BCS Changes Bad for VT and the Big East

by Will Stewart, TechSideline.com, 7/2/02

The changes announced to the BCS ranking system last week can be boiled down to one truth for Virginia Tech football fans: it's getting harder and harder every year for the Hokies to make the championship game, and this year, it gets even harder.

In case you missed it -- and I wonder how you could, given that the story was plastered all over the major web sites like ESPN.com, CNNSI.com, USAToday.com, and others -- the BCS, currently under the direction of Big East Commissioner Mike Tranghese, has decided to completely eliminate margin of victory from its computer polls.

That was the major change. There's another change, to the "quality win" component, which rewards a team for victories over other teams that are highly ranked in the BCS. But right now, it's the margin of victory component that drew the most ink and that we'll talk about here.

Now before you glaze over, hang in there, because this stuff is important to Hokie fans. To find out how important, check out this quote from Kenneth Massey, who runs one of the seven computer polls in the BCS system:

"'99 Virginia Tech wouldn't have gone to the championship game under this new system."

Now, are you hooked? You ought to be, because not only were the '99 Hokies your favorite team, but they were a great team too, and deserved to be in the championship game.

First, a quick primer: the BCS determines which two teams play for the national championship. It's comprised of five components: human polls (AP and USA Today/ESPN), computer polls, schedule strength, number of losses, and a "quality wins" factor. I'll stop there, and if you want to learn more, check out the Official BCS web site (scroll down the linked page) for a complete explanation.

In 1999, when the Hokies went to the national championship game under the BCS system as it was then, they did it with a weak schedule, but they compensated for that schedule by hammering the teams they played. Tech won 9 of their 11 games by 20 points or more, including whipping arguably their two toughest opponents, Syracuse and Miami, by a combined score of 105-10.

Even so, the undefeated Hokies faced a stiff BCS challenge late in the season from a one-loss Nebraska team that finished just over a point behind the Hokies in the 1999 BCS standings, 6.12 points to 7.42.

Nobody had a real quibble with the 1999 Hokies playing FSU in the Sugar Bowl that year, but in 2000, the Miami Hurricanes got jobbed when an FSU team they had defeated early in the season got to play in the national championship game against Oklahoma. So for the 2001 season, the BCS added a "quality wins" component that rewarded teams for victories over other teams ranked in the Top 15 of the BCS.

In other words, a variation of a strength of schedule component.

Also changed in 2001 were the computer polls. Of the eight polls (at the time), half were to have the margin of victory eliminated. Four pollsters volunteered, and the BCS was set for 2001.

Then the 2001 season was played, and a one-loss Nebraska team that didnít even play in its conference championship game was ranked #2 in the BCS, ahead of a Colorado team that defeated them and an Oregon team that was ranked #2 in both human polls.

The outcry began, and when the 'Canes used the Huskers to manicure the historic Rose Bowl field, it intensified.

The solution? More tinkering. Hence last week's changes to eliminate margin of victory from the remaining computer polls and to alter the quality win components.

Kenneth Massey's Take

So what does it all mean? To understand the BCS changes, I called Kenneth Massey, the VT Ph.D. student who authored and administers one of the computer polls used in the BCS.

Kenneth, I asked, what's it all mean?

"Strength of schedule's going to be way overrated," he answered. "Now the computers are going to go solely on strength of schedule and wins/losses, and strength of schedule's going to dominate that for sure.

"There's going to be no way for a team to compensate [for a weak schedule or a game against a weak opponent]. If Virginia Tech plays Temple and beats them 56-0, that shouldn't cause Virginia Tech's rating to go down, but under the new system, it will, just because they're playing a bad team."

In other words, a 7-6 squeaker over Temple is worth as much as that 56-0 blowout. Something's wrong with this picture.

"Last year [2001], they took margin of victory out of half of the computer polls," Massey recalled, "so four used it and four didn't. I agreed at that time to take it out of mine. I agree with the issue of public relations and not wanting to encourage people running up the score.

"But I think that if they force all the computers to take the margin of victory out, they're going to have some pretty serious problems at the end of the year when you have an SEC team with one or two losses ranked ahead of an undefeated Big East team, or maybe a Big Ten team that perhaps doesn't have a strong non-conference schedule."

Then he utters the line that makes Hokies fans sit up and take notice: "Like '99 Virginia Tech -- '99 Virginia Tech wouldn't have gone to the championship game under this new system."

Massey believes that an accurate computer rating system definitely includes margin of victory. He participates in an email list with his fellow BCS computer ranking programmers, like Richard Billinglsey, David Rothman, and Jeff Sagarin, and Massey says, "Those of us who are degreed mathematicians and have their doctorates, the ones that have technical backgrounds, we see the flaws in it and have gone over it, and a couple, like Rothman and Mathews, just refuse to go along with it."

The BCS will only employ seven computer rankings this season, because as Massey said, Rothman and Mathews took the mathematical high ground and opted out. That whittled last year's eight rankings down to six, and the BCS added The New York Times' ratings back in (after a one-year absence) to boost the computer rankings back up to seven.

"The rest of us, we're going along with it, because it's not our decision," Massey says. "It's college football's decision. We're just kind of along for the ride. We still have the right to publish what we feel to be our best ratings, and so, for example, I continue to publish the ones with the margin of victory on www.mratings.com, and provide the others [without margin of victory] as a service to the BCS."

Strength of Schedule

It's not a surprise that strength of schedule is getting yet another boost over margin of victory. All of the changes since 1999 have been geared towards it, geared towards pushing the BCS ratings to be more like NCAA basketball's RPI ratings, which are used to help seed the NCAA Tournament. RPI ratings, which Massey calls "silly" and "a hodge-podge formula," rely purely on strength of schedule and wins/losses, while ignoring margin of victory.

"The formulas that I use, and Sagarin uses, for example, they're documented in professional journals and based more heavily in mathematical theory," Massey sniffs. "The RPI is something an amateur would sit down and make up."

It's a little alarming that strength of schedule would get yet another shot in the arm during Big East Commissioner Mike Tranghese's oversight of the BCS (each of the BCS conference commissioners rotates through a one-year oversight of the BCS, and this year, it's Tranghese's turn). It's Tranghese's own conference that is perceived as being the weakest BCS conference, both by humans and by the computers, so any increase in the strength of schedule measure is going to reward conferences like the SEC and Big 12 and hurt the chances for Big East teams to make the national championship game.

The prudent thing to do, in this writer's opinion, would have been to order the computer pollsters to cap margin of victory at 15 or 20 points, and then inform the college football world, loudly, that the limit had been set.

According to Newport News Daily Press columnist David Teel, that very idea was suggested, but Tranghese shot it down.

"We think eliminating margin of victory is middle ground," Teel quoted Tranghese as saying, "because margin of victory is a component of the human polls."

Yeah, well, strength of schedule is also a component of the human polls. And a component of the computer polls. And a component of the strength of schedule component, for crying out loud.

"Strength of schedule is a more valid evaluation tool," Tranghese added.

Those are disheartening words, coming from a man whose conference includes Temple and Rutgers. He's in charge of the BCS, and road blocks for the Big East teams are getting bigger under his watch.

Never mind the Miami Hurricanes, who "get it" and have the luxury of signing anyone they desire to a home-and-home series (Florida, FSU, and Tennessee are all on the 'Canes schedule this year), plus the benefit of the doubt in the pollsters' minds. The damage here is being done to the Hokies, and perhaps the Boston College Eagles or Pittsburgh Panthers, if they were to field a strong team and make a run.

Hokie fans like to point towards 2003 as a good chance for the Hokies to make a run at the national championship. They'll have experience on both sides of the ball and a schedule that includes all of their tough games -- Syracuse, Miami, BC, and Texas A&M -- at home.

But that schedule also includes lightweights Central Florida, UConn, and JMU, in addition to the usual diet of Rutgers and Temple, not to mention that WVU might still be reeling. And now it doesn't matter to the computers how badly the Hokies pummel those teams (assuming they do). Tech will get knocked down a notch in the BCS computer ratings just for playing them.

In 1999, Tech might have played that schedule, won all the games, and gone to the championship game. Heck, they basically did play that schedule, minus Texas A&M.

But in 2003, the odds are stacked significantly against a championship run, no matter how Tech does.

Is this whining? No, not really. You can argue that champions should play tough schedules. But let's not go overboard and make it all about strength of schedule. The point remains that the 1999 Hokies, who belonged in the championship game, wouldn't make it in the new system, according to Kenneth Massey.

What if that fate befalls the 2003 Hokies? Or for that matter, any other team from 2002-2005 that might be one of the two best teams in the country, but doesn't have a murderous schedule or the luxury of belonging to the SEC or Big 12?

The Truth

Tranghese and others wanted to eliminate margin of victory to attempt to stop coaches from running up the score. The thing is, while the changes will indeed alter the BCS rankings outcome, they won't alter coaches' decisions to run up the score, because a margin of victory component is still present -- in the human polls.

If anything, now that margin of victory isn't part of the computer polls, coaches might be more likely to run it up, in order to crank up their precious human poll rankings.

The next time Bobby Bowden has Duke down 49-0 and the Noles have the ball at midfield, Bowden isn't going to think, "Hey, now that them dadgum BCS computers donít use margin of victory, we'll just take a knee."

No, he's going to think, "Dadgummit, I gotta score me a couple more TD's, so we can hang on to that #1 ranking in the AP and coaches' polls, because playing Duke's gonna kill me in those dadgum computer rankings. Rix, drop back and let her fly."

          

TSL Columnists Archives

TSL Home