No, the red/black count was some experimental code that I didn't use as generation of the graph data. The graph should be read "at 50% penetration, 20% of all decks have a positive EV for RM bet". That is just raw statistics with perfect knowledge of cards dealt. A count is used to recognize...