Today, I’m sharing with you another tool that can be used in a similar manner.
If you are still fuzzy on what the Pomeroy Ratings are, click here for a piece I did for Odds Shark on how to use them when handicapping college basketball.
The Bart Torvik Ratings are similar to KenPom’s. They both use Adjusted Offensive Efficiency and Adjusted Defensive Efficiency and their ratings tend to align with each other. For instance, both advanced analytics sites ranked Houston, Alabama, and UCLA as their top 3 schools heading into March Madness, but the Pomeroy Ratings have UConn No. 4, and Tennessee No. 5, while Torvik’s ratings had Tennesse No. 4, and UConn No. 5. The difference lies in each mathematician’s unique formula.
What makes Bart Torvik’s resource unique, and especially helpful this time of year, is that users can splice the data using specific dates. Curious how a team did during a 5-game stretch in February when they didn’t have their starting PG? This tool can help.
Because March Madness success is often dependent on which teams are hot and have built momentum entering the Big Dance, I like using Bart Torvik to chop off data from the earlier months of the season and see what teams have done lately according to the advanced analytics. Doing so has some fascinating results which you can use to help cash bets.
A Recent Example of Success
Earlier this year in the NIT, 93% of bets were on the Rutgers (-245) moneyline when they played Hofstra at one major sportsbook. It’s easy to see why the bets were so lopsided as Rutgers was getting a lot of sympathy (deservedly so) in the press about how they were a good team who deserved to make the tournament. Furthermore, a shallow dive into the analytics revealed that Rutgers was ranked 35th in the nation and Hofstra was ranked 91st. On a neutral court, take the team that’s better according to the analytics with a chip on its shoulder? It’s not a bad angle and was backed up by some stats.
However, a deeper dive into the analytics using the technique I’ve described revealed that paying a -245 price for Rutgers was very risky and there was actually value on Hofstra +205.
If you zoomed in on the data set from January 8th onward, Hofstra was the better team with a ranking of 49, while Rutgers was just the 68th best team in the country. Furthermore, from February 1 onward, Rutgers was the 84th best team in the country, while Hofstra was ranked 42nd.
With this data in mind, +205 odds on a team with better analytics the last two months sounded a lot better than -245 on a team with better analytics over the course of a whole season.
Oh yeah, and Hofstra won the game, 88-86. How’s that for data?