Greetings all and I hope you are ready for some analysis. I’ve spent most of the afternoon putting all kinds of numbers together. I know, I know – you wish you could have been here. Maybe next time. By the way, and I honestly don’t remember if I’ve mentioned this before, but my friends, who are as close to used to me as possible, I think, and by being so, not surprised by random trivia or statistical statements – sometimes even related to the conversation, but that’s probably more random than intentional. If we are at someone’s home or a restaurant or bar and I do this, one of them will give me a knife – usually a butter knife because they aren’t complete idiots. This “geek” knife has become something between a running joke and a badge or honor. Truthfully, it’s a bit harder to earn one than it once was. Today’s analysis is not just a deeper dive in the second half of November, but an attempt to earn a geek knife or ten.
First, let’s get the GCR picks out of the way. A friend of mine has already reached out to me about next week’s Texas A&M (31) at Georgia (6) game – UGA 72.9%. I look at every game (unless it’s an FCS team vs a lower opponent) and will be happy to give the GCR pct to anyone who asks. I can even do future or non-existent games – for example, who would win if Air Force (30) played at Pittsburgh (39)? It’s Pittsburgh at 52.7%. If Pitt traveled to Colorado, the Falcons would win 62.8% of the time. If they met in a bowl, Air Force takes it 55.1%. Of course these numbers change with each game entered, but it’s kind of fun to play “what-if”. Last week was a good week for the predictions overall (not the upset picks which were 2-9 straight up and 5-6 ATS). Here is the weekly followed by the YTD (since I started tracking it).
| Decile | FBS | FCS | Total | Expected Wins |
| 50.0-59.9% | 8-8 (.500) | 17-12 (.586) | 25-20 (.556) | 24.75 |
| 60.0-69.9% | 10-2 (.833) | 9-4 (.643) | 19-6 (760) | 16.25 |
| 70.0-79.9% | 15-5 ).750) | 7-2(.778) | 22-7 (.759) | 21.75 |
| 80.0-89.9% | 5-0 (1.000) | 4-0 (1.000) | 9-0 (1.000) | 7.65 |
| 90.0%+ | 1-0 (1.000) | 1-1 (.500) | 2-1 (.667) | 2.85 |
| Total | 39-15 (.722) | 38-19 (.667) | 77-34 (.693) | 73.25 |
| Decile | FBS | FCS | Total | Expected Wins |
| 50.0-59.9% | 52-50 (.510) | 74-59 (.556) | 126-109 (.536) | 129.3 |
| 60.0-69.9% | 60-25 (706) | 62-35 (.639) | 122-60 (.670) | 118.3 |
| 70.0-79.9% | 51-14 (.785) | 43-20 (.683) | 95-34 (734) | 96.0 |
| 80.0-89.9% | 31-3 (.912) | 22-4 (.846) | 53-7 (.883) | 51.0 |
| 90.0%+ | 13-1 (.929) | 17-2 (.895) | 30-3 (.909) | 31.4 |
| Total | 207-93 (.690) | 218-120 (.645) | 425-213 (.666) | 425.9 |
In the 300 games since I started, the GCR picked the winner 69% of the time and over 81% when predicted 70% or more in FBS. The FCS has been a bit more disappointing and over the summer, I’ll revamp the prediction model for that group. Overall, I’m just 0.9 wins below expectation after 638 games. So the model works, but there’s always continuous improvement possible.
But, that’s all fine and good. What about the games themselves? Ok, you’re right, you’re right. We had a new group of teams clinch bowl eligibility and a group that already know when their season ends. Before we jump into the analysis, here are the newly bowl eligible teams:
| Team | Conference |
| Arkansas St (67/65) | Sun |
| BYU (52) | Ind |
| Florida St (47) | ACC |
| Ga Southern (50) | Sun |
| Iowa State (32) | Big 12 |
| Louisville (29) | ACC |
| Miami – OH (58) | MAC |
| Utah State (129/106) | B12 |
And the ones who will lose the extra practice time and extra money and player SWAG, etc.
| Team | Conference |
| Bowling Green (167/118) | MAC |
| Houston (148/111) | AAC |
| Kansas (128/105) | Big 12 |
| Middle Tenn State (155/114) | C-USA |
| Mississippi (87/82) | SEC |
| South Carolina (94/86) | SEC |
| Texas State (151/113) | Sun |
Again, all fine and good, but there are a finite (78) number of teams that will get a bid. Let’s talk about that. First, some sort of rules of the road: There are 40 bowls (one of them is the Championship game) so that’s how we get 78 teams. To be eligible a team has to have 6 or more wins and only one of those wins can be against an FCS school. The team has to have a .500 record or better (unless they end of 6-7 because of a championship game loss). This is important because Army (121/101) is 5-6 with games @Hawaii (63) and Navy (46). They must win both games to be bowl eligible. Remember, any team that plays a game at Hawaii has the option of a 13-game regular season. If there are more than 78 teams that meet the criteria, some will not be invited (these are usually, but not always from the MAC, Sun Belt, or Conference USA). If there are not enough teams eligible, 6-7 teams are eligible next and then 5-7 teams will be eligible in order of their Academic Progress Rates. Whew – glad we cleared that up. All 130 teams are either already bowl eligible (64 so far), eliminated unless there are not 78 eligible (25 so far), one win away from eligibility (14 right now), or one loss from disqualification (27). Here’s the conference by conference breakdown:
| Bowl Eligible | Eliminated | 5 Wins | 6 Losses | |
| AAC | 7 | 4 | 0 | 1 |
| ACC | 8 | 1 | 1 | 4 |
| Big 12 | 6 | 1 | 1 | 2 |
| Big 10 | 8 | 3 | 0 | 3 |
| C-USA | 6 | 4 | 2 | 2 |
| Ind | 3 | 2 | 0 | 1 |
| MAC | 4 | 2 | 2 | 4 |
| MWC | 7 | 2 | 0 | 3 |
| Pac-12 | 4 | 0 | 4 | 4 |
| SEC | 6 | 4 | 3 | 1 |
| Sun | 5 | 2 | 1 | 2 |
| Total | 64 | 25 | 14 | 27 |
I’ll do this list after this week as well to see where we are going into rivalry week.
But, one of the things I was really interested in this week is how we could predict what the final final Playoff teams will be. So, here’s what I did. I took the 11 teams that have a semi-legitimate chance. Then I put the games in the GCR predictor. If a team had options for opponents (in championship games, I looked at the average to win pct). Note: these percentages and even a winner could change based on what happens this week and next, but we’ll just go with current data. From all of this, I predicted the final records of them, and with some controversy, came up with the final slate. Let’s take them one by one before showing the table. LSU (1) is 10-0 with games against Arkansas (173/121) and Texas A&M (31) before meeting UGA (6) in Atlanta for the SEC Championship. They should end up 13-0 with the top seed. Their opponent for the SEC, Georgia, will beat Texas A&M (what a nasty last two games they have) and waiting-for-this-season-to-finally-be-over Georgia Tech (150/112) before losing in the championship for the 2nd straight season. At 11-2, they will be in a New Year’s Six bowl. The other SEC team Alabama (3) is a special GCR situation, meaning that there is no adjustment for key player losses. The formula assumes all games are being played “now” and such things are not relevant to the ranking. That said, I don’t think Alabama gets in the playoff despite the 11-1 finish. First, they are playing not only an FCS school at the wrong time for optics, but they are playing a god-awful FCS school in Western Carolina (152/122 in the FCS). Even if they win 100-0 it won’t help them with a weak out of conference and cross-division schedule, compared to others. The game at Auburn (9) will be close, but it will be too little to late. They will get a New Year’s Six bowl, though.
The Big 10, which has been the strongest overall conference this year, has Ohio State (4) as the current leader. They play a pretty tough last two games – Penn State (8) and then at Michigan (26). If they win both of those they get the winner of the Minnesota (15)/Wisconsin (10) game coming up. If they lose to Penn State, the Nittany Lions will win the East. If that happens, both PSU (assuming they beat the team from the West) and Ohio State would be in – Penn State number 2 and Ohio State number 4). If Ohio State wins out (and they should) they will be the number 2 seed at 13-0. Speaking of Penn State after the Ohio State game they get Rutgers (79/75) to end the season. If they lose to Ohio State they are out and probably in the Outback Bowl. I think they should book tickets to Tampa since they will end 10-2. Like the SEC, the Big 10 has 3 contenders left. Minnesota (15) only gets in if they win the Big 10. To do that they have to beat Northwestern (137/107) on the road and then Wisconsin (10) followed by the winner of the Ohio State/Penn State game. They make the Big 10 Championship, but the final bell tolls and they end up 11-2 and in a New Year’s Six bowl.
The Big 12 is still down to 2 teams who played an instant classic last night. Baylor (5) despite the FG loss owns their own destiny. If they take out Texas (22) and Kansas (128/105) they have an opportunity to take on the Sooners (20) again for the Big 12 title. If they win that game, they are 12-1 and in. In fact, if the rematch is between two 1-loss teams, I think the Big 12 champion is in. It will be Baylor, though, as the #4 seed. Oklahoma will beat TCU (144/109), but face a tough challenge that the GCR thinks they will lose at Oklahoma State (16). If that happens and they then beat Baylor, the Pac 12 will sneak in at the 4th spot. But more on that in the next paragraph. Oklahoma ends up 10-3 and outside of the New Year’s Six.
Since we were just talking about the Pac-12, let’s move there. Utah (11) has an easy path to the Conference Championship game with a game at Arizona (91/84) followed by Colorado (92/85) at home. If Southern Cal beats UCLA (and they should) and Utah slips up, the Pac-12 will be left out of the Playoffs. More on that on the Oregon analysis. Utah should face and beat Oregon to finish 12-1, but they have the 74th most difficult schedule played and the 134th future schedule. Even if they do win out, they are looking at a New Year’s Six bowl. Oregon (14) on the other hand could kick out Baylor if they win out and if they get some help from Utah. If Utah loses, and Oregon faces Southern Cal (24), a 4-loss team, in the championship, the committee won’t care about the 12-1 record. But, if they beat Utah soundly, they have a chance at the #4 spot. It’s a long shot since the GCR says Utah wins on a neutral field. Oregon ends up 11-2 and gets the 5th New Year’s Six spot. By the way, the 6th spot is always given to the highest ranked Group of 5 school. That should be Cincinnati (12).
That leaves the final (#3) spot for the playoff. Clemson has run away from the rest of the ACC. Already 11-0, there are only 2 more contests left. First at South Carolina (94/86) then one of a trio of Coastal contenders – Virginia (17), Pittsburgh (39), and Virginia Tech (42). All 3 are tied with 2 losses in the conference and 7-3 overall records. If Clemson loses any of these games, they are out and the Pac 12 gets back in. They won’t, however, and will get the 3rd seed at 13-0.
To sum up: LSU (1), Ohio State (2), Clemson (3), Baylor (4). The New Year’s Six slate, in some pairing, will be Oregon, Alabama, Minnesota, Utah, Georgia, and Cincinnati. Penn State and Oklahoma will be at the next tier bowls. Here’s the chart with all of the analysis.
That’s it for the Explanation this week. Tune in Tuesday for G’s Exploration and see what may pop up. Send questions, comments, concerns, game pics, etc. And, as always, please continue to share so that we can keep growing.
Thanks, G
G,
What does your model do if you parlay teams? Is it a simple average of the win percentage? For example, let’s say I want to know the percentage chance for 4 teams (let’s use LSU, Clemson, Utah, Oklahoma for this one) all to win this weekend. Would it be the average of their percentage to win individually? Which would be 87.9% in this example….Or is there some other factor you’d need to add in?
So, Jason, I thought about this a lot. I don’t bet on college football, because being a father of a really good teenaged girl, I don’t expect people of a certain age to respond to things the way I’d like very often. That said, my first thought was purely statistical. Let’s say an event occurs 75% of the time (a known number) and you wanted to know the likelihood of the event occurring 4 consecutive times. It’s .75^4 or .316. In that known consecutive event you would expect the same outcome less than a third of the time. But a) football is not a known event. My current calculation shows Georgia beating South Carolina 92% of the time. Would the Gamecocks upset them again? Very, very unlikely. But it does happen but Georgia Tech beat Miami which was a 90%+ game. The point is the number I provide is not a known but rather a calculated expectation of the future based on past events and b) football games (in a parley sense) are not consecutive and don’t factor the same way. Let’s say the 4 games you chose all had a 75% win probability. The data through 300 games shows that the GCR picks the winner for 70-79% games around 75% of the time. So if you picked a 4-parley, the “average” would be 3 wins. A 6-parley would be 4,5 wins. It’s still very much a gamble. The problem is, because it’s college football, it’s impossible to know which of the teams will get that loss. Does that make sense or help at all?
“Their opponent for the SEC, Georgia, will beat Texas A&M (what a nasty last two games they have) and waiting-for-this-season-to-finally-be-over Georgia Tech (150/112) before losing in the championship for the 3rd straight season.”
Hey now, Georgia won the SEC Championship against Auburn in 2017, this MIGHT be the second straight loss in the championship.
Hi Alex. You get a safety for catching an error. You are right and I should have verified instead of relying on my memory. That game was a 28-7 Bulldog Victory. I will correct the blog tonight.