Welcome back after a long layoff, oh great and wonderful college sports fans! We’ve spent the last six weeks or so coding the 2026-27 football season in preparation for the preseason and the more than 1600 regular season games. Along the way, we’ve noticed a few oddball and/or interesting bits of data that we will share in the pre-preseason GCR.
…..
Today’s episode (#643) is introducing not only our proposed changes to the GCR for this season, but also to lay out the pre-pre and preseason schedule that will take us up to episode #675, aka the first day of the season (August 29th).
…..
The GCR will look similar in reporting to previous years for football with one exception. When we had our first full season of Men’s and Women’s basketball last season, we had a 2-step system for predicting games. We only used the models in the postseason, but they were fairly effective. The two pieces were the RPI Factor and the Score metrics. If both agreed on the winner, the GCR would predict (with ~70% accuracy in the conference championships and postseason tournaments), but if the metrics disagreed, we simply did not predict that game. There were no point spreads. In the past, we have used one prediction metric that was based on the Score metric for football. We will continue to predict both winners and who covers, but only if both Score and RPI (added now for football) agree. That means instead of predicting every game, we expect about 10-15% of the games to be listed as “NP” or No Prediction. This should correct for “loser” picks that are really just toss up games – mostly in the FCS world. At some point in the future, we may drop the “cover” prediction. To be clear, we do not bet on college (or pro) sports. This blog is not designed to be a betting site, although there are some who use it that way at their own risk. The “cover” metric was created as a comparison to how well Vegas picks winners which was used to validate the metric rather than a suggestion to place bets. Let us know what you think in the comments.
…..
The second change in the GCR is also based on the basketball postseason. In previous years, for football, we’ve listed all the games and predictions on the day of the first game of the week. In August and September that is either Thursday or Friday most seasons. In October and November, we shift to Tuesdays or Wednesdays. December and January are all over the place due to bowl games. The problem with that is each game that is played impacts the predictions for all future games – those early games could have an impact on games later in the week, especially those in conference. For the postseason basketball schedule, we decided to post daily. We were able to keep up with games played, keep things balanced, and predict with decent accuracy even through March Madness. We know our readers enjoyed the frequency because our viewership exploded. We hear you and are responding. This season, all games (football and basketball) will be predicted on the day of play, once we start the predictions. Football predictions will begin when each team has played at least one game (we may wait for two games for increased accuracy) and basketball when each team has four games played. That means in November and December, we may have three posts on some days (unless we get feedback that our readers would like to see a combined football, Men’s basketball, Women’s basketball prediction post – either way works for us). The Football Top 266 will be posted on the day following the last game of the week (usually on Sunday) while the Men’s Top 365 and Women’s Top 363 will be posted on a given schedule (TBD, but probably not Sunday).
…..
The final change only impacts the annual preseason Strength of Schedule (SOS) conference ranking. In previous years, our formula was based on two primary factors. First, each opponent’s win total in the previous year and second, an adjustment based on location (more points for an away game, some added value for a neutral site game). The adjustment still exists although the factors themselves have been recalculated to be on scale. Instead of wins by opponents, the new preseason SOS is based on last year’s GCR Score (which is used for ranking purposes). Non-division-I teams are treated similarly to the regular season including a slight adjustment to the base for the level of success obtained. Promoted teams are adjusted, if needed, whereas teams that have never played football at any level before (see Chicago St this season) are assigned a zero value. This change provides a more accurate description of what a team is facing, assuming that each team repeats last year’s performance (more on that on July 8th).
…..
Let us know what you think about these changes in the comments. We are equally, if not more eager, to hear your suggestions as how else we can make the GCR better for you.
…..
Here is the tentative schedule for the rest of May, June, July and August.
| Date | Topic |
| May 31 | Conference Realignment for 2026 |
| June 13 | Rule Changes for 2026 |
| June 26 | Conference Specific Scheduling Rules |
| July 8 | Intro to Strength of Schedule Preseason Ranking and Explanation of Calculation |
| July 9 | 25th Most Difficult Conference |
| July 11 | 24th Most Difficult Conference |
| July 13 | 23rd Most Difficult Conference |
| July 15 | 22nd Most Difficult Conference |
| July 16 | 21st Most Difficult Conference |
| July 18 | 20th Most Difficult Conference |
| July 20 | 19th Most Difficult Conference |
| July 22 | 18th Most Difficult Conference |
| July 23 | 17th Most Difficult Conference |
| July 25 | 16th Most Difficult Conference |
| July 27 | 15th Most Difficult Conference |
| July 29 | 14th Most Difficult Conference |
| July 30 | 13th Most Difficult Conference |
| August 1 | 12th Most Difficult Conference |
| August 3 | 11th Most Difficult Conference |
| August 5 | 10th Most Difficult Conference |
| August 6 | 9th Most Difficult Conference |
| August 8 | 8th Most Difficult Conference |
| August 10 | 7th Most Difficult Conference |
| August 12 | 6th Most Difficult Conference |
| August 13 | 5th Most Difficult Conference |
| August 15 | 4th Most Difficult Conference |
| August 17 | 3rd Most Difficult Conference |
| August 20 | 2nd Most Difficult Conference |
| August 24 | Most Difficult Conference |
| August 27 | FCS Day One G’s Expectation |
| August 28 | FCS Day Two G’s Expectation |
| August 29 | FBS Day One G’s Expectation, FCS Day Three Expectation |
…..
That’s it for today. We hope that you are as excited about the upcoming season as we are. Keep those comments and suggestions coming. It’s how we get better. Thank you for reading. It’s why we do this. Special thanks for all of the sharing with others over the last few months. It’s how we grow. See you in 13 days, JoJo and G.
Excited for a new season, Robert and JJ!
I, too, am not a better, but like you, I am interested in how the Vegas “experts” predictions compare with your computer’s predictions (in addition to actual results). We have to realize that the Vegas line is not a true winner prediction because the line is shifted to account for the number of betters on each side, and the bettors may not always be the most expert predictors (although large population prediction accuracy is a thing). For examples, there is school program tradition bias and fan-based population bias.
Have you ever done a final top 25 comparison between the GCR and the official college rankers? Of course, neither one is absolutely “correct”, but college fans could latch onto their favorite, based on how their school ranked in each of the Top-25 lists, lol. My off-the-cuff squinty-eyed comparison is that there is a lot of commonality.
Finally, how have you resisted calling the GCR an AI-produced database? All the big corporations say in advertising that their smartness is because of AI, when we both know most of it is just old-fashioned computer programing (but well-honed by human brains). Saying AI might get you 2 or 3 more followers.
We need to get together for lunch again.
Carl