# Problem with Sales Regression Residuals

I'm trying to build a model for the ticket sales for different sporting events over a period of 30 days before the game to the day of the game. The problem that I'm having is that I can't seem to fit the right model to it. I've tried negative binomial, linear, poisson, but nothing seems to fit right. A problem that tends to occur is I get a residual plot that looks like the one below. Does anybody have any suggestions for things that I can try?

Residuals for the negative binomial:

• Ticket sales are probably all nonnegative, and they may be integer, so you should expect residuals that are skewed (simulate negbin data, fit them with the correct model and look at the residuals). That does not need to mean that your models are incorrect. – Stephan Kolassa Jun 26 '14 at 16:10
• You might also try plotting the residuals against each of the predictors. It might give you some idea where there is homogeneity. Doing so could suggest other models additions to the model, or where transformations are necessary. – Chris Jun 26 '14 at 16:31
• Just to give you an idea, I added what my residual plot was for the negative binomial. – user3704120 Jun 26 '14 at 18:08
• What are SINGLE.SALES? Are you modeling the number of tickets sold on a given day, the amount of sales in \$s, or something else? Also, how is it that your negative binomial model predicts$\approx -40\$? There is something wrong with that. – gung Jun 26 '14 at 18:15
• Sorry, single sales is just differentiating it from group sales or season tickets. It's number of tickets. – user3704120 Jun 26 '14 at 18:19