Apparently, Jeff Sagarin doesn't think his new and "improved version" of his new formula for the BCS to use is improved enough. He made a significant change to it again during the week, so now the data that appears on his site is very different than the data that was used by the BCS last week.
You may recall that last week, Sagarin introduced a new version of his Elo-based formulas. I wrote then that the column named "Elo Score," which is now described as including margin of victory, bore a very strong resemblance to the "Elo Chess" version it replaced, which Sagarin clamed did not contain MOV. Meanwhile, a new non-margin version called "Pure Elo" was introduced and would be the version the BCS would use.
However, those results were very counterintuitive. Seven FCS teams were in the top 25, led by Bethune-Cookman at No. 4. However, because of Sagarin's mid-week change, the Wildcats are now 67th, but still the top-rated FCS school. That's a pretty dramatic change.
Among the top 25 FBS teams in his system, which are the ones that receive a BCS score, Oregon State and Virginia Tech switched spots. In the official BCS release, and in Sagarin's original rankings last Sunday, the Beavers were sixth and the Hokies seventh. Now, that's reversed. That would have improved Virginia Tech's BCS score slightly, but not enough to improve the Hokies overall BCS ranking. Sagarin was already the highest rating for OSU by far and is thrown out, so the Beavers' score doesn't change.
So, is this much ado about nothing? There's no way to know. His formula is secret. His process is unaccountable. It should be noted that the name "Elo Score" didn't last a week. It is described the same way as before, but is now called "Dimin Curve." Also, that rating, as well as his "Pure Predictor" ratings were also tweaked as part of his changes this week.
Hopefully, he's done, but unless he makes mid-week changes again, we'll never know. When he does, we can see different results with the same data, so we know the formula changes. It's disconcerting to know he's not satisfied with his formula yet and that it may continue to change.
I have contacted Jeff, but have not yet received a response. When I do, I will update this post.