Re: Moving Up In Stakes
Okay, enough. Aside from analyzing hand histories or speculating about leaks etc I wonder whether we can make some conclusions from Guthrie's graph (page 2 of the posts).
What I propose is a sequential correlation test on the data set. For a non-tiltable player (e.g. robot) the distribution of results for one day should look exactly like any other day. However if a player (e.g. me) often goes on tilt then today's session is likely to be poor if I did poorly yesterday. Conversely today's results might be more likely to be good if I did well yesterday.
I have a rough idea of how to approach it mathematically, but maybe a math/stats guy could chime in here? It seems to me such a correlation is likely to exist for most human players, and may add significantly to the risk of ruin, or at least extend really bad streaks.
|