BCS POLL IS AS GOOD AS ANY IN THE NATION.

Byline: KEVIN MODESTI

We interrupt America's annual BCS-bashing festival for this word of special interest to the college football fan in the house:

That ugly bucket of bolts, that computer, the one that picked Nebraska to face Miami in the Bowl Championship Series showcase game at the Rose Bowl on Jan. 3, might not be as stupid as you think.

The computer is as smart as the writers.

The computer is smarter than the coaches.

The computer is even smarter than you, Bucko.

Troubling, but true.

This is not meant to defend the BCS method for determining a national champion, because nothing can be as satisfying as a playoff system would be. Nor is this a knock on Oregon and Colorado, which make good arguments that they deserve the trip to Pasadena more than Nebraska does. Nor is this merely a matter of opinion.

This contrary notion - that the computer formula just might have got it right - is an objective fact based on the bowl-game results in the three- year BCS era.

For all the talk about the weaknesses of the BCS, has anyone stopped to ask how good a job the statistical hash of poll rankings, computer ratings and degree-of-difficulty factors really does in assessing football teams? Well, I guess I just asked.

Is there an answer? Sure.

Just like a human know-it-all, the BCS' electronic brain puts its reputation on the line every time one (or two) of the teams in its 15-deep ranking plays a game. If a ranked team plays an unranked team and the ranked team wins, then the computer's judgment is supported; if the unranked team wins, the computer is embarrassed. If two ranked teams play and the higher-ranked team wins, bully for the computer; if the lower-ranked team wins, it's egg on the computer's desktop.

So let's look back at the 1998-99, 1999-2000 and 2000-01 bowls and see how the BCS computer did. Each of those seasons featured 10 bowls that included one (or two) BCS-ranked teams.

In the first season, the higher-ranked team won only 5 of the 10 games. In the second, the higher-ranked team won 7 of the 10. And in the third - last winter - the higher-ranked team won 8 of the 10. Add it up, and the BCS computer had a 20-10 record at picking winners.

Is that good? By way of comparison, let's see how the Associated Press (writers') poll and the ESPN/USA Today (coaches') poll did in forecasting the outcomes of the same 30 games.

The higher-ranked team in the AP poll went 20-10, and the higher-ranked team in the ESPN/USA Today poll went 19-11. (The polls' disagreement came in the 1999-2000 Alamo Bowl. AP ranked Penn State No. 13 and Texas A&M No. 18, while ESPN/USA Today had the Aggies No. 13 and the Nittany Lions No. 17. Penn State won.)

So the BCS rankings seem to be just as accurate as the writers' rankings - and a little more accurate than the coaches' rankings. Both of the traditional polls are components of the BCS formula.

Here's another comparison that might surprise you: In the same 30 bowl games, the point-spread favorites went only 16-14.

Most notably, the oddsmakers had Tennessee as a five-point underdog against Florida State in the 1999 Fiesta Bowl and Oklahoma as a 10-point underdog against Florida State in the 2001 Orange Bowl. The BCS, AP and ESPN/USA Today polls predicted the upsets in those games.

Distressing news: Since the final point spread is, essentially, determined by the flow of the betting, and the betting is done by fans like you and me, the poor showing by bowl favorites suggests you and I know less about football than the BCS computer we love to criticize.

What else does it all mean?

This year, again, there will be 10 bowls that include teams in the BCS top 15. The glaring disagreement between the BCS rankings and the other two rankings involves Nebraska (No. 2 according to the BCS but No. 4 in the other two) and Oregon (No. 4 according to the BCS but No. 2 in the other two). Between those teams is Colorado (No. 3 according to all).

Critics are taking the other two polls' higher opinion of Oregon and lower opinion of Nebraska as evidence the BCS rankings are screwed up. But the results of the past three years say the two polls own no greater claim to accuracy than the BCS rankings. The computer has as much of a right to be right.

Even if all of America had faith in the computer's judgment, of course, the BCS would remain a terrible way to set up a championship game, especially after a season with as many ambiguities as this one.

All that math is too complicated for even the most sophisticated fan, player or coach to understand. (If the BCS' rankings are exactly as accurate as the AP's, why not just use the AP's?) Teams are made to feel their fates are in the hands of some guy behind a pocket protector. It's too bad the title isn't decided entirely on the field.

What's needed is a playoff, but until that fantasy comes true, college football could do worse than to trust the BCS computer.

CAPTION(S):

box

Box: SIX BRUINS HONORED