So I blew it. In my five years of being the BCS Guru, this was the worst flub I've ever had. Because one of the mistakes is so egregious, I owe you folks an explanation.
The first mistake was fairly minor, that Auburn went ahead of Oregon by .0002. I didn't foresee the Tigers gaining that much ground in the polls after a game that they were truly lucky to win (Alabama really should've had a 35-0 lead and put the game to bed in the second quarter). I knew Auburn would be No. 1 in all the computers, so as soon as both polls came out, I knew the exact margin and Tweeted it as much.
Now, the serious mistake was having Ohio State ahead of Wisconsin in the original projection. That of course goes beyond flip-flopping teams in the 5 and 6 spots because of what's at stake - a Rose Bowl berth.
First, there's the obvious question of whether this was a publicity stunt. Fair question. If it were, then that means I have no integrity and was going for the cheap thrill of getting undue attention. That's absolutely not the case. I've been toiling at this for too long and built up too much goodwill and trust than to risk it for some one-night stand.
Then it becomes a question of competence. Having your credibility in doubt isn't all that much better than having your integrity savaged. When no WMD was found in Iraq, President Bush was called either a liar or a dunce by his critics. Same here, but of course with much less at stake.
The erroneous projection happened because of a miscalculation of the polls and a simple data entry error. If only one of the two occurred, it would not have happened. That the fact both of them happened on this one projection (on Ohio State) made it a perfect storm.
First, the poll miscalculation. The polls are always harder to project than the computers because it involves psychology. You have to form an educated evaluation on the behavior of the voters. That's much more difficult than analyzing the computers, which follows a certain formula and ironclad logic. I had believed that by virtue of LSU's loss to Arkansas, thus removing it as a buffer between Ohio State and Wisconsin, the Buckeyes would gain significant ground in the polls.
That didn't happen in the margins I had forecast. Wisconsin lost only 26 points in the coaches poll and a completely unforeseen 24 points in the Harris poll, out of 114 voters. I had projected the Buckeyes to pick up about 100 points in the Harris poll with LSU gone, but that simply didn't happen.
Now you must wonder, given that I had to know Wisconsin would stay ahead of Ohio State in the two polls, did I completely screw up on the computers?
The truth is, I didn't. That's why this was such a mind-numbing debacle.
I simply transposed two projections. on Massey and Sagarin, I assigned the values of 20 (for No. 6) to OSU instead of OU on my spreadsheet, and 10 (No. 16) for OSU to OU.
By making these mistakes, Ohio State's computer numbers were inflated to .790 instead of .680 (OSU's actual BCS computer score was .660) and Oklahoma's was deflated from .780 to .640 (OU actual: .760). This explains why OU was also initially projected to be a spot lower (No. 10) than it ended up (No. 9).
I don't expect everyone who reads this blog to get all this - and you shouldn't have to. It was incumbent on me to check and double-check my own data before publishing the projections. Perhaps I was in too much of a rush to get the projections out (this has become a bad habit since I moved back to the West Coast), or maybe because I did so well the last two weeks (1-10 in exact order two weeks ago and 1-13 last week) that a little hubris got the better of meticulousness.
Whatever, it was my mistake and I own it. The only thing to do is to acknowledge it and make sure it won't happen again. I still maintain that my record speaks for itself over the years, against Jerry Palm, Brad Edwards, anybody. I have a report card dating from 2006 to prove it.
And finally, I don't mind the criticism and even the hate mail. I deserve it, and I won't erase any of it to cover my own ass. You can't make me feel any worse than I already do.