I have data from a league over a season. The season was conducted as a round robin so everyone played each other. Each game was a best of 5 so at the least a game produced 3 matches and at the most produced 5. During each match 25 features were collected for each team. There are teams who lost all their matches, teams that won all their matches, and teams that are mixed. What I want to do is find out if there is a statistical difference between a certain feature and how it is correlated with other features in a winning match versus a losing match. I have thought about treating it like two groups (ie male/female) or like pre-post but it doesn't seem to really fit either of those because some teams appear in both groups but not all. Would enjoy any advice!

  • $\begingroup$ If you are interested only in whether a team lost or won then you assign a binary response for this, and afterwards you perform logistic regression on all the other variables. $\endgroup$ – user2974951 Nov 30 '18 at 13:30
  • $\begingroup$ I've already set a logistic regression model for prediction what I'm asking is for help analyzing statistical significant differences between winning and losing $\endgroup$ – JJ Stamp Nov 30 '18 at 19:24
  • $\begingroup$ The output of the logistic model will tell you that, that is which variables are significant and in what way they impact the classes (won | lost), that is do they increase the odds of winning. $\endgroup$ – user2974951 Dec 1 '18 at 8:48

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.