Let's start with this question. Is there a systematic method for making 68 predictions on an event that is over 48 weeks away?
Every team in the country performed at a certain level for 30-plus games last season. Each team returns some percentage of minutes played from the players responsible for that level of performance.
So we start with an assumption -- albeit inexact -- that returning players will not regress in their established level of performance. More specifically, we assume that the combination of improving players will at least equal the performance of those who regress or lose time to injury and other absences.
Next, we look at the departures. In addition to graduating seniors and transfers, players projected in the first round of the NBA draft are considered gone, even if they've yet to announce their intention for early entry (that's you, Jabari and Joel). We can add them back into team calculations later, as needed.
ESPN recruiting grades -- or prior school statistics, in the case of transfers -- are the baseline for determining the performance level of newcomers, in combination with an estimate of the available minutes each new player will assume.
I no longer make any qualitative adjustments to this data. This bracket is strictly a numeric seed list of 68 teams (including projected automatic qualifiers) placed into the regionals and subregionals of a new year. Over the summer and into the fall, I'll begin to make subjective determinations based on things like coaching changes, projected schedule strength, news reports and other roster issues. Thankfully, there isn't as much conference realignment to potentially skew the data as was the case a year ago.
This method does seem to occasionally overvalue teams with an unusually high number of returning players (On, Wisconsin!). And recruiting rankings tend to be haphazard, at best, while undervaluing the most important newcomers. But I've found the numeric approach to be more accurate, on balance, than the typical "eyeball" method for such things. It also provides a legitimate baseline for the many personnel changes that inevitably occur between now and November.
One area in which our calculations have been a slight net negative is in the ability to forecast No. 1 seeds. So I offer a little of my own way-too-early eye test to set the following top-line odds for the 2014-15 season. Here are my odds-on favorites to land on the top line come Selection Sunday 2015: