The "accuracy" table for each year presents a scoring of the preseason magazines' conference predictions. There is one row in the table for each preseason magazine, and one column for each conference (or conference division). The magazines (rows) are sorted by overall accuracy (total of scores for all conferences), with the most accurate magazines (lowest scores) listed at the top.

Each cell in the main body of the table is a scoring of one magazine's (row) predictions for one conference (column). The score is the sum (over all teams in the conference) of the difference between predicted and actual finish of each team. The more accurate the prediction, the lower the score.

When a team in a conference finishes tied for a position, the score for that team is computed from the most favorable of the tied positions. For example, if Michigan finished in a three-way tie for 3rd place, Michigan would be counted as finishing in any of 3rd, 4th, or 5th place -- whichever was closest to the predicted value for each magazine. A prediction of 1st would score two points (1st is two places away from 3rd); 2nd would score one point (2nd is one place away from 3rd); 3td, 4th, or 5th would score zero points (exactly right); 6th would score one point (6th is one place away from 5th); etc.

For example, consider Preview Sports' 1996 predictions for the ACC, versus the actual final results:

Those magazines which did not have picks for all conferences are disqualified from the overall standings, and are presented at the end of the table. However, overall scores have been computed for them anyway -- by adding the "average" score for each non-predicted conference (plus a two-point penalty per non-predicted conference).

The best overall score in each column is presented in bold emphasis. (Note that bold emphasis has a different meaning for the "total accuracy" file; there it indicates magazines that did not pick every conference.)