So I've developed this algorithm to predict scores and overall points based on two very, very simple factors. Running it against Iowa's season so far, it correctly predicts the total points scored in Iowa's games to 99.88% accuracy. Here's the predictions versus reality:
PREDICTION: NIU 32, IA 29 (final: 30-27)
PREDICTION: IOWA 30, ISU 21 (final: 27-21)
PREDICTION: IOWA: 33, WMU 17 (final: 53-3)
PREDICTION: IOWA: 27, MN 23 (final: 23-7)
PREDICTION: MSU 23, IA 19 (final: 26-14)
Weighting by points, the accuracy on points is 94.06%, but if I add home field advantage, I bump up to anywhere from 98.21 to 99.88% accurate on points, depending upon the coefficient I use in weighting the point differentials from prior games.
Anyway, the point is to get the Super Action Prediction Machine's prediction on record for Ohio State. Survey says:
If I use my standard coefficient of 1.5, I get 48-23 Buckeyes. If I double the coefficient, I get 37-24 Buckeyes. Reading between the lines, the math says that Iowa performed poorly on offense against legit defenses, and while weighting the points more helps the defense, there's just not enough firepower on offense.
For reference, using a standard coefficient of 1.5, the Buckeyes are predicted to score 48 and Iowa scores 23.
Using a double coefficient of 3.0, the Buckeyes score only 37 and Iowa scores 1 more point (24).
It doesn't matter how high I jack up the coefficient, the logorithmic scale combined with Iowa's mediocre offense doesn't ever get us anywhere.
The MOST optimistic model I can arrive at is OSU 31, Iowa 24. The logorithmic effect of the weighting coefficient tapers off so quickly, the record difference between the two teams is just never overcome.
BEST CASE: OSU 31, IOWA 24
WORST CASE: OSU 56, IOWA 21
STANDARD COEFFICIENT: OSU 48, IA 23
ADJUSTED COEFFICIENT: OSU 37, IA 24
The math says take the over and the points.