10 years, or 1 for every 5 years they finished top 5, or 1 for every 4 years they finised top 3. Under Scheyer, they haven't yet had enough total years or years in the top 5 or top 3 to say they're necessarily doing worse.
I think this also highlights how surprisingly low the championship odds are even for top tier programs, and how the path to a championship is to get many years with a top 3 final KenPom ranking.
