The Dr. K. NCAA Football Forecasts

2018: Season Over


The following predictions are based completely on the Kambour football ratings .

The results are sorted by start time.

The numbers in parentheses represent the point-spread. The over/under pick is in parentheses with a U or O.

The next 3 columns represent the estimated probabilities. The first number is the probability that the team picked to win actually wins. The second is the probability that the team picked to beat the spread beats the spread. The third is the probability that we beat the over/under.

Printer friendly links and links with games sorted by value according to spread and over/ under are listed at the bottom of the page.


2018 Record
                    Straight-up                  vs. Spread                Over/Under 
Last Week         0-1-0      0.000              1-0-0     1.000           0-1-0      0.000
Season           570-202-0   0.739            371-389-11  0.488         409-348-14   0.540
                             Best Bet (Spread)         Best Bet (Over/Under)         
Last Week                     1-0-0     1.000             0-1-0      0.000            
Season                       11-5-2     0.647             7-10-1     0.441  

"Printer-friendly page"

"Points spread page (sorted by probability of beating the spread)"

"Over/Under page (sorted by probability of beating the over/under)"

To take a look at the underlying rankings click here.

To review previous weeks predictions: week 1, week 2 week 3 week 4 week 5 week 6 week 7 week 8 week 9 week 10 week 11 week 12 week 13 week 14 week 15 week 16 week 17

Note: The ratings are the result of a Dynamic Hierarchical Bayesian Linear Forecaster. The author has a Ph.D. in Statistics from Texas A&M. He specializes in Bayesian Forecasting. The forecasting method has been presented at four technical conferences, the 1997 and 1998 Conferences of Texas Statisticians, as an invited presentation at the 2001 Joint Statistical Meetings , and at a 2003 Houston INFORMS meeting. The powerpoint slides from the INFORMS talk are available here.