…are included. I use a model to combine polls. It tries to model pollsters' different house effects, on the assumption that these on average cancel out. The model has the race like this: