NBA Preseason - Season Correlation

[Latrinsorm]

I was glancing over the preseason standings today and noticed that the Lakers had an average margin of victory of -14.2, and I wondered: does preseason margin of victory give us any information about regular season margin of victory? So I took the last five full years of data and as it turns out, yes!

http://img.photobucket.com/albums/v456/johnnyoldschool/NBAPreseasonCorrelation_zps04f220f4.png

More broadly, we can look at four factors...
1. If a team with a positive MoV in the preseason had one in the regular season,
2. the same for negative,
3. if a team with a positive MoV in the regular season made the playoffs, and
4. the same for negative missing the playoffs.
These factors over this sample turn out to be 63% ± 10%, 67% ± 12%, 95% ± 5%, 87% ± 8%.

Thus...
If your team finishes the preseason with a positive MoV, you'd have a 60% ± 10% chance of making the playoffs.
If your team finishes the preseason with a negative MoV, you'd have a 58% ± 12% chance of missing the playoffs.

These numbers are indistinguishable from chance, but because each individual factor is distinguishable from chance I anticipate that with more data a better-than-chance outcome will arise. Stay tuned! :)

[Latrinsorm]

Okay, espn.com goes out to the 2002 season and with that data the values become...
63% ± 7%, 66% ± 8%, 95% ± 3%, 86% ± 5%
If your team finishes the preseason with a positive MoV, you'd have a 60% ± 7% chance of making the playoffs.
If your team finishes the preseason with a negative MoV, you'd have a 57% ± 7% chance of missing the playoffs.

I also thought it would help to visualize if we look at it a couple other ways. First let's have a double graph, the number of teams that missed and made the playoffs by preseason MoV...

http://img.photobucket.com/albums/v456/johnnyoldschool/NBAPreseasonCorrelation2_zps717b8a58.png

...and the % of teams that made the playoffs by preseason MoV...

http://img.photobucket.com/albums/v456/johnnyoldschool/NBAPreseasonCorrelation3_zpse1a4d52f.png

Pretty sweet, huh? Each point of preseason margin of victory is worth 3% more chance to make the playoffs! The bins outside the range shown are 31% for all less than 6 points and 68% for all more than 6 points, which sorta fits, but it's only 15/33 and 17/8 which is pretty small for such a range. (Plus the regression obviously can't be linear forever, otherwise a +15 margin would mean 100% chance and -19 margin would mean 0% chance.) If we look at the teams currently outside of the sixes range this year, we have...

IN: Boston, Golden State, Utah, New Orleans, Minnesota, Phoenix, Portland
OUT: Lakers, Clippers, Oklahoma City

I would be very surprised if all these teams ended up as predicted, but at 70% it means 3 teams should go the other way: Clips, Thunder, Utah. Boom! Congrats on the 8 seed, Boston! And there may yet be movement in the last couple games. Check out the predictions from last year with errors bolded...

IN: Chicago, Toronto, Brooklyn, Charlotte, New Orleans, Phoenix, Houston, Minnesota
OUT: New York, Philadelphia, Orlando, Milwaukee, Atlanta, Denver, Utah

11 for 15, 73%! :D So there you go.