Salutations and Happy New Year!!!! We are in the final push for college football this season with just 11 games to go (9 more regular bowls, the FCS Championship game, and the FBS Championship game). The last two weeks have been a flurry of activity not just with the holiday season, but also with games nearly every day. I mentioned just before the bowl season started that, prior to bowl season, some teams (e.g., Clemson, Auburn, Alabama, et al) are higher ranked than logical while others (e.g., Oklahoma et al) are under ranked. I did some review into the code and found the issue. While it’s fixed for 2020, pending the analysis I’m hoping to tease you with, I did not incorporate the changes into 2019. Switching formulas, even correcting errors, that have existed all season would not only be unfair, it would undermine the integrity of the ranking system. I, like everyone else who reads this blog, have my favorite teams to love or hate, but I strive to make this blog based in fact without my personal biases. If the GCR says my favorite team will lose, or my most hated teams will win, so be it and I’ll report it that way even explaining why I think the formulas made those calls. Again, it goes to the integrity of the tool. To be completely honest, I don’t always pull for the GCR picked team, even though a high percentage of correct picks makes me happy. Now here’s the teaser: I looked back at the first 1510 games played this year (there will be 1555 this season) and reviewed the current GCR pick (they change each week: remember, the GCR thinks all the games occur at the same time) in both the 2019 formula and the corrected 2020 formula, only looking at correct/incorrect picks. If the corrected formula (which put Ohio State, LSU, Clemson, and Oklahoma as 1-4), is more accurate, then it should also be more predictive. This hindsight test should a) demonstrate how good the predictor is with (nearly) complete season information and b) demonstrate the corrections improve (keep them) or reduce (retool) the prediction accuracy. After today’s powerhouse games, there are 2 tomorrow, then just 3 more over the next 4 days. During the 5 days from January 7 through the FCS Championship on the 11th, I’ll talk through the results of this test.
…..
I had already decided to write that paragraph today before yesterday’s games. It was time, but they gods of college football put a little irony in the mix as I had my worst prediction day of the bowl season. For those statistically minded folks out there, I’m thinking it was an application of the Central Tendency Theorem. After starting 20-5 and feeling more smug than I should have, Kentucky with last 15 second heroics, Navy’s determination, and Utah’s utter failure to appear resulted in a 2-3 mark and 22-8 overall. Still at .733 so not complaining, but need some better results today.
…..
There are 4 blockbuster games today which is appropriate for a new year and what most are calling a new decade (that’s really next year, but who’s counting). All 8 teams are ranked in the AP top 25 and the top 28 in the GCR including 5 in the top 10. Before we get into this a few things. One: the rankings for these teams are as of today – all previous games have been included which impact current teams either through SOS or through attrition/significant jumps. Two: these are 2019 numbers with the calculation error. Here’s a couple of examples. Prior to the bowl games, Auburn (9-3) was ranked 4th. Oklahoma (12-1) was ranked 20th. After the corrections, Auburn, whose 3 losses came against both teams in the SEC Championship Game and Florida (11-2), came in at #10 – appropriate as the best 3 loss team this season. Oklahoma came in at 4th – appropriate for the best 1 loss team. Three: since there are no Group of 5 teams playing today, I’ll only update the Power 5 numbers. The average OffEff is 114.2, Def Eff is 126.1. The average rank is 48.2. We will use those numbers for comparison.
…..
Before we get into these great games, I want to do a little Voice of the Customer. I am such a geek, that even if I didn’t have this blog – in fact, the 6 seasons I had the GCR before I started the blog – I would do different analysis and cuts of the data just for fun, synthesize it, and tell my friends during what would more than likely be an excruciating 3-4 hours for them. That may be a little exaggerated, but I doubt it. Before the blog, I would send an email (until last season, I didn’t even have a way to include the post-season – long story for another blog) with whatever analysis piqued my interest the most. This post-season, I’ve tried to look at the pertinent information (again, that seems interesting to me) to help explain why team A should beat team B. Here is my ask: I need feedback. Not just for this season or the bowls, but things that will help the blog – things you would like to see. It can be about anything from the data to the presentation to the style to the frequency – you are my customers and I need to hear your feedback. This is not a “G- you’re great” session – especially since I know there are many improvements needed. Tell me the things you’d want to keep as is, certainly, but I can improve only if you hit me between the eyes with feedback (we call that interocular feedback in my profession). Any help you can give me is not only appreciated, but also very welcome.
…..
Now to the main event. Every year the Big 10 and the SEC play what are usually the 2 biggest non-new year’s six bowls in the Citrus and the Outback. They are almost always two games college football fans want to watch because of the matchups. The issue I have with them is they are both played at 1:00 (all times Eastern). If you have multiple TVs next to each other, like one of my nephews, then you can watch both at home. Or you can find a bar, but otherwise there’s a lot of bouncing back and forth hoping you don’t miss a big play. Today’s games are no exception and feature two very intriguing contests. First, let’s talk about the Citrus Bowl played in Orlando. Michigan (18, 9-3, 3rd in the Big 10 East) takes on the Crimson Tide of Alabama (5, 10-2, 2nd in the SEC West). The Wolverines lived up to their mascot going 3-3 against top 25 teams. Overall they had the 5th most difficult schedule facing 8 good (6+win) teams and 5 great (10+win) teams. Overall their opponents were 90-63. Alabama is an example of the formula issue and you can see that in these numbers. They were 0-2 against top 25 teams (losing to LSU and Auburn), only faced 6 good and 1 great team, and they’ve faced opponents with a losing record – 71-77. In the new model (in all cases in this blog, these are pre-bowl rankings), Michigan is 13th and Bama 9th – again, seems closer to accurate. I’ll use both numbers as comparison today, just for fun. When Michigan has the ball, they will take a significantly above average offense (OffEFf 144) against the #25 defense (145). When the Elephants have the ball it will be strength on strength: Alabama’s #3 offense (172) vs Michigan’s #20 (152) defense. This will be power football where every yard will be contested. Vegas likes the Tide -7.5 and I agree. Alabama 68.1%.
…..
Around 85 miles away, in Tampa’s Outback Bowl, Minnesota (14, 10-2, 2nd in the Big 10 West) will battle Auburn (3, 9-3, 3rd in the SEC West). In the updated formula, Minnesota is #14 and Auburn is #10. Both teams have one win against the top 25, the Golden Gophers had 3 chances and the War Tiger Eagles had 4. Minnesota played 6 good and 3 great teams while facing oppenents with a losing record (73-77). Auburn had the second most difficult schedule facing a 95-57 cumulative record so far, 9 good teams and 5 great ones. We see the difference in these teams in the efficiency rating, really the DefEff. Offensively they are almost identical – Gophers scoring 137, good for 21st in the country, and the Tigers rating 138 at 20th. Auburn has the advantage on defense with a 13th best 178 while Minnesota is unranked and a little below average at 119. That 119 against the schedule they played leads to an Auburn victory – hard fought at 54.7%. Vegas agrees Auburn -7.0.
…..
After a short break, we have our final New Year’s Six Bowls. First the Rose Bowl in good old Pasadena kicking off at 5:00 featuring a solid match up of Wisconsin (10, 10-3, Big 10 Runner up) and Oregon (28, 11-2, Pac-12 Champions). Interestingly, in the new improved formula Oregon is 11th and Wisconsin is 16th. Both of these teams had fantastic records against the top 25 (keep in mind the top 25 has a combined .762 winning percentage) with Wisconsin going 4-2 and the Ducks going 3-1. Both teams played 9 good teams with the Badgers doubling up the great opponents, 4 to 2. Both squads faced strong competition: Oregon’s opponents were 93-72 and Wisconsin’s were at 11th best 99-67. Both schools have fantastic defense that are nearly identical. Wisconsin has a DefEff of 185 (9th best) and Oregon 184 (11th best). Both offenses are above average and in the top 25: Ducks (135, 24th), Badgers (149, 9th). These teams are very close and it should be a close and fun game to watch. Vegas thinks so too with Wisconsin -3.0. I agree. Badgers 56.8%
…..
The nightcap and final NY6 game is the Sugar Bowl out of New Orleans, home of this year’s FBS championship game. Kick off is at 8:45 when Baylor (6, 11-2, Big 12 Runner up) faces Georgia (4, 11-2, SEC Runner up). In the new formula UGA is 8th and the Bears are 12th – and the new GCR picks UGA to win (although it’s not exactly comparable because it has 45 fewer games included, but the SEC has fared well in those 45 games and the Big 12 has faltered, so it’s probably close). Baylor was 2-0 against top 25 teams (Oklahoma is 34th), and the Bulldogs were 3-1. UGA played 9 good teams and 3 great teams against 96-67 competition. The Bears had 6 good and 2 great foes and played a below average 82-82 set of opponents. On offense, Baylor has a slightly above average team (123) vs the 2nd best defense in the country at 229. When the Bulldogs have the ball, a slightly better offense (131) will face an above average (19th in the country at 152) defense. Everything says UGA will win. Vegas thinks so at Georgia -5.0. But, in the upset of the day, Baylor wins a squeaker 50.8%.
…..
That’s it. I hope all of you are recovering from the night of revelry. Thank you again for helping me launch this blog. Please don’t forget to provide feedback – I need it. Please share with others and have a wonderful New Year. Here’s to a safe and peaceful 2020, G
