Monday, February 21, 2011

Reasonable Post-All-Star Break Expectations For the Atlanta Hawks

The Atlanta Hawks have won 61.8% of their basketball games so far this season. If they maintain that pace for the remainder of the season, they'll (rounding up from 50.7) win 51 games.

Every day publishes John Hollinger's playoff odds, the results of 5,000 simulations of the remainder of the season. At the All-Star break, the playoff odds predict the Hawks will win 46 games.

A five game difference over a 27-game sample? That's pretty severe. Which method is likely to provide the more accurate prediction?

Because neither method includes all relevant information, my guess is that they're equally likely to provide an accurate prediction and the Hawks will win (barring a major trade or major change, either positive or negative, to the rotation) between 46 and 51 games this season.

You see, Atlanta's current winning percentage does or does not (depending on your preferred semantic perspective) reflect that the Hawks played the league's easiest schedule prior to the All-Star break. Furthermore the Hawks played the easiest schedule by a wide margin. The difference between Atlanta's strength of schedule (opponents' winning percentage : .459) and 29th-ranked Miami's strength of schedule is 18 percentage points, equal to the difference between Miami and 22nd-ranked Philadelphia. From here on out, the Hawks will face opponents with an average winning percentage of .549 which is 31 percentage points higher than the league's toughest schedule faced (by Houston) so far this season.

Hollinger's playoff odds know about this strength of schedule discrepancy which goes a long way toward explaining its relative pessimism about the Hawks over the final 27 games of the regular season. But the playoff odds should not be taken at face value because, in their sterile, mechanical state, they do not account for the fact that the two losses, by a combined 75 points, the Hawks suffered in the last month came in Al Horford's absence and are unlikely to be replicated if the team's best player remains healthy from here on out.

Judged by their season-long efficiency margin, the Hawks are a 30-25 (54.5%) rather than a 34-21 (61.8%) team. Remove those two home losses from the measure, and the Hawks are 33-22 (60.6%) team. Three-quarters of the difference between Atlanta's record and their expected record is attributable to those two losses. One cannot dismiss those losses entirely, they exposed both the teams greatest weaknesses as much as they revealed Horford's importance to the team but any purely statistical predictions will give those two games too much importance.

So I say, consider 46 wins to be the minimum number of games the Hawks should be expected to win this season and 51 wins to be the maximum number of games the Hawks should be expected to win this season. Should their final total fall outside that range, be it high or low, serious credit or blame must be apportioned.

No comments: