I'm new to statistics.

  1. What type of model should be used to predict a basketball team's winning percentage? I am currently using a multiple linear model, but the residual vs. predicted plot shows my data points clustered in a circle. I standardized my explanatory variables since they were either percentages or totals.

  2. Also, what model should be used for the Pythagorean Expectation formula? I would like to see what that would look like in R.


  • $\begingroup$ I think for starters you may want to consider a logistic or a probit regression where you dependent variable is a basketball team's winning percentage. The Wikipedia articles on both subjects are quite good starting points. $\endgroup$ – usεr11852 says Reinstate Monic Jan 12 '14 at 14:03
  • $\begingroup$ Are you sure? A win percentage is continuous and can take on more than 2 different values. For both of those models, it suggests usage when dependent variables are binary, e.g. take on a value of 0 or 1. $\endgroup$ – user2205916 Jan 12 '14 at 14:21
  • $\begingroup$ Even though win percentage is continous, your dependant variable (did they win or not) is not - it can be either 0 (lose) or 1 (win) (what you do with ties is up to you). Using a logistic regression or linear probability model would probably be your best bet. Then predict from the estimated model. $\endgroup$ – ltronneberg Jan 12 '14 at 16:03
  • $\begingroup$ @ltronneberg The dV will be a proportion if you have a time series or some other aggregated form of data. In that case you could take the log of the dV and then use a linear model instead of logit or probit. $\endgroup$ – tomka Jan 12 '14 at 16:05
  • $\begingroup$ @Itronneberg Thanks for that explanation. I will go with logistic. $\endgroup$ – user2205916 Jan 12 '14 at 16:21

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.