<

Note:

This year, three conferences (AAC, CUSA, PAC-12) switched from tracking conference standing as two independent divisions, to standing within the conference as a whole. AAC and CUSA made the call well in advance, but the Pac-12 reached their decision relatively late in the preseason. As a result, only seven predictors submitted picks with the Pac-12 lacking divisions.

There is a way to convert division-less predictions into divisional ones -- just filter out the members of each division into two separate lists. For example, if someone predicts the Pac-12 division-less as: 1. Oregon, 2. Utah, 3. UCLA... then we could sort Oregon as 1st in the North division, and Utah as 1st and UCLA as 2nd in the South division.

The downside of this is that it loses some "fidelity" of picks. It gives us an apples-to-apples way to compare standings to predictions, but only by converting to what is essentially a slightly lossy format. The other option would be to come up with a way to interleave the divisions into one division-less list, but that would be partially fabricating the predictions from some external source, which is even less desirable.

This issue has arisen once before (in 2020). When it did, I adjusted scores by a factor to ensure that the average score for divisional matched the average score for division-less picks. But this comes with two issues: (1) it implicitly assumes that the division-less predictions were on average as good as the divisional ones; and (2) it assumes that the "spread" between good and bad scores will be somewhat equivalent. After further thought, I've felt it best to do with the lower-fidelity (but more fair) scoring method.

This page ranks the accuracy of the preseason magazines' conference predictions for 2022, by comparing them to conference finish.

Lower scores indicate more accurate predictions; the lower the score, the closer the predictions were to the actual results. Magazines are sorted by total score for all conferences, with the best predictions (lowest total scores) at the top. The best prediction for each conference/division is highlighted in the list. For information on the scoring method, see the notes file.

<