Using “Conference-Only” Stats to Get Accurate Reads

Using “Conference-Only” Stats to Get Accurate Reads

As a stathead, probably the most disappointing element of the “debate” this season about who belonged in college football’s Final Four was the absence of discussion about statistics. It’s as if everyone decided looking at stats was just too complicated, so let’s just follow some simple rules:

Any undefeated Power 5 team gets in (Florida State)

The SEC champ gets in (Alabama)

Any Power 5 champ that wins a lot of blowouts gets in

If there are too many Power 5 champs that win blowouts, talk in circles and invent tiebreakers on the fly

Oregon, Ohio State, TCU, and Baylor were all either champs or co-champs of a Power 5 conference. Oregon rated best in the national rankings, which was their tie-breaker. That left the other three fighting for one spot. Ohio State got the nod. We’ll probably never know for sure what tie-breakers were used behind closed doors to put the Buckeyes at #4. It sure didn’t hurt that they won a blowout as an underdog the last time they took the field.

I wish I could tell you that I had the solution to this. There just isn’t enough interconnectivity in the schedules to clear out the fog. But, a quick and simple way for an analyst (handicapper/bettor) to get a decent postseason read on teams is to do the following:

Focus only on CONFERENCE stats

Make a good faith effort to evaluate how the conferences rank against each other

The first step takes out the garbage games against non-conference cupcakes, while giving you about a nine-game sample size to draw from. Not perfect…but a GREAT way to evaluate teams WITHIN a conference. This is particularly true in the Big 12 where everyone plays each other once. But, you’ll see a fairly accurate picture even if the schedules don’t line up perfectly.

Then, once you have teams rated within their conference, take the steps necessary to adjust for conference strength. Some analytics sites rank the conferences for you. You can pick one of those, do a composite of them, or do your own evaluation.

Here’s a quick example. TCU and Baylor both played nine games in the Big 12 (it’s a 10-team league remember), here are their yards-per-play stats with Turnover Differential thrown in for added context

Conference Only Yards-per-Play

TCU: 6.9 on offense, 5.4 on defense (+18 turnover differential)

Baylor: 6.2 on offense, 5.7 on defense (+12 turnover differential)

Against the same eight teams…and then with one game against each other…TCU won YPP differential +1.5 to 0.5, and did a much better job forcing turnovers (36 takeaways compared to 24 for Baylor). Baylor’s 70-6 scrimmage against Northwestern State isn’t counted because it wasn’t a conference game, nor was anything from TCU’s pre-conference slate. The numbers show what happened when two good teams played extremely similar schedules (though the deck was loaded a bit for Baylor because the head-to-head meeting was in Waco).

TCU grades out very well here. It would have been cool if this was in the media discussion about how to separate the two teams. Instead we had, literally HOURS of:

Baylor won the head-to-head meeting, that qualifies them…or

TCU lost by less than the value of home field advantage, that qualifies them

Here’s how some of the other teams in the Final Four discussion performed in their conferences…

Conference Only Yards-per-Play

[box] Alabama: 6.4 on offense, 4.7 on defense (+3 TO’s) in the brutal SEC

Oregon: 7.1 on offense, 5.4 on defense (+12 TO’s) in the decent Pac 12

Ohio State: 7.2 on offense, 4.7 on defense (+10 TO’s) in the soft Big 10

Florida State: 6.5 on offense, 5.5 on defense (-2 TO’s) in the soft ACC

Ole Miss: 5.9 on offense, 5.1 on defense (+4 TO’s) in the brutal SEC

Mississippi State 6.4 on offense, 5.8 on defense (-1 TO’s) in the brutal SEC [/box]

Alabama has great numbers with such a big YPP differential in a killer conference. Oregon obviously hit a home run as well. Ohio State did the same, but in what’s widely perceived to be an easier conference. At least now, observers can really pin down the issues.

How much should Ohio State be penalized for playing in the Big 10?

If you’re trying to get “the best four teams,” would FSU fall behind the SEC or Big 12?

Does Ole Miss have a good case for inclusion with those SEC stats and a win over ‘Bama?

Is TCU’s stat line more impressive than either Florida State or Ole Miss in context?

There was time for this, and a lot of other stat-related material in the non-stop studio jabber. We really had to hear Kirk Herbstreit say that Ohio State “had done enough to get in” 50 times? Stats aren’t poison!

They’re not a cure either. At least by themselves. Complicated issue. But, I do believe you’ll get a clearer picture of how college football teams measured up this year by evaluating “conference only” stats as your postseason baseline. Then, adjust for how respected analytics efforts had the conferences rated.

You have plenty of time to do that for ALL teams in ALL conferences before the bulk of the major bowls get started. Utah of the Pac 12 is the only Power 5 team to have taken the field yet. You can find “conference only” stats at www.cfbstats.com, a fantastic source for college football statistics. You have to go team by team (or category by category) to find the numbers I outlined above. But, there are also many other stats that you may prefer to mine. Do the work, and see if it helps you make better decisions once the major conferences are ready to take the field this postseason.

See you again next week to talk some more about the Final Four, which is set for New Year’s Day and evening.

Jeff Fogle is a freelance writer living in Austin, Texas. He writes about college and pro football, college and pro basketball, and MLB on his blog StatIntelligence. You can follow Jeff on Twitter @JeffFogle.