good stuff Clint and Cody
Kellner: Louisville, Houston, Oklahoma, Notre Dame
Cody Kellner has been doing football ratings since 2002 and has worked on various media relations projects for the Cotton Bowl Media Relations Team and the College Football Playoff.
Kellner is a PhD student in Applied Technology and Performance Improvement at the University of North Texas. Kellner bases his measurement on the average of each opponent's individual rating that each team has played up to that point during the current season.
His results are currently being published in the Waco Tribune Herald and the USA Today.
Wins, losses, and margin of victory are considered. Each team is given a game rating that is reflective of the result of the game (win or lose) and the strength of the opponent for that game. More points are given for defeating a higher rated then than a lower rated team. More points a deducted for a loss to a lower rated opponent than a higher rated opponent. Results from the previous season are considered, but diminish as a factor as the season progresses.
Not much of a difference, I have mostly been using the same approach since 2007. However, starting this season, I will be giving an additional penalty to a team if they have suffered a loss at any point during the season.
Haha, I don't think so, but those were great times!
Oklahoma has lost two games, but they were against two teams (Ohio State and Houston) that can very well make it to the playoff. Situations like this increase the importance for ratings systems such as what I have as a 2 loss Oklahoma team should still be ranked higher than a team with a better record playing an easier schedule.
Based on the success Houston had last season, I feel more confident they can continue sustaining their success during the season in comparison to Louisville. However, Louisville pretty much controls their own destiny being in a Power 5 Conference.
If Louisville can go undefeated, they will make the playoff. However, based on being in a Group of 5 Conference, Houston must go undefeated and will also need help in order to make the playoff (such as Oklahoma winning the Big 12 and several Power 5 Conference Champions having multiple losses).
My top 4 teams that year were the same four teams that actually made it to the playoff. My system had eventual national champion Ohio State at No. 2 going into the playoff. Based on that, Ohio State was an easy choice to be in the playoff as they were the only playoff team that year that didn’t have an FCS team on their schedule, so their schedule was naturally tougher.
Outside of the top four, I had TCU at No. 5 and Baylor at No. 6. Based on the lack of wins against quality competition the Big 12 had in non-conference play that year, it was difficult to justify having TCU or Baylor in the playoff that season.
I believe the SEC is the best overall conference right now, but the difference between the SEC and other conferences has been shrinking over the last couple of years. The Big Ten is continuing to make progress and I see them having success such as the SEC has recently enjoyed.
Normally, my results have been in line with the BCS and CFP over the years. Overall, I would say Oklahoma, Texas, Alabama, LSU, and USC have been among the best in my poll during this time. I will say there was a few years there that my model always seemed to underrate some really good Ohio State teams back when the rest of the Big Ten was struggling and hurting Ohio State's strength of schedule.
Notre Dame's loss to Duke will hurt Texas' resume as they defeated a Notre Dame team that Duke defeated also, so the quality of win isn't as great as what it could have been. However, Notre Dame always has a tough schedule, so if they can bounce back from this start and win some more games, that will eventually help Texas.
Not necessarily, each week is so different and my model is only designed to compare teams as to where they are right now. However, over the last two years, 21-22 of my top 25 teams will appear in the CFP selection committee's top 25 each week.
Kellner is a PhD student in Applied Technology and Performance Improvement at the University of North Texas. Kellner bases his measurement on the average of each opponent's individual rating that each team has played up to that point during the current season.
His results are currently being published in the Waco Tribune Herald and the USA Today.
You're ranking system came to be in 2002. For those who aren't familiar with it, would you break down what all figures into it, what fine tuning you've done for this season, and some tweaks you think are still possibly yet to come?
Wins, losses, and margin of victory are considered. Each team is given a game rating that is reflective of the result of the game (win or lose) and the strength of the opponent for that game. More points are given for defeating a higher rated then than a lower rated team. More points a deducted for a loss to a lower rated opponent than a higher rated opponent. Results from the previous season are considered, but diminish as a factor as the season progresses.
How much different has it been for your using your system and trying to accurately size up programs in the playoff era compared to the BCS era?
Not much of a difference, I have mostly been using the same approach since 2007. However, starting this season, I will be giving an additional penalty to a team if they have suffered a loss at any point during the season.
Do you think a lot of people know you got your start playing back yard football on Alexander street?
Haha, I don't think so, but those were great times!
Help us out with "quality losses.” For example, Oklahoma is considered to have two, but they still have two losses 4 weeks.
Oklahoma has lost two games, but they were against two teams (Ohio State and Houston) that can very well make it to the playoff. Situations like this increase the importance for ratings systems such as what I have as a 2 loss Oklahoma team should still be ranked higher than a team with a better record playing an easier schedule.
Explain the difference in Louisville and Houston right now in your mind and who is more likely to hold the season together or do you think they are both just early season frauds?
Based on the success Houston had last season, I feel more confident they can continue sustaining their success during the season in comparison to Louisville. However, Louisville pretty much controls their own destiny being in a Power 5 Conference.
If Louisville can go undefeated, they will make the playoff. However, based on being in a Group of 5 Conference, Houston must go undefeated and will also need help in order to make the playoff (such as Oklahoma winning the Big 12 and several Power 5 Conference Champions having multiple losses).
The inaugural season to the playoffs left a lot of Baylor fans in dismay. How did you grade the teams in that year and what were your thoughts on both TCU and Baylor being left out?
My top 4 teams that year were the same four teams that actually made it to the playoff. My system had eventual national champion Ohio State at No. 2 going into the playoff. Based on that, Ohio State was an easy choice to be in the playoff as they were the only playoff team that year that didn’t have an FCS team on their schedule, so their schedule was naturally tougher.
Outside of the top four, I had TCU at No. 5 and Baylor at No. 6. Based on the lack of wins against quality competition the Big 12 had in non-conference play that year, it was difficult to justify having TCU or Baylor in the playoff that season.
From your system, or even your own studies, do you see any conferences so far who are incredibly over or underrated?
I believe the SEC is the best overall conference right now, but the difference between the SEC and other conferences has been shrinking over the last couple of years. The Big Ten is continuing to make progress and I see them having success such as the SEC has recently enjoyed.
Per your system who are the 5 best teams over the past 14 years?
Normally, my results have been in line with the BCS and CFP over the years. Overall, I would say Oklahoma, Texas, Alabama, LSU, and USC have been among the best in my poll during this time. I will say there was a few years there that my model always seemed to underrate some really good Ohio State teams back when the rest of the Big Ten was struggling and hurting Ohio State's strength of schedule.
Is Notre Dame a legit fraud? How does this mess with Texas and their ranking?
Notre Dame's loss to Duke will hurt Texas' resume as they defeated a Notre Dame team that Duke defeated also, so the quality of win isn't as great as what it could have been. However, Notre Dame always has a tough schedule, so if they can bounce back from this start and win some more games, that will eventually help Texas.
Do you find your advanced knowledge of metrics helps you make accurate yet bold predictions? Do you have any for us?
Not necessarily, each week is so different and my model is only designed to compare teams as to where they are right now. However, over the last two years, 21-22 of my top 25 teams will appear in the CFP selection committee's top 25 each week.
Never miss the latest news from SicEm365!
Join our free email list