Why I don't like the BCS computers

[Dan Thompson]

The Harris poll, AP poll, and the Coach's poll all have

1. LSU
2. Oregon
3. OU

The stupid BCS computers have

1. LSU
2. Kansas
3. Oregon
4. ASU
5. Missouri
6. Ohio St.
7. OU

[mdklatt]

The Harris poll, AP poll, and the Coach's poll all have

1. LSU
2. Oregon
3. OU

The stupid BCS computers have

1. LSU
2. Kansas
3. Oregon
4. ASU
5. Missouri
6. Ohio St.
7. OU


Why does everyone assume that when there's a difference between the computer and the polls that it's the computers that are wrong?

[Dan Thompson]

Because I was a computer programmer for over 32 years and I know that garbage in garbage out.

[Flagstaffsooner]

The computers are still eating our azz for N texass and Utah state.

[ouradu]

It'll play out, but you know, the human polls have been wrong a lot too. The same polls last week said Ohio St. was the best team in the country. They were wrong, but for some reasons the computer polls and the human polls still have Ohio St. ranked ahead of Illinois. That's gotta be wrong, right? I don't necessarily have a problem with the way the computer polls are working out. Both the human and computer polls have their flaws. The computer polls look at everything very coldly, while the human polls sometimes let emotion and current perceived success cloud judgement. Without a playoff it's all opinion, whether it be a human ranking the teams or a programmer deciding what's more important and letting that formula rank them. I don't think we can complain too much either way.

[OKLA21FAN]

Because I was a computer programmer for over 32 years and I know that garbage in garbage out.
computers don't make errors....it is the programmers that make them



just sayin :pop:


but until the final computer polls come out, i will not pass judgment on them. OU will get some much needed computer love if they win out and KU and MU don't stumble in their other games.

[Paperclip]

We lost to a bad Colorado team. We're lucky to even be sniffing a chance to play for a MNC.

[mdklatt]

We lost to a bad Colorado team. We're lucky to even be sniffing a chance to play for a MNC.

Exactly. And we didn't exactly look like world beaters at Ames.

[mdklatt]

Because I was a computer programmer for over 32 years and I know that garbage in garbage out.

How is this any different than the thought processes used by biased and subjective human voters who only see a fraction of the games?

What is crappy about the computers is that they took out margin of victory, but that was a human decision.

[TopDawg]

The computer polls look at everything very coldly, while the human polls sometimes let emotion and current perceived success cloud judgement.

Yep. In my post, titled "Why I don't like the human polls" I'd like to point out that this week the humans (AP, Coaches, Harris) have given 148 first place votes to LSU, 57 to Oregon, 14 to OU, and 16 to Kansas.

If OU, LSU, Ohio State, USC, Florida, Texas, etc. had done what Kansas has done so far this year, they'd have every single first place vote. But since it's Kansas...ehhhhh.

[TopDawg]

2nd verse, same as the first.

[OK2LA]

The stupid BCS computers have

1. LSU
2. Kansas
3. Oregon
4. ASU
5. Missouri
6. Ohio St.
7. OU

Which stupid BCS computer standings are you checking?

http://www.sportsline.com/collegefootball/polls/bcs


1 LSU 0.98 1 2742 0.9793 1 1457 0.9713 0.99
2 Oregon 0.938 2 2651 0.9468 2 1407 0.938 0.93
3 Kansas 0.909 4 2498 0.8921 4 1344 0.896 0.94
4 Oklahoma 0.854 3 2580 0.9214 3 1366 0.9107 0.73
5 Missouri 0.81 6 2248 0.8029 6 1194 0.796 0.83
6 West Virginia 0.786 5 2278 0.8136 5 1238 0.8253 0.72
7 Ohio State 0.774 7 2100 0.75 7 1145 0.7633 0.81
8 Arizona State 0.75 8 1919 0.6854 8 1042 0.6947 0.87
9 Georgia 0.672 9 1840 0.6571 9 1005 0.67 0.69
10 Virginia Tech 0.613 10 1665 0.5946 10 923 0.6153 0.63

:pop:

[StormySooner-IN]

Which stupid BCS computer standings are you checking?

http://www.sportsline.com/collegefootball/polls/bcs


1 LSU 0.98 1 2742 0.9793 1 1457 0.9713 0.99
2 Oregon 0.938 2 2651 0.9468 2 1407 0.938 0.93
3 Kansas 0.909 4 2498 0.8921 4 1344 0.896 0.94
4 Oklahoma 0.854 3 2580 0.9214 3 1366 0.9107 0.73
5 Missouri 0.81 6 2248 0.8029 6 1194 0.796 0.83
6 West Virginia 0.786 5 2278 0.8136 5 1238 0.8253 0.72
7 Ohio State 0.774 7 2100 0.75 7 1145 0.7633 0.81
8 Arizona State 0.75 8 1919 0.6854 8 1042 0.6947 0.87
9 Georgia 0.672 9 1840 0.6571 9 1005 0.67 0.69
10 Virginia Tech 0.613 10 1665 0.5946 10 923 0.6153 0.63

:pop:Uh.
He posted the BCS Computer Rankings. Not the overall BCS ratings.
Overall BCS=#4.
**** Computers=#7.
:twinkies: ;)

[Stitch Face]

The stupid BCS computers have

1. LSU
2. Kansas
3. Oregon
4. ASU
5. Missouri
6. Ohio St.
7. OU

The logic is simple. We beat Missouri, but we lost to a team that is worse than we are. Since Missouri lost to a good team (us) and we lost to a mediocre team, Missouri is obviously better than we are.

Makes perfect sense, right?

[OK2LA]

Uh.
He posted the BCS Computer Rankings. Not the overall BCS ratings.
Overall BCS=#4.
**** Computers=#7.
:twinkies: ;)


Ah.

What "BCS Computer Rankings Bowl" do we go to if we finish # 7?

:twinkies:

[Stitch Face]

Aren't the computers just supporting the fallacious notion that wins and losses are transitive? E.g. OU and Missouri play in the same conference and have the same record. OU beat Missouri, but lost to Colorado, who Missouri beat handily. Therefore Missouri must be better than OU. The fact that we beat Missouri in an actual game against each other is apparently given less importance than how we each fared independently against Colorado.

Maybe instead of a playoff or championship game the top four teams in the country should each just play Colorado and then we compare the scores.

[birddog]

all this bcs crap is making my head spin. all the "what ifs" just get so tiresome.

i remember a thread about the bcs ruining people's enjoyment of college football and i agree with it. it's beginning to get comical.

[possumfat]

If we had taken care of business at Colorado this thread would not exist. We are fortunate to possibly still have the opportunity to play for the MNC.

[TopDawg]

Aren't the computers just supporting the fallacious notion that wins and losses are transitive? E.g. OU and Missouri play in the same conference and have the same record. OU beat Missouri, but lost to Colorado, who Missouri beat handily. Therefore Missouri must be better than OU. The fact that we beat Missouri in an actual game against each other is apparently given less importance than how we each fared independently against Colorado.

Again, this comes down to the question of "deserving v. best." I don't think the computers are trying to tell us which teams are best, they're just each giving us a different way to look at who is more deserving.

If all other things were equal, it might be crazy to have Missouri ranked above OU, because they lost to us. But not all other things ARE equal. We lost on the road to a crappy team, they lost on the road to a top-ranked team. Our most impressive non-conference opponent just got shut out at home. Their most impressive non-conference opponent just beat #1 on the road.

[Dan Thompson]

Richard Billingsley has USC ranked 7th (8-2) and OU ranked 9th.

[soonersn2007]

It's funny how some TV bobble heads thinks these debates make college football so great.............when in fact it only sours me on it.

[Stitch Face]

Again, this comes down to the question of "deserving v. best." I don't think the computers are trying to tell us which teams are best, they're just each giving us a different way to look at who is more deserving.

If all other things were equal, it might be crazy to have Missouri ranked above OU, because they lost to us. But not all other things ARE equal. We lost on the road to a crappy team, they lost on the road to a top-ranked team. Our most impressive non-conference opponent just got shut out at home. Their most impressive non-conference opponent just beat #1 on the road.

I think most people agree that there is no perfect way of determining "best," whether by playoff or the insane conglomeration of opinions and algorithms we use now. But any halfway-legitimate notion of "best" still seems more palatable than "deserving," which is what the pollsters seem to rally around.

I can dig the computers for taking into account all the other things (SOS, etc.) that humans naturally overlook due to their own flawed sense of reason or out-an-out biases. It just seems like the outcome of actual on-field competition is increasingly diminished. To be fair you could also say it took a hit when KSU busted our azz right before we waltzed into the NCG. That year it worked to our favor, this year it's not (so far.) It's screwed up either way.

[Stitch Face]

It's funny how some TV bobble heads thinks these debates make college football so great.............when in fact it only sours me on it.

Agreed. The one, single, isolated, only thing I prefer about the NFL is the absence of polls. People/sportswriters/commentators can talk all they want about who is the best team in the NFL/AFC/NFC but that's all it is...talk. You can listen to it, you can engage in it, you can ignore it. The outcomes are the same.

In college ball all that biased hot air actually translates to team standings at the end. And that's bull****.

[mdklatt]

I can dig the computers for taking into account all the other things (SOS, etc.) that humans naturally overlook due to their own flawed sense of reason or out-an-out biases. It just seems like the outcome of actual on-field competition is increasingly diminished.


Is CU better than us? They beat us head-to-head.

oSu beat Tech. Tech beat A&M. A&M beat oSu. Who's better?

The computers attempt to answer the questions that cannot be answered by head-to-head play.

[Jdog]

Aren't the computers just supporting the fallacious notion that wins and losses are transitive? E.g. OU and Missouri play in the same conference and have the same record. OU beat Missouri, but lost to Colorado, who Missouri beat handily. Therefore Missouri must be better than OU. The fact that we beat Missouri in an actual game against each other is apparently given less importance than how we each fared independently against Colorado.

Maybe instead of a playoff or championship game the top four teams in the country should each just play Colorado and then we compare the scores.

:confused: Related to OU I'm getting the impression that the Computers calculate only for the QUALITY OF LOSSES, NOT THE QUALITY OF WINS?????:confused:
WOULD WE BE IN BETTER SHAPE IF WE LOOKED AT QUALITY OF WINS?

[Stitch Face]

Is CU better than us? They beat us head-to-head.

oSu beat Tech. Tech beat A&M. A&M beat oSu. Who's better?

The computers attempt to answer the questions that cannot be answered by head-to-head play.

Yeah, I get it. But CU doesn't have nearly the same record we do, do they?

Tech and aTm have identical conference records at this point and their overall W/L record is comparable so that's a wash for all I know. OSU is close behind them both. You're just using the transitive fallacy I mentioned above to cloud the picture.

OU and Missouri are in the same conference and have identical records, which is about as 'equal' as two teams can get for comparison purposes in this day and age. The fact that OU beat Missouri in an actual game (a shocking thing to consider, I know) would seem to dictate that OU is perhaps slightly 'better' than Missouri at this point. I know it's crazy.

But if you agree with the computer rankings then more power to you.

[Stitch Face]

Again, I appreciate the fact that the computers take into account the things that human opinions and simple W/L records don't. My point is that, in this case, the math is off. We can tweak the formula, we can increase/decrease the weight of the polls. It's still all just made up based on homerism, commericialism, and arbitrary numbers in the guise of objectivity.

[Jdog]

Again, I appreciate the fact that the computers take into account the things that human opinions and simple W/L records don't. My point is that, in this case, the math is off. We can tweak the formula, we can increase/decrease the weight of the polls. It's still all just made up based on homerism, commericialism, and arbitrary numbers in the guise of objectivity.

How true -

[aero]

Here's what gets me the most. Below is a copy of the bcs computer poll of the top 25 teams. The "RB" poll, which I think is the "college football research center", (correct me if I'm wrong), and look at the discrepancy between it and the others. My questions are - what information are they using to come up with the ratings? It seems the "computers" are no different than human polls. In other words, if some programmer deems a certain stat or category important or more pertinent than another stat or category, how is that not bias? I've never understood why there needs to be 5 or 6 computer polls. Does ANYBODY know how each of these computer polls are derived? Why can't the BCS just come out with a criteria of stats, s.o.s., margin of victory, etc. to compute and use 1 computer poll? The fact that they actually use 6 polls tells me they know the computers are only as good as what's programmed into them and they really aren't reliable so they use enough that they can derive an average. Either that or they are getting a hefty kickback from the computer polls.



BCS Computer Polls updated 12-Nov 4:43 PM
High and Low ranking are thrown out

Ave BSC
Team KM JS RB ST PW WC Rank Pts
1 LSU 1 2 1 1 2 1 1.25 0.99
2 Kansas 2 1 6 3 1 4 2.50 0.94
3 Oregon 3 3 3 2 3 2 2.75 0.93
4 Arizona St 4 4 5 4 5 3 4.25 0.87
5 Missouri 5 5 10 5 4 6 5.25 0.83
6 Ohio State 6 6 2 6 6 5 5.75 0.81
7 Oklahoma 8 7 9 7 7 9 7.75 0.73
8 West Virginia 9 8 4 9 8 7 8.00 0.72
9 Georgia 7 10 11 8 9 8 8.75 0.69
10 Virginia Tech 10 9 8 14 10 12 10.25 0.63
11 Florida 11 11 15 11 13 10 11.50 0.58
12 Virginia 14 16 24 12 11 11 13.25 0.51
13 Southern Cal 18 14 7 10 14 19 14.00 0.48
14 Clemson 12 12 17 16 17 13 14.50 0.46
15 Boston College 13 13 22 19 16 14 15.50 0.42
16 Illinois 16 17 13 15 20 20 17.00 0.36
17 Cincinnati 21 20 16 18 19 16 18.25 0.31
18 Connecticut 19 18 20 20 18 17 18.75 0.29
19 Tennessee 15 15 23 17 23 21 19.00 0.28
20 Texas 22 23 21 13 21 15 19.75 0.25
21 Michigan 20 21 18 22 15 23 20.25 0.23
22 South Florida 24 22 33 26 22 18 23.50 0.10
23 Kentucky 17 19 30 24 30 22 23.75 0.09
24 Penn State 25 27 19 21 25 26 24.25 0.07
25 Wisconsin 23 26 14 23 31 25 24.25 0.07

[Stitch Face]

BCS Computer Polls updated 12-Nov 4:43 PM
High and Low ranking are thrown out

Whew! Good thing they toss the outliers! Keep any questionable data from gettin' in there.

[MI Sooner]

Don't forget that Missouri actually has a quality non-conference win (Illinois on a neutral field).

We played Texas, but Texas isn't getting much love (and rightfully so) from the computers (zero big wins).

One thing that's unfortunate is that we're penalized for playing a crappy 1A (I refuse to use the "Bowl Subdivision" BS) team, but other teams aren't penalized for playing 1AA teams. Sagarin ranks 1AA teams, why don't the other computer rankings? Or if they do, why aren't they used?

[Crucifax Autumn]

Blah-blah-blah to all this...

The simple fact is that in a couple of weeks the human polls and the computers are both gonna be showing us a LOT more love if we win out. The human polls for obvious reasons.

The computer polls will show us better because we will have a few quality wins added to our stats and some teams ahead of us will lose, lose some SOS, etc.

Think of it this way. John gets a check for $1,200. Fred gets the same check. John pays his bills as soon as he gets his check. Fred figures, what the hell, I have until the 1st to pay and goes out drinking and celebrating over how much he has in the bank. John sees what he has left and cuts back, eats Ramen noodles, and waits for the next check. Fred spends half his money having fun then on the 1st decides to finally take care of all those bills. He's got $600 left and $850 in bills. Fred is in debt...John is in the black and goes to the bar to drink beer all night and watch the Sooners go to the BCS championship!