I remember a guy who was successfulat picking winners used a form of your statistic but would also correlate it into the % increase from your points allowed
For example 28ppg and giving up 14ppg wasn't as good as averaging 21 ppg and giving up 10 ppg even though the gap was greater. Anything over 1.00 was a really good team
He could predict super bowl and playoff winners with simple math especially late in the year
Using these numbers
Alabama would be 31 divided by 18. 1.72
Clemson. 28 divided by 18. 1.55
Notre Dame. 21 divided by 17 1.23
Ohio state. 24 divided by 23 1.04
Essentially he would then come up with a pick It looks like there is a gap from Each team especially Bama and Ohio State