How the Playoff Committee uses Data to Place Teams

First off, let me say I hate that the committee has no firm, fixed, transparent criteria that one can easily judge and see how they got what they got. I would prefer a list of the criteria, even it was 100 items long, with a weighting that would require a computer to keep track of it all, but at least all would know and could figure out what is going on, but that’s not what we have.

The CFP committee uses a company called SportSource Analytics to provide them with data to compare and rank the teams. If you have ever looked up college football team stats on https://cfbstats.com/, you are using their data.

One could compare two teams, say Wisconsin and Florida just using the data at that website.

Their standard data sheet for Florida is at this link: https://cfbstats.com/2019/team/235/index.html
and here’s the one for Wisconsin: https://cfbstats.com/2019/team/796/index.html

You could look at the two of them side by side, and if you did so, you’d see that Wisconsin is better in more rows than Florida.

But that’s tedious just for two teams, and almost impossible for 25, so instead go straight to the SportSource Analytics Rankings Tool (at least the free one open to the public, I’m sure the committee’s version has more bells and whistles) and compare for yourself at https://sportsourceanalytics.com/rankingtool/

In the base ranking there are 13 attributes that can be weighted however you want. In the advanced setting there are 19.

Here is what their default advanced comparison looks like with all 19 sliders at the default 50%


Notice Wisconsin is #4 and Florida is #11 using this advanced base ranking.

But the committee does not just look at this, they play around with settings and look at many different things.

For instance, what is strength of schedule? it comes out WI #9, FL #36 (set all sliders to 0, put SoS at 100)

If you just hit reset to go back to the unadvanced base comparison (all 13 sliders at 50), it’s WI #5, FL #18.

You can do this for any one, or any number of parameters and see what the committee sees.

Say you want to look at just the 4 opponent adjusted rankings (Opp. Adj. Scoring Offense/ Defense & Total Offense/Total Defense) in their advance set, set all the other 15 sliders to 0, the result is WI #3, FL #5

You can play around all day long and Wisconsin beats Flotida more often than not on their ranking tool. That’s why, with the exact same 10-2 record, Wisconsin ends up ranked one spot ahead of Florida.

This is the paint by numbers comparison approach they use, we just never know what sure what numbers they value more than others and how consistent they are.