It would probably be more interesting if I understood what you were talking about. Unfortunately I dont. however, good job taking the time to put out a thorough analysis. For that ill at least throw you a TP.
I don't think I worded my post in the clearest way, to be sure.
On November 15th, we were debating if a team's win-loss success at that point in the season was an accurate representation of how good a team was. Or, if Hollinger's rankings was a better representation of how good a team was.
The assumption was that better teams would win more games in the ensuing months than bad teams. If the team with the #1 win-loss record is really the best team, you'd expect them to win the most games over the next 2 months.
I decided to try to track a simple way to see if what win-loss record said was the "best" or "Worst" team at a given time ended up accurately predicting how good or bad a team played in the next few weeks/months, or if Hollinger's rank was a more accurate rank of how good teams really were (and therefore how successful you'd expect them to be going forward.
For example, both Hollinger and win-loss record said New Orleans was the best team, so they get ranked as 1 in both lists. However, from November until now they had the 16th best record in the NBA, so for the span from 11/15 to 1/10, they get ranked 16. This means that both Hollinger and Win/Loss as predictors were 15 spots off from their Rank compared to the actual Rank in future games played. Contrast to Miami. Win-Loss record predicted they were the 10th or 11th best team, but they had the 1st best record in the league from 11/15 to 1/10, so the win/loss prediction was 9.5 ranking spots off. Hollinger had Miami pegged as #2, and they ended up #1, so he was 1 spot off there.
When I averaged that out for each team, Hollinger was off an average of 5.13 spots per team; win-loss off by 6.07 spots on average.
