It's the dawn of a new year, and roughly one-third of the 2012-13 NBA season is complete. By now we have a pretty decent idea of which teams are truly good and bad, but while NBA teams' records tend to stabilize over far fewer games than their counterparts in football and baseball, scheduling and early-season randomness can still play a role in team records, even 30 games into the season.
Cutting through the fog, a good analyst still needs to account for a team's point differential (usually a better indicator of true team strength than basic win-loss records) and the strength of opponents faced thus far, with a healthy dose of regression to the mean included for good measure. After a third of the season, even a power rating based on schedule-adjusted margin of victory needs to be regressed about 45 percent toward preseason expectations.
After taking all of that into account, we can construct a set of predictive ratings which can then be fed into a simulator for the purpose of projecting future outcomes. So which teams will ring in 2013 with a surprising January? And, conversely, which teams are primed for a disappointment? Let's go to the simulation results: