We’re getting close to filling out brackets and although winning the pool (and the cash) is everyone’s goal at the beginning, there is another way to gain bragging rights in the office during the madness. Everyone wants to be the guy who picked the big upset. Even if your championship team gets knocked out in the Sweet 16, you can still gloat about picking that 14-seed that ruined everyone else’s brackets.
So what can we look for when searching for these elusive upsets? We recently added some new filters to college basketball within Bet Labs and I used them to find a certain quality you should be looking at before filling out your bracket.
I started out with these three filters in a spread system:
I then opened up the new pace filter within Bet Labs and saw the following data:
Each past data point is plotted and if it is profitable it is listed in green. As you can see, there is a lot more green on the left-hand side where the pace is slower. So if we set the maximum pace to 70.5, we end up with a 86-70 record, which is a 55.1% win rate. Now that’s a good result, but this is a spread system. Covering the spread is fine and dandy, but it doesn’t do you any good when you need to pick the winner of the game for your bracket.
So I created the exact same system, only this time for a moneyline system:
These teams have won less than a third of the time, but the payouts from the upsets have given this system a +9.4% ROI since the 2010 tournament. You’ll notice that the “Seed” filter includes teams seeded 9th through 16th. As you probably already know, no 16-seed has ever beaten a 1-seed. We can quickly edit the filter to remove 16-seeds from the system and get these results:
By eliminating the 16-seeds, the ROI jumps to 20.5% while gaining another twelve units of profit over the same period of time. Another great feature of Bet Labs is the Past Matches button which allows us to go through the system on a game-by-game basis.
Recent upsets include Georgia State last year and Mercer taking down Duke in 2014. But by inspecting the data, you’ll notice a few games that look a little out of place. Tennessee as a -425 favorite as an 11-seed? That’s because that is a play-in game (also called the First 4). Tennessee technically fits all of the filters we specified, but those games don’t really fit with the idea that we were wanting to test.
It’s not a big deal, we can simply look at the Favorite/Dog filter and make a quick change:
By eliminating the favorites that were in the system, we end up with a system that is +35 units and has an ROI over 30%. If you would like to create your own NCAA Tournament systems, you can do so with a 6-day trial by clicking here. All new signups will automatically gain access to 10 professional college basketball systems that will give you picks from now through the end of March Madness.