When you compare two systems or games in BJ you want to have all the gain show up in one statistic so you can make a linear comparison. What you have done has gain showing up in both reduced RoR and win-rate. As for CE/WR, it tells you that your approach to the first game is more optimal. But that doesn't make it better. c-SCORE tells you that, despite betting less optimally, the second approach is better when risk is factored in. Standard deviation is the stat you didn't include that would bring how these things relate into focus for you. The LS rule is quite valuable in reducing SD. It allows a counter to bet at a lower risk, which usually allows for larger bets at the same RoR.
Without all the necessary information to say for sure, it looks like you used the same ramp for each system when the optimal ramp with the LS rule would bet more aggressively. When making comparisons both systems should use their respective optimal bet ramps. RoR, spread and BR should be kept constant over both sims. This allows all gain to show up as an increase in win-rate so the gain can be compared directly. If you have a need to compare when the gain showing up in multiple places, c-SCORE allow for a which is better comparison.
I am not sure what the "1x" and "2x" mean in the rule sets for the games.
Thanks a lot, Three. I'm experimenting with using one betting strategy for various rules and various penetrations in shoe games. I understand the best way to compare games is if I keep my RoRs, BR, and bet spreads constant, but because of the nature of the experiment I can't keep my RoRs constant. So I wanted a way to compare the games with my non-optimal betting strategy with different RoRs and different c-SCOREs but using one value to compare. I wondered if there might be a meaningful way to combine c-SCORE and RoR.
1x means one hand, 2x means two hands.
Last edited by MercySakesAlive; 04-25-2018 at 11:00 PM.
I like to differ SCORE/RoR questions like this to Don. He is the man when it comes to these things. I know what I would do. One thing is to E-mail Don before I get into a lengthy analysis if there is something I am sketchy about. I have tried workarounds that didn't follow his primary recommendation. Usually he was right about needing a better way to analyze the problem. Sometimes approximation was good enough. We are all lucky to have him as a living resource.
"We are all lucky to have him as a living resource."
Uh-oh. Do you know something I don't know?! In any event, thanks for the kind words. (Below comments do not apply to you personally.)
And yes, a bit too much reinventing the world in these above posts. Either you are interested in just how much money you can make (e.v.), risk be damned, or you take a more rational approach and temper that number with the risk you need to take to attain your goal (s.d. or variance). The ratio of the two, e.v./s.d., was christened the Sharpe ratio by a famous economist who won the Nobel prize many years ago. SCORE simply squares the Sharpe ratio (Desirability Index) to allow for linear comparisons and to express a number that corresponds to an hourly win rate when certain assumptions are applied.
As with the I18, people always want to take a simple and elegant concept and look to make it more complicated and convoluted. Pity.
Don
Cool, I appreciate your help. If you're willing--- back to my SCORE confusion... hopefully, I won't cause you too much more headache but I think I'm having some basic concept issues regarding CVCX and SCORE. Now that I've reread Chapter 9, I understand SCORE is restrained by a 10k BR, optimal-bets, 13.5% ror, with Ill 18 and Fab 4, with a particular number of people at the table depending on deck numbers. Is it possible I misunderstood you when you... you said that CVCX always gives me SCORE and not c-SCORE. How can it be a SCORE and not a c-SCORE if I run a 3 person sim on a 6 deck or a use a 20k bankroll? Thanks for any help.
in cvcx. on a sim results page. in top bar. left side. click 'customize.' then click 'customize columns.' you can turn on or off "c-score". this is the same stat that is called "score" on the results page. I think this was the question you may have been asking.
edit: actually, I think I was not answering the question you were asking, NB10.
Last edited by MercySakesAlive; 04-26-2018 at 04:13 PM.
Bookmarks