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.