Someone created a computer algorithm that has analyzed the selection committee's tendencies over the last two years, and predict their next move. As someone who hates selection committees, and loves computer algorithms, I think this is super cool.
Every Tuesday in the second half of the college football season, the College Football Playoff Selection Committee releases its ranking of the 25 best teams in the country. The top four teams at the end of the season will go on to a four-team playoff to determine the national champion.
CFPMachine uses historical data and machine learning techniques to predict what the committee will (or would) do each week of the season (but not what they should do!). Not only does CFPMachine consider on-the-field success, it also considers off-the-field information (program prestige, average kickoff time, conference strength, etc.).
Current top 5:
- Alabama
- Clemson
- Michigan
- Washington
- Louisville
VT currently sits at 23.
Comments
Only a machine could put Clemson number two after the flat performances and extremely lucky wins over NC State (especially this one) and Louisville.
I think the Louisville win is very high-quality, but maybe it has more to do with the Auburn win, while Michigan's best win so far is Wisconsin or Colorado. I'd take Auburn and Louisville over that.
Yes, but what I'm saying is, machines can't fully understand how lucky Clemson got that NC State went on a 6 minute drive to end the game and missed what is an extra point in the NFL, from the dead middle of the field to not lose that game. Clemson's defense couldn't stop the drive, they didn't block that field goal, they didn't even influence it with a big push, the kicker just choked and Clemson got super lucky.
Actually advanced stats are designed to account for that, they don't care about the result of the game, only the result of the plays, and Clemson is still 4th overall in S&P+.
If you care about resumes and wins, you can't put too much stock into how lucky you got.
Conversely, if you care about quality of play, you can't put too much stock into one game and instead have to look at the overall body of work.
You can't have it both ways to make a team look bad.
Yeah I am familiar with the S&P+, FEI, etc but that doesn't mean that those metrics don't come out flawed sometimes. I don't think any other top 10-15 team would find themselves begging god for NC State to miss a field goal in order to beat them, and that's exactly what Clemson had to do. I think they should have already been eliminated from playoff potential with that loss. But here's to hoping Florida State beats them this week and maybe Syracuse gives them the orange sting the week after.
Louisville got lucky to beat Duke, Washington got lucky against Arizona, Ohio State got unlucky and lost to Penn State, Baylor and Nebraska and WVU have played nobody, Boise State got absurdly lucky against BYU, Auburn and Wisky and Florida State have two losses, Florida got lucky against vandy (and look atrocious generally), A&M just got blown out. The only two teams in the top 15 who don't look like they'd fall victim to the same trap are Bama and Michigan. Bama is a clear number 1, while Clemson has higher quality wins than Michigan.
At worst, Clemson should be ranked 3rd, but second is pretty reasonable if as a voter you value wins against quality opponents over magnitude (or luckiness) of a win.
Louisville didn't get lucky to beat duke they averaged almost 9 yards a play against them. They just played keep away and it still wasn't enough to beat them.
That said, luck is certainly a part of college football but Auburn was totally out of sorts and only looked good after they played one QB and stopped letting Malzahn call plays and Clemson looked average in that game. Then they beat Troy by 6, at home. Then they got super lucky TWICE in very unfairly officiated games against Louisville (constant PI and def holding not called) and against NC State where at one point NC st had 13 penalties and Clemson had zero including a no call super PI on Clemson's interception and an obvious facemask that was a no call while NC st was in the redzone that would have given them a 1st down inside the ten and instead they got no points.
Michigan has thoroughly beaten every team they have played even though Wisconsin was close their defense looked superhuman. The Colorado game was close for a bit, and Colorado has turned out to be a pretty good team, but they pulled away in that game as well. They also went on a ridiculous streak of not giving up any points. They have absolutely looked like a better team than Clemson from week to week.
Also of note: Clemson in it's last two home games (Louisville and NC State)
Opponent penalties - yards = 24 for 224
Clemson penalties - yards = 11 for 99
Stats according to ESPN
If I were betting, I'd take Clemson (or Louisville too honestly) over Michigan or Washington on a neutral site.
That being said - I'm of the belief that the most 'deserving' team should get the higher ranking. Clemson has the best win of any team right now (that auburn win is looking more impressive each week), and no losses. They pulled out an ugly game. It was ugly, but its still better than any other team's 'quality loss.'
EDIT: added response to above comment
My argument is that the committee members should not consider this at all. It doesn't matter how 'lucky' a team is - they got a win against an average (but decent) ACC team.
I would argue that NC State is not an average team, and Louisville dispatched them as such. Clemson wouldn't have beaten Louisville in a fairly officiated game or if James Quick wasn't a moron and went inside or through the defender.
But Clemson has beat Louisville and Auburn, two very good teams. Michigan and Washington have wins over....Colorado and Wisconsin. I would prioritize wins against good teams over lucky wins every day of the week.
I mean to be fair almost every national champion as long as i can remember has some kind of slip up where they're lucky to survive. I mean Terrance Cody's block against Tennessee is one that stands well in my memory
The difference is they have had to get lucky multiple times.
Another difference, Terrance Cody blocked TWO field goals in that game that would have won Tennessee the game, that is an Alabama player making a play, not getting lucky.
Clemson blocked one field goal, but the NC State kicker missed two field goals including an NFL extra point from the dead center to win the game. On this drive, the Clemson defense had multiple opportunities to stop NC State and instead they gave up a 6 minute drive that ended regulation. Then he just missed the chip shot, no play made by a Clemson player won that game in regulation, they just watched the field goal miss.
agree to disagree, it takes a certain amount of screwing up by the kicking team for a block to occur. Be it a bad snap, terrible blocking, or a bad kick. Both of those blocks were AWFUL kicks. kickers are important too and I hope NC State realizes that now after that game and their botched field goal this week too
I still think Clemson is a great team, and I still think Ohio state is a terrific team (probably outside of washington the few teams that can matchup well and put points up on bama). one bad performance does not change my mind.
I'm curious to find out exactly how they weigh program prestige. Is it a ranking system? Assigned weight? Etc.?
shit
what? our average should be confounded by these 7pm kicks
Ha. Just making a joke. We've only had two noonish kickoffs this season which is better than years past. The nooner joke will never die.
noon:30 has got to improve our overall chances lol