Overview: Writing code to do a monte carlo simulation on different blackjack strategies. Version History 1.2 -> Blackjac.py is the file to run. EVPH is around .32% with single deck counting. 1.0 -> Blackjackr2.py is the real file. Others are tests/revisions. Examine runtest for parameters (count = True, countaccuracy = 100, etc.) Runlayer1() is the overarching function to call runtest. Not necessary. Currently produces roughly a 50% higher EV than not counting. Roughly +.3% per hand using one deck. Bugs: Counting accuracy is currently embedded within the runtest function. Need to extricate it out and add it to runexperiment function. Todo: Add splits, doubles, surrenders Make the runtest method be able to change counting accuracy. Possibly implement genetic algorithm to fine tune count. Plot/draw data from text files. Thoughts: Currently counting EV is nearly double of single deck non counting at around .32% EVPH According to Blackjackaudit by Deepnet Tech, 1-6 players has no effect on EV. There are arbitrary thresholds for truecount indices. This is problematic.