Jeff Anderson knows most college football fans are wary of the computer rankings that make up one-third of the Bowl Championship Series rankings.
Anderson and Chris Hester team up on one of the six computer rankings used by the BCS, Anderson & Hester.
"The knee-jerk reaction, which has been fueled from the start by some people in the media, (computers) are an easy target," Anderson said. "It's an inanimate object, what does it know?"
Yet, Anderson believes the computers get a bad rap.
"A lot of the people who follow the sport more closely over the years have found the computer rankings to be a better reflection of what teams have done than the (human) polls offer," he said. "They're both crucial, I think.
"You've got to have some sort of objective way of looking at these teams. The objectivity of the computers, I think, tends to improve the way the poll voters vote.
"If there's some team they have wildly off from where the computers have them, they probably take a second look. LSU, for example, is a team that has been high in the computers all year, but was not very high in the polls. Now, finally, this week you see a bump with their win over Alabama. LSU is a team that's been winning all year long, however miraculous it has been."
'The BCS trusts we're professionals'
Since Anderson had a lot of useful information about the computer rankings, which I couldn't include in today's story due to space limitations, I'll try to get to some more of his points here.
One of the criticisms lodged against the computers used in the BCS rankings is that only one of the six publishes the formula used for the rankings. This leads to obvious questions among coaches and fans about the methods being used for the rankings, as well as who is feeding the information into the computer.
I contaced Asher Feldman, who runs a website (bcsknowhow.com) that analyzes the BCS standings, for his take on the computer rankings. Feldman attempted to look into the formulas used by the six computers last summer. The only one to publish its formula, Feldman said, is the Colley Matrix.
"The computer formulas are pretty well-kept secrets," Feldman said. "Five of the six don't publish their formulas. Even the one that does, it's hard to replicate. You need to be pretty well versed in computer science or Excel documents if you want to replicate that formula. They're all very hard to predict."
Anderson said there are proprietary reasons for keeping the formulas secret, but that the rankings used are all well-established and respected.
"I think the BCS trusts we're professionals," Anderson said. "We've been doing this, all of us, since before the BCS came in existence. Everybody publishes what their rankings are based upon. I don't think anybody would be able to make much sense of the specific math. Nobody probably wants to dive into that anyway."
Non-conference slate hurts UW
As for UW's low computer ranking — it has a computer average of 10, including lows of 14 in the A&H and Colley Matrix — Anderson made it clear the Badgers' easy non-conference schedule is to blame.
Other than wins over Ohio State (No. 13) and Iowa (No. 16). UW's best win is over No. 70 Purdue (based on A&H rankings). Four of the Badgers' eight victories are either against an unranked Football Championship Subdivision opponent (Austin Peay), or teams ranked worse than 100: UNLV, San Jose State and Minnesota.
As Feldman pointed out, the A&H also puts a heavy emphasis on strength of conference. It ranks the Big Ten fourth, behind the Big 12, Southeastern and Pacific-10. Other computers have the Big Ten third. That also has an impact on the Badgers' ranking in A&H.
"It's an interesting year for the Big Ten," Anderson said. "There's tremendous parity at the top. It's certainly making it interesting. I think it's going to be a lot of fun to watch it the rest of the way."
Here are the A&H rankings for the Big Ten:
8. Michigan State
13. Ohio State
27. Penn State
By comparison, the Big 12 has seven teams — more than half of its conference — ranked in the top 25.
The Badgers' strength of schedule is also not helped by not playing Penn State and Illinois in the conference this season.
Computers don't project matchups
One point Anderson wanted to clear up about computers is the fallacy of trying to gauge who would win a head-to-head matchup between two teams.
"That's another thing fans tend to get confused about, and the media can be misleading on this sometimes: The computer rankings are trying to reward teams for the accomplishments that they've achieved on the field," Anderson said. "They're not looking forward and trying to project who would win some hypothetical future matchup. There's a difference between those things.
"The point of our rankings is try to reward teams for what they have actually done to date."
Anderson was actually fairly complimentary about the Badgers and their remaining schedule against Indiana, Michigan and Northwestern. If the Badgers win out, they would pick up two more solid wins against Michigan and Northwestern.
"Wisconsin has been good," Anderson said. "Winning two out of three of the high-profile games (Michigan State, Ohio State and Iowa) is clearly impressive, with two out of three on the road.
"Once the Badgers get a couple supplemental wins under their belt, something under those two (Ohio State and Iowa), against the top 70, our rankings are going to reflect the fine season they're having more clearly."
Because the Harris and coaches' polls, which each count one-third, tend to be similar, the computer rankings typically have a big impact on the BCS rankings. But Anderson believes giving two-thirds weight to human polls and one-third to the computers is the right mix.
"I think you have to have that subjective element," he said. "It's the fans' game. In the end, you need to have a national championship game the people believe is the right matchup."
Anderson believes the mix of the objective computers and subjective human voters is a good balance.
"Not only do I think there's room for both, I think it's essential to have both," he said. "You don't want it to be a popularity contest and I don't think anybody wants to just have the computer rankings spit out a verdict without the subjective evaluations of voters. I think the blend of the two is essential."