Last year's record and stats should have no effect on the current year. That is a very poor modeling decision, likely just an effort to have some data, no matter how inapplicable, to make up for the lack of data early in the season.
Why is using data from a prior season bad? It promotes the status quo which favors the blue bloods. It discounts coaching changes, and quality of coaching. It discounts the effects of transfers. It discounts the effects of conference realignment, and the resulting recruiting boost and increased revenue.
Other than that, let's throw in last year's results to downplay the current results of the upstarts.