Computer rankings help teams like South Florida because they don't take prestige into account. Chris Livingston/Icon SMI

Want to know more about the polls and computer ratings used by the BCS? Check out Stewart Mandel's new book Bowls, Polls and Tattered Souls: Tackling the Chaos and Controversy that Reign Over College Football.

MAILBAG Submit a question or an opinion to Stewart. Your name: Your e-mail address: Your home town: Enter your question:

Well, the first BCS standings are out. Stewart, would you like to comment on why the computer rankings differ so much from the Harris Poll and Coaches' Poll? And being that the computer rankings only accounts for 33 percent, why include them?
--George Winstead, Chesapeake Beach, Md.

I'm no math or computer expert, so you're going to have to look up the individual Web sites for Richard Billingsley, Wes Colley, et. al., if you want to know their exact methodology (and each one is different from the others). Generally speaking, however, I can think of four major differences between the computers and human polls -- some of which are good, some of which aren't.

1) Silly intangibles like prestige and recent history play no factor. It's no surprise USF is held in higher regard by the computers (where the Bulls are No. 1 in five of the six, No. 2 in the other) than they are by humans (where they're No. 3 in both the coaches and Harris polls). To the computers, "USF" is just another set of data, not some unfamiliar, 11-year-old Big East program. And that's a good thing, because apparently such biases are still alive and well. Harris Poll voter Eddie Crowder, Colorado's coach from 1963-73, told CBSSports.com this week that, "I haven't even seen a brief highlight of [USF]," and that he couldn't name a single Bulls player. If his final No. 2 vote came down to deciding between USF or LSU, Crowder said he'd likely side with the Tigers because, "They're better stabilized ... They've been there, done that for 100 years."

Are you kidding me? Did I just walk into some sort of time warp where it's 1963 again and there's only one game on television all week? All but one of the Bulls' games have been available to anyone with basic cable. And what on earth do the last 100 years have to do with the 2007 season? The sad thing is, Crowder is probably par for the course among the Harris panel, which consists heavily of retired coaches and ADs. So in this regard ... advantage: computers.

2) There is no arbitrary starting point for the teams. Some of the computer pollsters do put out preseason editions, but they become meaningless once the games begin. In the computers, Ohio State and Arizona State are currently tied for fifth, which makes sense because they have nearly identical credentials. Both are 7-0. The Buckeyes' wins have come against opponents with a combined 15-18 record (Division I-A games only), while the Sun Devils' foes are a combined 16-20. In the human polls, however, OSU sits 12 spots ahead of ASU, the single biggest reason being that the Buckeyes started the season 10th in the coaches poll while the Devils were unranked. Again, advantage: computers.

3) Margin of victory is not a factor. As you may recall, prior to the 2002 season, the BCS honchos asked its participating computer gurus to remove any margin-of-victory component from their formulas. While their intention was to dissuade coaches from running up the score (and for the most part, they've succeeded), in doing so, they kind of made a mockery of the whole rankings concept. In the computers' eyes, USF's 64-12 victory over UCF last week is no more or less impressive than Texas' 35-32 escape against the Golden Knights earlier in the season.

That doesn't make much sense to me, nor did it to several of the mathematicians, a couple of whom refused to comply and walked out. Another, Jeff Sagarin, did make the adjustment but posts both those results and his true rankings every week. How different are they? In the BCS version, Oklahoma checks in at 12th (and has an average of 11). In the original Sagarin version, the Sooners are No. 1. Advantage: humans.

4) Early season ratings aren't necessarily accurate. Like with any statistics, the more data that's made available, the more accurate they become. Which is why the computers have always included a disclaimer that their ratings in say, Week 6, may skew drastically from those at the end of the season. Basically, you're not getting a complete picture, hence such "wacky" anomalies as Michigan checking in seven spots higher than Oregon -- the same team the Wolverines lost to 39-7 -- in Billingsley's computer while barely registering in some of the others. In other words, there's no "common sense" override. Advantage: humans.

That last one was more of a problem back when the computers carried greater weight. In the original BCS standings, the human polls only comprised 25 percent of the overall formula, but after AP and coaches No. 1 USC finished No. 3 in the standings in 2003, organizers revamped the thing so that the coaches and Harris polls now account for a combined 66 percent of a team's score. As a result, it would take a pretty drastic deviation by the computers for the final standings not to mirror the final polls. You can decide for yourself whether or not that's a good thing.

Stewart, please do your best to defend your obviously East Coast-biased media colleagues with regard to their love affair of Boston College QB Matt Ryan and their dismissal of Texas Tech's Graham Harrell as a system guy
--Matt Johnson, Dublin, Calif.

I shall not do that, because this is a rare instance where I actually believe the "East Coast bias" thing to be 100 percent true.

No doubt Ryan is a very good quarterback, but his purported status as a Heisman front-runner is completely a product of gushing East Coast media types. That starts and first and foremost with ESPN's Connecticut-based NFL draft gurus, who have attached a certain level of credibility to Ryan as opposed to some other quarterbacks (Harrell, Chase Daniel, Dennis Dixon, Tim Tebow) who have produced similar or better stats against similar or better competition but are not necessarily viewed as future pros. Meanwhile, the New York Times has covered Ryan considerably, as has my esteemed publication, both of which are also East Coast-driven.

Don't get me wrong, Ryan is having a fine season (63.2 percent completions, 2,148 yards, 17 touchdowns, six interceptions) ... but we're talking about the nation's 37th-rated passer here. His biggest "Heisman performance" to date was throwing for 435 yards against 4-3 Georgia Tech. All of the quarterbacks I mentioned above -- along with more prototypical candidates Brian Brohm and Andre Woodson -- are ranked higher than Ryan, and have all turned in bigger performances in bigger games. So obviously there's an X-factor keeping Ryan above most of those guys on the various Heisman lists, and Matt, you nailed it.

Now, if Ryan goes off against No. 11 Virginia Tech's defense next Thursday night -- then we should start talking Heisman.

3 of 4