Run new mixed effects model (pRisky ~ goalCondition * overGoal + points + (1|id/game)) and create plots with points earned on the x-axis and pRisky on y axis (I think here I will create bins of points to then actually get a probability instead of 0 and 1 for the pRisky).