I have no interest for trolling, but sometimes I post a new technique to seek comments from experts here. Often times, this leads to warnings me of trolling. Earlier, I posted my ASC technique saying that it may dictate playing one-hand or two-hand blackjack, but haven’t heard any reasonable comments on it, except Don’s firm no.
What do you think about my ASC technique?
I think your system is very good. It has a high PE and IC but low BE.
BC = 0.9152
IC = 0.9165
My recommendation is that instead of using the other secondary count, try to keep a side count of aces. In this way, you could increase the BE from 0.9152 to 0.9818!
Sincerely,
Cac
I already use the ace side count. I was just curious if a 5+3/7+1vs. X-1 sidecount would be of any value to the main playing count?
I was just wondering if you knew if it was theoretically possible to get your playing efficiency up to "as good as" 80 percent with this or any sidecount?
Last edited by Jack Jackson; 07-19-2022 at 11:40 PM.
http://bjstrat.net/cgi-bin/cdca.cgi
Download this analyzer and play around with side count numbers-
https://web.archive.org/web/20080630...ficien/eff.htm
"Don't Cast Your Pearls Before Swine" (Jesus)
In my opinion, the mental effort of maintaining that secondary count is unnecessary. Of the R22 main plays, the only one that is better correlated than the primary count, is 16vT.
In this case the correlation changes from 0.6892 to 0.7767. I don't think it's worth the effort.
On the other hand, the tags from the secondary count are already being counted in the primary count.
Regarding the other question, somewhere I read that theoretically you can't go beyond 70%. I think it was in TOB.
But, based on some research we're doing with Zenfighter, I think it's possible to go further. However, I don't know what that limit is.
PE is a parameter that nowadays needs to be updated. It was originally intended for SD and for 71 plays that do not include soft doublings or splits.
Sincerely,
Cac
Agreed, the concepts of Playing Efficiency and Betting Correlation are dated and new definitions must be used. One plausible approach is to compare simulation results using a given counting system to simulation results using perfect play combinatorial analysis where the exact deck composition is used to calculate the optimum playing strategy and pre-deal EV for betting.
Naturally, these sims take a really long time because a CA calculation is done at every step and one would need to resort to heavy parallel computing to run these sims. I was able to recently run an 80 million rounds sim for 6D shoe S17 DAS with 40 rounds per shoe(to eliminate cut card effect). I would like to do more work on this but preliminary results show an increase of 25% in SCORE compared to a Hi-Opt2 ASC. So we have about 75% PE assuming BC is one.
Chance favors the prepared mind
An excerpt from a book by John May and Frank Scoblet-
Super Counters Beyond Level Four
Popular systems are already close to the "ideal bet", so to be substantially better you need to break the 0.70 PE barrier. To do this, you must use a multi-parameter system. This means that you must carry more than one account at a time. Some commercial systems recommend using the Aces side count, as the Ace behaves like a small card when playing hands, and like a big card when determining correct bets.
But a focused mind can go further, at least in theory. Can a system of greater complexity push the boundaries further?
Bert Fristedt and David Heath published an article entitled "The Most Powerful Blackjack Strategy Devised" in Winning Magazine. This article suggests counting cards in three groups: 2, 3, 4, 5 were simultaneously counted against 6, 7, 8, 9, and 10s. The player had to use a 2D strategy chart in order to play using Heath's regular betting account. The system's playing efficiency is .75, which looks impressive, until you consider it's only a tenth of one percent advantage! The Heath system outperforms the best commercial systems in shoe games.
Peter Griffin approached the problem from a different angle. He proposed to improve the Hi-Opt I system with no less than a five-way count. He presented his ideas at the First and Second International Gambling Conferences in June 1974-75. Griffin's method results in an efficiency of 0.89 percent in the game. Extensive simulations by John Gwynn showed that Griffin's strategy did achieve 0.2 percent over existing Aces sideways systems. But again, virtually nothing has improved in multi-deck games. And the tables that require you to memorize the Griffin method are too long, which adds to the difficulty of remembering almost half of the deck of cards in the actual game.
David Sklansky put forward another suggestion in "Gambling Times", August 1977 in his article "Getting the Best Of It: The Best Of It: The Key Card Concept—An Extra Edge At The Blackjack Table". Sklansky's approach is less systematic, as he suggested memorizing a few key rules using knowledge of the cards you see dealt in a one-deck game. However, as Sklansky later clarified, the concept of a keycard is not an actual system, only a set of general guidelines that could be included in the system.
Surprisingly, although the development of computers has made system modeling much easier, there has never been any outstanding work published on multiparameter systems.
Last edited by Gramazeka; 07-21-2022 at 09:08 PM.
"Don't Cast Your Pearls Before Swine" (Jesus)
"Don't Cast Your Pearls Before Swine" (Jesus)
In shoes the use of CA is practically impossible and forget completely if the system to be used is higher than level 2. However, an alternative is possible using "perfect EORs".
I know it's not the same and we probably won't get the most out of it. But it is somewhat faster.
Now, there is something I want to clarify regarding what I call "perfect EORs". The idea is to create a set of EORs for each decision instead of using the difference between
two types of plays (hit vs stand, doubling vs hitting, etc.). For example, 11vT is going to be made up of three different EORs instead of just one:
11vT (Hit)
11vT (Stand)
11vT (Double)
For 88vT we will have four EORs:
88vT (Hit)
88vT (Stand)
88vT (Double)
88vT (Split)
Do you understand what I'm pointing to? Every time you have to evaluate a play, you will have to decide which one has the highest expected value according to the calculated EORs.
Of course, this implies the use of many tables loaded into memory.
Combinatorial analysis is still necessary, but only to calculate new EORs.
Sincerely,
Cac
Bookmarks