I may be missing something here, but I imagine all our simulation programs keep track of points played in a game, as well as the winning post, et. al. Why not just crosstab the posts of winners by number of points played? Given a big sample, the first non-zero cell will be runouts. Those games could be flagged for comparison with non-runout games for question 2. As far as Q-3 goes, I don't think it is possible to identify runouts from the published data.
