If you took BYU's preseason predictions from Ford, he had BYU with the following likelihood of winning
- 75% vs SIU
- 15% vs SMU
- 57% vs Wyoming
So that adds up to 1.47 expected wins and a 6.4% chance (based on preseason and assuming independent results) of being 3-0
Meanwhile UNLV had the following odds
- 44% vs Houston
- 99% vs Utah Tech
- 12% vs Kansas
Which adds up to 1.55 expected wins (more than BYU) but (with same caveats as above) only a 5.2% chance of being 3-0. It's just that UNLV had a guaranteed win, whereas BYU did not.
Same story with NIU
- 99% vs Western Illinois
- 2% vs Notre Dame
- 71% vs Buffalo
Means 1.62 expected wins (more than either BYU or UNLV), but a miniscule 1.4% chance of being undefeated
Bottom line - we excel here by having more toss-up games instead of a mix of expected easy wins and unexpected upset victories