[Degenerate]

I want to work out seedings & potential matchups for the NBA playoffs, using an estimate of how many wins i think each team will get, does anyone know of anyway to be able to do this?

I have been trying to use binomdist function in excel to do this, but can't figure out how to put it all together

These are my win estimates for the East (They probably aren't that good)

Boston 63.47
Detroit 59.4849
Orlando 52.3134
Toronto 47.0697
Cleveland 46.6865
Washington 38.536
Atlanta 37.0533
Philadelphia 35.2909
Indiana 34.1956
New Jersey 34.3223
Chicago 32.7354
Milwaukee 31.5777
Charlotte 27.3574
New York 26.6976
Miami 15.8829



Any help would be appreciated

[pat venditto]

cut everyone off accept the top 4 there. cleveland has no shot

[Data]

Here is a formula that gives a good estimate:
SeasonWins%=0.5+2.7*(OffensiveEfficiency-DefensiveEfficiency)

Or, you can just use these estimates:
http://www.knickerblogger.net/stats/2008/

[Ganchrow]

I want to work out seedings & potential matchups for the NBA playoffs, using an estimate of how many wins i think each team will get, does anyone know of anyway to be able to do this?

-snip-

The quickest, easiest method is probably just to run some Monte Carlo simulations and use outcome frequency to estimate probability. I posted a basic Monte Carlo simulator written in Perl about a week ago which could be repurposed for your needs.

You will need to decide exactly what you're looking to estimate (for example, is it the probability of team A finishing as the nth seed or the joint probability of teams A-H finishing as the 1-8 seeds, respectively?) You'll also need to determine to what extent you're willing to make simplifying assumptions to speed up the simulations process. For example, would you be comfortable assuming a single season-wide win probability for each team or do you want to specify multiple win probabilities for each team based on (say) home/away status, or perhaps even more complicated criteria?