Championship Subdivision Football | FCS Football | Stadiums | Blogs | Forums

1:20 in...

http://sports.espn.go.com/ncf/news/story?id=5680983

Gotta love when computers decide who plays for the BCS "National Championship."

How anyone can support computer rankings is beyond my comprehension...

disregard - technical issues on my side

disregard my post

CrackerRiley wrote:How anyone can support computer rankings is beyond my comprehension...

Computers don't have ridiculous bias like human voters do.

DJH wrote:Computers don't have ridiculous bias like human voters do.

Throwing darts isn't biased either. I guess you think that's an accurate way to do a poll too.

Um...no duh...with the likes of West Chester, South Dakota St., and Duquesne on the schedule...who wouldn't?

Hey, I didn't say they are a good thing.

But I do believe that if formulated the right way, computers are often a better ranking system than humans.

DJH wrote:Hey, I didn't say they are a good thing.

But I do believe that if formulated the right way, computers are often a better ranking system than humans.

Formulated by whom? Humans?

Hey, I'm as big a UD fan as you get & there isn't any doubt these computer rankings over-inflate how good some of the FCS teams are. Nobody is going to argue if you say UD or ASU are the top FCS teams so far this year. To call UD the 35th best team in the country, including FBS, may be stretching things too far. If you wanted to say they're better than half the FBS teams, maybe. Granted Ohio St has played a weak schedule to date but UD isn't in OSU's stratosphere.

Keeper

1 Appalachian State 5-0
2 Delaware 6-0
3 James Madison 4-1
4 Western Illinois 4-2
5 Jacksonville State 6-0
6 Stephen F Austin 4-1
7 Massachusetts 4-1
8 Montana State 5-1
9 Eastern Washington 4-2
10 William & Mary 5-1
11 SE Missouri State 5-1
12 Northern Arizona 3-2
13 Georgia Southern 3-2
14 Wofford 4-1
15 Chattanooga 3-2
16 Montana 4-2
17 North Dakota State 4-2
18 Bethune-Cookman 5-0
19 Northern Iowa 2-3
20 South Carolina State 4-1
21 Furman 3-2
22 Villanova 3-2
23 Cal Poly 4-2
24 South Dakota State 1-4
25 Sacramento State 3-3
26 Southern Illinois 3-3
27 Liberty 4-2
28 Central Arkansas 3-2
29 Missouri State 2-3
30 Texas State 3-2

DJH wrote:

CrackerRiley wrote:How anyone can support computer rankings is beyond my comprehension...

Computers don't have ridiculous bias like human voters do.

Yes they do. They have whatever bias the programmer has. If the programmer thinks strength of schedule is important the computer rankings will reflect that. If the programmer thinks margin of victory is important the computer will also reflect that.

If I'm a Boise St. fan (or any non-BCS conference school) and I'm writing a program to rank college football teams I'm going to give margin of victory more weight than strength of schedule.

If I'm a fan of a BCS conference school I'm going to give strength of schedule more weight.

BlueHen86 wrote:

DJH wrote:
Computers don't have ridiculous bias like human voters do.

Yes they do. They have whatever bias the programmer has. If the programmer thinks strength of schedule is important the computer rankings will reflect that. If the programmer thinks margin of victory is important the computer will also reflect that.

If I'm a Boise St. fan (or any non-BCS conference school) and I'm writing a program to rank college football teams I'm going to give margin of victory more weight than strength of schedule.

If I'm a fan of a BCS conference school I'm going to give strength of schedule more weight.

But thats not a bias. Thats just weighting the data differently.

A bias is when someone thinks Ohio St is better than Boise St in a given year because OSU is traditionally a better program.

Computer rankings take into consideration actual data and actual results, without the bias.

DJH wrote:

BlueHen86 wrote:
Yes they do. They have whatever bias the programmer has. If the programmer thinks strength of schedule is important the computer rankings will reflect that. If the programmer thinks margin of victory is important the computer will also reflect that.

If I'm a Boise St. fan (or any non-BCS conference school) and I'm writing a program to rank college football teams I'm going to give margin of victory more weight than strength of schedule.

If I'm a fan of a BCS conference school I'm going to give strength of schedule more weight.

But thats not a bias. Thats just weighting the data differently.

A bias is when someone thinks Ohio St is better than Boise St in a given year because OSU is traditionally a better program.

Computer rankings take into consideration actual data and actual results, without the bias.

It's bias if the program writer has an agenda, just like it's bias if the pollster has an agenda. One is manipulating data to make his team look better, the other is filling out a poll and placing his team higher. Either way, it's bias, the computer is a tool just like a pen is a tool.

It's also bias if the people who select the computer polls to be used in the BCS also have an agenda.

And just to be clear: I'm not saying that there actually is bias in computer polls, I'm just saying that there is the potential for it.