Unfortunately, if you polled half the country, you'd probably get hundreds of different top 10's. Which top 10 you gonna use? [/quote] I don't think the BCS is all it should be and I'd like to see a less "human" (biased/flawed/ignorant) influence on ranking.[/quote] I'd like to see rankings eliminated completely. Example. After the LSU Florida game, peope every where began the "LSU and Florida are the 2 best teams in the country talk". You know why? Because people already had in their minds that we were number 1. So if Florida could play us that tight, they must be number 2. Did we really earn the #1 spot? No, we were given a REALLY high pre-season ranking, and were able to move up one spot. If we had been ranked # 15 in the pre-season, everyone would say Florida was over rated. Because of these polls/rankings, people are influenced before they ever even see a game played. Sounds an awful lot like the attitude people had when we switched from the old way to the BCS. Now everyone hates the BCS. I understand problems with seeding will alwas exist due to the fact not enough of these teams play each other, or have like opponents. But before an arguement can be made for seeding, you have to find the most fair and un-biased way of seeing who gets in. Any poll voted on by any person or machine will be biased and flawed, and will eventually be considered just as much a failure as the current BCS. How do you know? Most people, myself included, didn;t think Boise (or Utah back in '04 could hang with the big boy. Guess what. They did hang. So us sitting here, saying these teams don't belong, just because we think they don't, is already a major problem. Who get's in and who doesn't should not be based on anyone's opinion. I have no problem with rematches happening, if they happen to result due to a bracket being played out. So that is zero improvement from the BCS in this department. If we're not going to eliminate the bias, we're not fixing ANYTHING. The problem with the system prior to the BCS was biased opinion decided who was got a chance at the title. The problem with the BCS, is that Bias weighs heavily on who get's to play for the title. And you just admitted that bias opinion dictates who get's to play for the title. Do you see a pattern developing? Which is also just as flawed as the BCS. With the number of upsets that have already happened, I would say no one in the top 15 is safe. Everybody has a shot. Which is all the more reason to completely scrap any sort of ranking system. Auburn got absolutely abused by the press and pollsters for playing a close game against K-State, and losing to USF (and this was before losing to Moo U.). Why? because we THOUGHT K-State and USF weren't even poll worthy. We were wrong. Because of our pre-concieved notions, we penalized a team much more harshly than we should have. An example going the other direction. In '03 or '04 (I forget which) Arkansas was listed as "others recieving votes". Then they beat Texas. So naturally, we shoot them all the way up to #14. They must be a darn good team if they beat Texas is what the pollsters thought. They were wrong. Because before the game, everyone THOUGHT Texas was a sure fire top 5 team, Arky had to be an awesome team to beat them. Translation: we have no clue who would beat who on a neautral field. So any system based on opinion would be invalid at best. It was exciting and important to any team that currently has one or fewer losses. Most of those teams wouldn't be invited to a playoff, because we THINK they're not good enough. Kinda like we THOUGHT South Florida wasn't going to be anything special this year. Without looking at any other polls right now, let's take a quick vote. Let's get 20 people just on this board, and let them all list their top 12. I'm willing to bet some teams will be on some lists and off of others, and vise versa. You know why? Because we have no clue who the 12 most deserving teams are. It's all guess work. Someone is going to be left home because we THINK they're not good enough. (Personally, I think 12 is too many for a playoff, just pointing out that a ranking system is so severly flawed, that it should be eliminated from deciding a champ) possible, but I have my doubts as to whether or not the Presidents would sign off on a system that didn't garauntee their champ a spot. It would be similar to saying that the NFC South champ isn't going to the playoffs this year, because we feel a team from the NFC East is better.
You're preaching to the choir, Nutriaitch... if fact, somewhere around here I've proposed my own ranking system that exclusively uses results on the field. To recap in case you missed it, here's a quick summary: A team that wins at home gets a point. A team that loses on the road loses a point. A team that wins on the road gets 1.25 points. A team that loses at home loses 1.25 points. At the end of the season each team will have a raw points total (could be negative, positive, or zero). A percentage of each of your opponent's raw points total (maybe 5-10%) is added to your raw points total to give your team an adjusted points total. This gives teams that beat teams with good records more points than teams that beat losing teams, and more points for winning on the road, and more penalties for losing at home. BUT... maybe I'm just too cynical, but I don't see any sort of ranking system that doesn't use human pollsters being implemented anytime soon - we've actually backtracked, giving the human polls more emphasis. SO... in lieu of an "ideal" ranking system, I'll take a 4-12 team playoff using the flawed ranking system over the 2-team system we have now.
See, now you're on to something. I actually had a similar system that I put together a while back. I wouldn't call it a ranking though, because that implies votes/bias. I think your points system may need to be tweaked a little though. 2 things. 1st the points system has to be easy enough for the average Joe sitting at home to be able to figure out rather easily. 2nd, without seeing it play out, I don't know if your adjusted point will properly weight the games played. Here's my example. (all points figured off opponents FINAL won/loss record) 1 point home victory against team with worse record than you. 2 points home victory vs team with equal or better. 3 point rd victory vs team with equal or better record. 2 points rd victory vs team with worse record. -3 home loss vs team with worse record -1 home loss vs. team with better record -2 road loss vs. team with better record -1 road loss vs. team with better/equal record But you still run across a small problem. Unbeaten LSU would probably get equal credit for a home win vs. 1-win UL-M, as what 1-loss Cal would get for beating 2-loss Oregon at home. Nobody would consider those games even remotely equal.
I don't know that either. I know the NCAA has all of the game results, and if I get some free time I may see if I can come up with a way of quickly taking that data, importing it into a database, and then be able to run the various output reports I'd need. So your example has some similarities to mine. The major differences that I can see are: 1) I give extra points (and take away extra points) based on winning on the road/losing at home, but I do that regardless of what your opponents' final record winds up being. I do take into account the opponents' records by factoring in a percentage of each opponents' raw points total to each team's final adjusted points total. I'm not saying my way is superior to your way though. 2) I don't give (or take) as many extra points as you do for winning on the road/losing at home. That's just something that I think we'd have to figure out where the right balance is.
Of course the rating system has a subjective element involved. The only way you could have a completely objective system is to base it strictly on win/lose records without any regard for strength of schedule. I don't believe that would be such a good thing. But the current system is not completely subjective. The win/lose record of a team is a factor as well as its defensive and offensive stats.
The reason I weight each game, is so that a team get penalized harsher for losing @home to Tulane than they would for losing @ home to Florida. And likewise, you should get ore credit for winning @Florida than @Tulane. Do you have some type of formula figured out yet for adjusting the points at the end of the season? If so, I'd love to see it. Maybe I can figure out a spread sheet, and at least put up some of the points for the top 25, and see what it's looking like compared to the polls currently in existence.:thumb:
Sorry for the ridiculously long manuscript... but here it is anyway. My point is that the regular season of college football is unique. It is more like a season long playoff than what any other sport has, and I believe that is a key component of what makes college football so beloved to so many fans out there. By the end of the regular season "playoff", most unbiased observers feel confident that only a handful of teams are national title contenders. An efficient playoff would begin where the regular season ended. I believe at this point in time the true number of contenders is less than 8, and while that may be growing with parity, I don't think we're beyond that point yet. It is always easier to start small and grow - it is easier to add games than it is to take them away. You've pointed out more than once that you realize a smaller playoff is more likely to be implemented at first, so I won't go on about that. A larger playoff would in effect be a second season where all the teams start over, at least to some extent. There may be seeding handicaps, but participating teams still have some semblance of a new life. I truly believe this second season would take excitement away from the regular season the way march madness takes away from college basketball's regular season. Sure, there would be some excitement over teams trying to make the playoff, but the focus would go away from each game to making the playoff. Like I said before, I believe the bowls would benefit from an efficient playoff and be damaged by a larger playoff. A larger playoff takes more teams from their pool, takes more attention away from the bowls, and really reduces the bowls to NIT status or worse. An efficient playoff might actually add teams to the bowl pool and be less of the sole focus of the post season. Secondary bowls, which are important to college football, would be better off. Also, a larger playoff makes life more stressful for every program. Imagine if LSU missed out on the playoff for a number of years while still going to a respectable bowl. Fans would be unhappy and feel like not making the playoff is basically a total failure. Stress on coaches and expectations to perform would be through the roof. The idea of a playoff is very attractive, especially when obviously deserving teams are left out, but we must be careful. While a playoff may answer the problems that we see, it may cause new, unexpected problems. We must be conservative in calling for a playoff and try to be as thoughtful as we can to just what getting a playoff would mean. I think this is quite right and I pretty much welcome it, though I'd love see LSU always dominant. Too much parity might not be everything it is cracked up to be (for the traditional powers). I hate that nonBCS teams don't get more of a chance, too. I believe that a nonBCS team making a top 10 ranking is equally (if not more) impressive as a BCS team making a top 5 ranking. Rating systems have inherent biases against nonBCS teams that aren't easy to accurately adjust for. I agree with this. Still, I would love to see the computers play a larger roll in the BCS ratings, but I would also like to see only well designed computer ratings used. I question some of the computer rating systems that the BCS uses. Thanks for the civil and intelligent discussion. If things keep going like they've been going, there are going to be a lot of critical voices at the end of the season wondering why only two teams are playing for the NC. It really is an embarrassment that insults the intelligence and passion of college football fans. I'd use a combination of the best computer rating systems. Computer rating systems may be biased in that the weight given to some components of the total available data changes based on what the computer rating designer believes is important, but that is totally different than rating teams based on name recognition and historical relevance. Computer rating systems are as unbiased as you can get, and intelligently designed formulas are more accurate and more consistent than opinion polls. On the other hand, computers can't see everything, so I'd still put some weight on the opinion of professionals much closer to the pulse of college football than the average fan. I personally believe the AP Poll to be the best of the opinion polls. A smart combination of computers and opinions is much more effective than using strictly conference champions. By the end of the season, there is enough data to pinpoint the top teams in the nation. When LSU won the SEC in 2001, we had a nice team, but most unbiased observers would agree that LSU did not belong in the national championship conversation. That is based on educated opinion, but also on a complete season of on the field data that should not be ignored in lieu of an upset in one game at the end of the season. I don't disagree that opinion is fickle and changes with the wind, especially during the season, but by the end of the season, things have been played out and can and should be examined. I wouldn't suggest relying solely on human opinion, but educated opinion should not be ignored. You criticize the bias of human opinion, and I don't totally disagree. But when you criticize the bias of computer ratings (not rankings, which imply integer positioning, and not polls, which imply opinion), I ask what exactly do you think your simple rating system is other than a simplified computer rating system? You are coming up with a formula, weighing specific data which, in your opinion, is important. I wonder why you believe that a formula that the "average joe" can easily follow is a more accurate simulation of the reality of college football. It seems that the more precise the formula used, the more accurate the rating system is, assuming the formula is well thought out. I say this from experience, as I have a formula significantly more complex than what you've come up with that I believe more accurately represents what is important in rating teams. I'd be happy to share them with you. I don't mean to toot my own horn, but a lot of work has been put into them. My rating system has been effected by many hours of thought and collaboration with others more or less interested in rating teams. I appreciate the dialogue with you and always hope to be intrigued by fresh and innovative ideas. Sorry if I sound full of myself. I just want to let you know I'm not totally pulling these arguments out of my rear. In the case of nonBCS teams I totally agree. Often, and arguably through little to no fault of their own, successful nonBCS teams just don't have the same amount of meaningful data draw conclusions from. Which is why I promote special consideration for nonBCS teams. BCS teams, on the other hand, have every opportunity during the regular season to prove the worth. If we want every game to be as important as possible, BCS teams that have dropped the ball on one or more occasion have only themselves to blame. Most people believe the regular season in college football is the best in organized sports anywhere, and that is in large part because of the importance of every single game. If we start putting more importance into one game at the end of the season (which is sometimes only of regional significance) than into each game played, the feel of a "regular season playoff" is diminished, which I believe is bad for college football. I think most would agree that rematches are less desirable than non-rematches, and while I am not totally opposed to rematches in a playoff, if steps can be taken to avoid rematches by more accurately recognizing the significance of the first match-up, I'm all for that. I'm afraid we don't live in a perfect world. We can strive for perfection, but attaining it is a rare endeavor. We can strive to eliminate bias, but we have to realize we live in a finite world with failures related to the palpable limitations of the world. To say we have no clue is wrong. The science of statistics has made possible and clear many things which before it were murky and confused. In statistics, there is no such thing as a sure thing, but some things are more likely than others. First of all, I don't believe conferences should be guaranteed a position to play for a NC, I think that should be earned. But I also believe that as diverse a field as possible playing for the NC is a good thing. The NFL's playoffs are far from efficient, and the regular season is significantly less exciting than college football's as a result. I could put my formula up if you're interested. I'd also like to scrutinize with you the fields of my playoffs from 1995 to current, if you're interested.
