It's futile to try to find any one factor (offense v. defense) that dictates a single elimination tournament winner. The best teams have the best chance to win, but in any one and done format, there are going to be randomness to the results (makes the "hire Odom"; "hire Moser" posts based on NCAA results beyond stupid). Here are the last 8 champs, and their KP rankings entering the tournament and the O and D ratings after the tournament:
2011 - UCONN #16 entering the tournament; entered the NCAAs as 3 seed; final O and D ratings: 19/15
2012 - KY #1 entering the tournament (had been the #1 rated team since mid-February; also, the #1 team to start the season) entered the NCAAs as a #1 seed; final O and D ratings: 2/7
2013 - L'ville #2 entering the tournament; #1 seed; final O and D ratings 7/1
2014 - UCONN #25 entering the tournament; (#44 in the nation when conference play started); #7 seed; final O and D ratings: 39/10
2015 - Duke #6 entering the tournament: #1 seed (the last #1 seed); final O and D ratings: 3/11
2016 - Nova #5 entering the tournament: #2 seed final O and D ratings: 3/5
2017 - UNC #3 entering the tournament: #1 seed final O and D ratings: 9/11
The finalists tonight:
Michigan - #10 entering the tournament; during the season, bottomed out at #44 after a 17 point loss to UNC in late November; have made a steady climb during the year; #16 entering the Big 10 tournament; #10 entering the NCAAs (#3 seed); now up to #7; O and D ratings: 31/3
Nova - #2 entering the tournament (behind only UVA); has never been rated lower than #2 all year: current O and D ratings: 1/14
KP projects a 74-68 final and gives Nova a 69% chance to win. Vegas line is Nova -7 O/U is 144.5. FWIW, the #1 team entering the tournament has one the NCAAs only once over the last 8 years.