Quote Originally Posted by bjanalyst View Post
CC showed HL w AA78mTc and 5m6c beat HO2 w ASC with PE but HL was 2% below HO2 - 2*(Adef) for BC and when HL + (1/3)*(5m6c) was used SCORE gap was reduced but HL system still lost because BC was still over 1% below HO2 w ASC.
No. You want to understand how accurate something is look at the standard deviation of the actual data points that a sim collects around the data points average for each decision bin. Hiopt2/ASC wins because it has a tight standard deviation around its averages. CC has no way of indicating what matters. It is just a starting point for making a system. Then you try to tighten all the decision bell curves around their averages to get a low SD for the betting or playing decision. BC is a correlation. It has nothing to do with accuracy. You want to know why your system won't beat Hiopt2/ASC look at these graphs.

Post #17:
https://www.blackjacktheforum.com/sh...-at-tc-1/page2

Post #30:
https://www.blackjacktheforum.com/sh...-at-tc-1/page3

Look at the scale at the bottom of each graphs. The Hiopt2/ASC graph shows accuracy and is very tight around the average. The Hilo graph is over twice as wide with much longer tails but the BC isn't very different. That is the difference between betting accuracy and BC. The more accurate bets also allow you to bet more at the same RoR when everything else is kept equal. For example a Hilo optimal bet S17, DAS, LS, 6 deck/1 cut off with the SCORE standards for everything else other than bets is 2x1 to 2x130 and a c-SCORE of 73.90, while Hiopt2/ASC 2x14 to 2x140 and a c-SCORE of 85.75. Now all of these are for full indicesKO preferred gets 2x12 to 2x120 and a SCORE of 62.33.