The Essential Guide to Your Membership


Your support lets me continue to do the research on more accurate methods for predicting football and March Madness.


The homepage for your membership contains the primary predictions for each week during football season. Also, I post updates at the top of the page.

To check out these updates and predictions, click here (although there’s a good chance you just came from that page).

During football season, I post predictions for the spread in each game for both college football and the NFL. These are ensemble predictions that combine predictions from many data sources, which include game and market data.

Later in the season, I post predictions for total points (after week 6 in college football, after week 7 in the NFL in 2016). From the total and spread, you can work out a predicted number of points scored for each team.

Beyond these game predictions, your membership also includes these three major items.

  • College football win totals. I release these computer predictions in early June, about when some sports book issues their win totals. You see these numbers before the public gets them in July.
  • Bowl predictions. These appear after championship week in college football.
  • March Madness bracket advice. The win probabilities for each team are available after the bracket gets announced on Selection Sunday. My full bracket advice appears on Wednesday, the day before the Round of 64 begins.

Your membership also entitles you to my book How to Win Your NCAA Tournament Pool. To get your copy, click on “Download:”


Let me know what you think.

Team rankings in football

In the drop down menu under Members, you’ll find links to the primary pages for college football and the NFL. These pages contain the primary ensemble rankings upon which my predictions are based.

These rankings pages also break down teams into offense and defense.

In college football, I use yards per play adjusted for strength of schedule to evaluate offense and defense. Since these quantities tend to persist from year to year, these rankings include data from last season as well as the current season.

For the NFL, I use yards per pass attempt adjusted for strength of schedule to evaluate offense and defense. My research shows that pass efficiency correlates with winning more than even yards per play.

To learn more about how passing dominates the NFL, check out my article on the insignificance of rushing, one of my favorites.

Pass efficiency doesn’t persist from year to year in the NFL, so I only use data from the current season. During the first three weeks of the season, the numbers on the NFL rankings page are raw yards per pass attempt. Starting week 4, this pass efficiency gets adjusted for strength of schedule by my algorithm.

The rankings pages for college and pro football also have links to team pages, which let you look at match ups in games.

How to instantly evaluate a game

The interactive visuals on the team pages give you a snap shot of the match ups in any game. To see this, first click on an opponent in the schedule (upper right corner).

Then to understand how the visuals work, check out this visual:


Better defenses appear further to the right. When an offense and the opposing defense appear next to each other, the unit further to the right is predicted to have an advantage.

The visual comes before the 2014 College Football Playoff championship game. Oregon was the favorite in the game, but the Ducks had a poor rush defense. The numbers suggested it wouldn’t hold up against Ohio State’s top ranked rush offense.

Ohio State’s Ezekiel Elliott ran for 246 yards on 36 carries (6.8 yards per carry) against Oregon. Despite 4 turnovers, Ohio State won 42-20.

It is the start of the fourth, and it is creepy how on point your predictions are.

commenter on Deadspin based on above prediction

For college football, the visuals allow you to look at match ups on offense and defense, as well as these units broken down into passing and rushing.

For the NFL, the primary visual for offense and defense uses yards per pass attempt. The other visuals show rushing by yards per carry and pass rush by sacks per pass attempt. For both of these quantities, I use a least squares ranking algorithm to adjust for schedule after week 3.

Why humans get fooled by small sample size


I posted the article I sent in the mail to members in August, 2016. You know it’s dangerous to judge based on small sample size, but this article digs into why humans have such an uneasy relationship with randomness.

To check out the article, click here.

Thanks again

If you have any questions, email me here.