Now that football is now officially over and done with, at least until September (IMO…I really don’t care about the draft or anything else that happens between the Super Bowl and the first Thursday Night Football game), it’s time to move on to NCAA Men’s College Basketball (MBB). As mentioned in previous blog posts, I’ve been posting win probabilities for Top 25 teams on my Twitter account for the last few weeks. In the last couple of days I’ve shown win probabilities for all the games played on any given day. I may continue doing that dependent on just how many Division I games played; I’ve seen days where nearly 50 games are played in one day (usually Saturdays).
This particular blog post is one that I hope will be a regular weekly series throughout the rest of the regular season. Basically I’ll be comparing a ranking of the Top 15 teams that I’ve compiled using a Bayesian approach with 10,000 simulations of the remaining games in the regular season. As of this writing, there are 1479 games left to be played.
Now, obviously, you can’t make an Apples to Apples comparison of the AP Poll versus a mathematically generated ranking. The rankings for the AP Poll come from various sportswriters and broadcasters from across the United States. Each voter provides their own ranking (typically based on the previous week’s results as well as the season so far). I’m not sure that any of those polled actually make a decision based on simulations or mathematics, but anything’s possible. The AP Poll, though, is what we have so that’s how I’m making my comparison…for now.
Here is my first published results. The last column shows the percent chance that the team in the Simulated Final Rank column will have of being the best team in the nation on the morning of March 11th.
| Rank | AP Poll | Simulated Final Rank |
Chance of Finishing #1 |
|---|---|---|---|
| 1 | Tennessee | Gonzaga | 35% |
| 2 | Duke | Nevada | 32% |
| 3 | Virginia | Houston | 13% |
| 4 | Gonzaga | Duke | 9% |
| 5 | Kentucky | Tennesee | 5% |
| 6 | Nevada | Virginia | 4% |
| 7 | Michigan | Michigan | |
| 8 | North Carolina | Virginia Tech | |
| 9 | Michigan St. | Marquette | |
| 10 | Marquette | Michigan St. | |
| 11 | Virginia Tech | Cincinnati | |
| 12 | Houston | Buffalo | |
| 13 | Kansas | Kentucky | |
| 14 | Villanova | LSU | |
| 15 | Purdue | North Carolina |
There are definitely differences in the rankings here. The Top three teams in my rank are Gonzaga, Nevada, and Houston. The one thing to note for all of these teams is that none of them are necessarily in “top tier” conferences. The three of those teams finishing out the rest of the season undefeated seems much more feasible than say Duke, Tennessee, or Virginia. After all Duke will be playing Virginia this Saturday afternoon, has games on the road against Louisville and Virginia Tech in February, as well as play North Carolina…twice! Duke has a rough road ahead of them as does Virginia, the same conference, with games against the same opponents as Duke. As you can see, though, their chances of being #1 are a little smaller than Duke’s.
Gonzaga doesn’t play against any other AP ranked opponents for the rest of the season. Neither does Nevada. Houston plays Cincinnati twice; Cincy is currently ranked #25 on the AP Poll, but my ranking seems to indicate that they are much better than that ranking at #11.
I’m a little surprised that Tennessee doesn’t have a higher chance of being #1. Granted they still have two games to play against Kentucky, but the only other currently AP ranked opponent they have is at LSU.
In any case, this is my first published attempt at this. Just like everything else I’ve built over the last several months, I’ll make improvements and see how things shake out with those improvements!
Feel free to reach out with comments or questions. For sure, if you feel your deserves a higher ranking, let me know your thoughts!