Discussion: Why are the CFP spots decided via a committee vs. average of polls (i.e. BCS rankings)?
It is something I think of every season as I see the inconsistencies in the CFP committee's criteria and ranking logic.
I remember the old Bowl Championship Series (BCS) system from 1998-2013, and everyone's gripe with how at the end of the day the system only allowed for two teams to compete for the National Championship. So birthed, the College Football Playoff system...
However, it never made sense to me why the BCS ranking system of taking a conglomerate of computer-driven and human-drive polls was abandoned. It seemed to average out the opinions of multiple sources, and at the end of the day spit out a pretty agreeable top-4 teams.
Was there actually a legitimate argument against using the BCS rankings to determine the playoff teams?
November 10th, 2021 at 11:39 AM ^
You aren't wrong. Let's think about this, we essentially moved from:
(1) not necessarily perfect but absolutely known and transparent algorithm to
(2) the decision being made by a committee that can basically do what they want.
I think that was 100 steps backwards. Things like the Colley Matrix aren't perfect --- no mathematical ranking system is --- but at least the algorithm was publicly available and transparent (sort of like the old RPI and current NET).
I don't even know why they interview Gary Barta (or whoever the committee head is in a given year). The committee head basically can say whatever they want, and 75% of the time they don't really answer the question anyway.
November 10th, 2021 at 11:51 AM ^
Remember back when the 'computer polls' decided things? Sagarin's algorithm for CNN was the first of these, and it was eventually determined to be biased towards huge winning margins - so teams ran it up as much the could. It also scored 'well known teams' (ie: Bama, Florida, OSU, etc) on a slight curve. Other polls, later used in BCS rankings, were found to have similar issues. The committee approach was considered a step up from the BCS crap when it came about...
November 10th, 2021 at 11:55 AM ^
That's partially true ---- the BCS DID use computer algorithms that included margin of victory for the first few years. But all of those algorithms were purged by 2003.
That includes the Sagarin ratings. Sagarin ratings were part of the BCS algorithm from 2003-2013 but he published a set of numbers for the BCS that explicitly considered only wins and losses (every single algorithm was binary wins and losses from 2003-2013).
November 10th, 2021 at 11:52 AM ^
In 2003, Oklahoma was guaranteed to play for the BCS title BEFORE the Big 12 title game. Kansas State (w/ 3 losses) blew them out by 28 points, but still went to the Sugar Bowl over 1-loss USC even though the human polls had them #3. Why? Because of the computers.
That system should've been abolished after the season.
November 10th, 2021 at 11:57 AM ^
But what was really the issue in 2003? Was it the algorithms, or was it the fact that only 2 teams made the "playoffs"?
I'd argue it was the latter. If a 4-team playoff existed in 2003, we don't have a Oklahoma, USC and LSU issue, of 3 teams for 2 slots.
The "best" system right now, IMO, would be a 4 (or higher)-team playoff where the teams selected are all algorithm based.
November 10th, 2021 at 12:55 PM ^
I would argue that there was a major issue with the algorithm. USC was #1 in the human polls and dominated the competition, except for an OT loss at Cal.
People can argue about who should've been #2 between LSU and Oklahoma and expanding the playoff, but any algorithm that put that USC team at #3 is extremely flawed.
November 10th, 2021 at 1:17 PM ^
2004 had an even stronger case for a 4-team playoff, as USC, Oklahoma, and Auburn all went undefeated.
Auburn got left out and blubbering SEC shills have been punishing college football fans for it ever since
November 10th, 2021 at 11:59 AM ^
Because people bitched about using the polls (especially when intentional poll manipulation seemed to be a thing) so they decided to go to a selection committee like NCAA basketball, since no one complains too much about that.
However they failed to realize that no one complains about that because of the 64-team format, not because of the selection committee.
November 10th, 2021 at 12:34 PM ^
The vast majority of AP voters are too arrogant to ever admit when their pre-season rankings are wrong so they continue to reward teams like UC and MSU for early season victories over at the time highly ranked, but now clearly bad teams like IU and Miami. The other voters are just slappies for their teams who also down-vote rivals and other conferences.
And the coaches poll was just self-serving crap with most of the ballots being filled out by low-ranking members of the AD who don't watch any games.
The committee isn't perfect but it's a bit better than the previous idiocy of the polls.
November 10th, 2021 at 12:57 PM ^
The real reason is that the College Football Playoff was created by and for the purpose of serving the Power 5 conferences and their member schools, and a committee is the best way to maintain that cartel. They made the playoff and want to control it at their own terms. Their strong bias toward power 5 teams is evident in Cincinnati's rankings so far this year.
There are no referees, or higher power to make appeals of fairness to, in the world of college football. No one has the power to force them to use any objective system for ranking and seeding.
November 10th, 2021 at 1:26 PM ^
Cincinnati is not doing itself any favors. When you look like you are struggling to beat 2-7 Navy and 3-6 Tulsa, and you already play an incredibly weak schedule, things are going to be tough for you.
Yeah they beat Notre Dame, and that is obviously good.
They probably deserve to be in the playoff. But probably not any more than a 1 loss OSU, Michigan or MSU do. That is the problem--there are probably going to be more than 4 deserving teams.
November 10th, 2021 at 1:59 PM ^
To make a college football play off work fairly, the following would need to be done for an 8 team playoff modeled off the NFL.
1. Take humans out of the decision making process. No humans voting in polls, no humans writing code and no humans on committees making decisions. Let's face it, none of us are Solomon.
2. Make everybody who wants to participate join a conference. Everybody plays the same number of regular season games.
3. Level all the conferences with the same number of teams in each. If you want divisions within the conferences, all the conferences have to have divisions. I'd recommend not having divisions as that will skew some of the results in 6 below.
4. Make determining the conference winner work the same in every conference.
5. The 5 conference winners go to the playoffs. If you end up with 6 conferences after all the realignments in 2 above, then the 6 winners go to the playoffs.
6. You fill in the remaining 3 slots with the remaining 3 best teams (call them wildcard spots), just like the NFL does. Head to head competition (if applicable), no head to head?. No worries, move on to other criteria: best won-lost-tied percentage in games played within the conference, best won-lost-tied percentage in common games, minimum of four, strength of victory, strength of schedule, best combined ranking among conference teams in points scored and points allowed.
In other words, it's all objective math.
For the 3 wildcard teams (or 2 in the case of 6 divisions), you'd have to tweak the criteria as there aren't the number of games played across conferences as there are in the NFL.
November 10th, 2021 at 3:03 PM ^
Overall this does sound more fair
So why do I find college football endlessly fascinating, while the NFL is bland as hell?
November 10th, 2021 at 7:46 PM ^
1) The NFL is designed for mediocrity. They want as many fanbases engaged as possible. This leads to a ton of teams in the 8-9 to 10-7.
2) The NFL is set up to minimize football and maximize commercials. When they had issues with games running long due to all the commercials, they just went to a running clock all the time to cut out more football.
November 10th, 2021 at 2:21 PM ^
Money and power
November 10th, 2021 at 2:28 PM ^
1) FBS is too big to support a common sense playoff.
2) Outside of a couple dozen teams (most years) P5/G5 is a distinction without a difference that's manipulated to continue making the rich richer and the poor poorer.
Once the conference moves are complete good luck figuring out which of the ACC/AAC/BIG12/Sun Belt are P5 and which are G5.
November 10th, 2021 at 5:14 PM ^
Because polls don’t take $$$ into consideration
November 10th, 2021 at 5:55 PM ^
The BCS rankings were horrible because the computer standings were stripped of being able to use any reasonable criteria to determine rankings due to politics. They weren't allowed to consider margin of error (because it might encourage poor sportsmanship) to name one, and by the end of the BCS era they were absolutely meaningless.
If you re-did the BCS standings and used actually worthwhile ranking systems like SP+ or what have you and then averaged them with the AP Top 25, you'd have something sensible.
November 10th, 2021 at 6:37 PM ^
Is this system perfect? No. But I believe it’s an attempt at something that makes sense - football people who watch the important games.
the coaches poll? Some low level person who has been cobbling film all weekend puts this together for a coach. Next to worthless as these coaches maybe get to see what percentage of games that aren’t theirs in a weekend?
AP poll - is it better? Maybe but better than a panel? No.
As for quips about “people getting screwed” in a larger format is a stretch. There’s 6-10 legit teams a year on average. After that - yes you may be last team out looking in - but screwed is a stretch. Such teams likely had multiple losses or a bad schedule or insert any number of other issues here. So right now, legit good teams might get left out such as Cincinnati. Could be a paper Tiger or legit. But have an argument. Spartans maybe depending on how year ends. Same for Michigan. After that, meh.