Optimal Betting/Sim Information Request

[ohbehave]

"11% is not enough"

Is that what you said ?

I knew I'd get some heat for that statement.:grin: But for my bet ramp, etc it amounts to a difference of $300 in 10,000 rounds.

I would make at least $300 in mistakes with ZEN. So, for me at least for now, its not enough of a difference considering the advantages of how simple OPP is.

Just offering an option, not saying everyone should go with OPP.

 

[QFIT]

I knew I'd get some heat for that statement.:grin: But for my bet ramp, etc it amounts to a difference of $300 in 10,000 rounds.

I would make at least $300 in mistakes with ZEN. So, for me at least for now, its not enough of a difference considering the advantages of how simple OPP is.

Just offering an option, not saying everyone should go with OPP.

OPP has a SCORE of about 35% of KO Preferred. About 50% of KO with no indexes. It is arguable weather KO-Rookie or OPP is simpler.

 

[ohbehave]

I still don't quite understand SCORE.

 

[gibsonlp33stl]

What's a good book to learn the Zen counting method?

 

[QFIT]

I still don't quite understand SCORE.

Essentially, it is the hourly win rate with a $10,000 bankroll, optimal betting and a 13.5% risk of ruin. But you can just look at it as a relative measure of the win rate given standardized circumstances.

 

[ohbehave]

OK gotcha. For a SCORE comparison in my sims (1 billion rounds) I get:

Win Rate(per round)/SCORE
ZEN/.13/13.1
HiLo/.13/13.3
OPP/.11/11.3

Above is for a DD game. I did not do a shoe sim for all 3 but for OPP & HiLo I get:

Win Rate(per round)/SCORE
HiLo/.20/20.9
OPP/.09/9.1

OPP really sucks at shoe games but not so much for pitch games.

 

[ohbehave]

I forgot to mention the bet spread in the DD game is 5:1 and for 6 decker 10:1.

 

[QFIT]

OK gotcha. For a SCORE comparison in my sims (1 billion rounds) I get:

Win Rate(per round)/SCORE
ZEN/.13/13.1
HiLo/.13/13.3
OPP/.11/11.3

Above is for a DD game. I did not do a shoe sim for all 3 but for OPP & HiLo I get:

Win Rate(per round)/SCORE
HiLo/.20/20.9
OPP/.09/9.1

OPP really sucks at shoe games but not so much for pitch games.

Something's quite wrong. OPP is nowhere near that close to HiLo or ZEN and is worse at DD than 6D. You can see SCOREs at http://www.card-counting.com/cvcxonlineviewer.htm. OPP is the last strategy in the list.

Card-Counting.com​

 

[ohbehave]

Something's quite wrong. OPP is nowhere near that close to HiLo or ZEN and is worse at DD than 6D. You can see SCOREs at http://www.card-counting.com/cvcxonlineviewer.htm. OPP is the last strategy in the list.

Card-Counting.com​

I think I have to respectfully disagree. DD and a 5:1 spread gets a c-SCORE of 2.95. Is that correct?

I'm using the OPP version of PowerSim directly from Carlos.

In both live play and computer-simulated play OPP far outperforms at DD.

 

[QFIT]

I think I have to respectfully disagree. DD and a 5:1 spread gets a c-SCORE of 2.95. Is that correct?

I'm using the OPP version of PowerSim directly from Carlos.

In both live play and computer-simulated play OPP far outperforms at DD.

Well that explains that. Why would you use a program from the system author with no reputation for sims?

We are talking about relative performance. HiLo Illustrious 18 has a c-SCORE of 14. 4.7 times that of OPP. At six decks, HiLo is vastly better, but not 4.7 times better. OPP performs badly under all circumstances. But much worse in pitch games. Read Theory of Blackjack to understand why.

 

[ohbehave]

I know nothing about sims myself but PowerSim is open-source. Have you checked out Carlos' version for flaws in the code?

A question... In OPP the IRC is +6. How does the count on your viewer relate to the running count in OPP?

 

[QFIT]

I know nothing about sims myself but PowerSim is open-source. Have you checked out Carlos' version for flaws in the code?

A question... In OPP the IRC is +6. How does the count on your viewer relate to the running count in OPP?

PowerSim cannot even calc True Counts correctly. It is the worse code I have seen in decades. No joke. It is 1950s style code. (The author jokes about modern coding techniques.) This is not an exaggeration. I asked one of the other BJ sim authors to look at the code that is a computer science prof. He said it was too painful to read. Unfortunate as it has been pushed on students as a good way to start writing sim code.

 

[ohbehave]

How does your software determine the OPP count? And how does the count shown in the table on your viewer relate to the OPP running count?

 

[ohbehave]

"The true counts employed by PowerSim are "floored," which means the running count is divided by the number of unseen cards remaining, then multiplied by 52 to get an index scaled to full deck, rather than half or quarter deck. Then the INT math function is used to get an integral index, which in Basic (as opposed to c++) is never higher than the exact fractional index.

Examples of floored TCs:
* RC = 20, unseen = 64, TC = INT(20/64 x 52) = INT(16.25) = 16 (nice to dream, isn't it?)
* RC = 20, unseen = 65, TC = INT(20/65 x 52) = INT(16) = 16
* RC = 20, unseen = 66, TC = INT(20/66 x 52) = INT(15.7575..) = 15
* RC = -7, unseen = 122, TC = INT(-7/122 x 52) = INT (-2.9836) = -3
* RC = -7, unseen = 118, TC = INT(-7/118 x 52) = INT(-3.0847) = -4

Choosing accuracy of decks remaining to nearest quarter deck, or nearest half deck was not done. Instead the fractional TC is precise and then floored. Again here, nearest 1/x deck accuracy could be written in by a programmer, and I would be glad to help, if this is desired.

Doing the entire TC calculation "in line," rather than splitting it up into parts, guarantees the accuracy of this flooring technique, under the IEEE 754 floating point standard employed by XBasic. Another advantage is simplicity and readability of the source. A disadvantage is speed in execution."

This is quoted from the Powersim instructions. Are you saying this is not the correct way to calculate true count?

 

[johndoe]

I don't understand why Powersim gets such a bad rap. It's completely free, and does exactly what it purports to do. The simulation results have been thoroughly verified as being just as accurate as any other sim.

If the code is ugly, it's only a big deal if you wanted to modify it for your own purposes. But I really didn't think it was that horrible, as far as Basic goes.

Sure, it's not a polished professional package, and it's missing some features I'd like, but it does work.

 

[QFIT]

How does your software determine the OPP count? And how does the count shown in the table on your viewer relate to the OPP running count?

I really don't remember what IRC I used. Has no effect on the SCORE.

 

[QFIT]

This is quoted from the Powersim instructions. Are you saying this is not the correct way to calculate true count?

Last I looked, PS TC routine had one line of code. CVData has about 1,000 lines of code for TC calc. Humans do not take the precise number of unseen cards and divide into RC to get a precise TC. This overestimates advantage. Humans use many different methods. I discuss some of this at http://www.qfit.com/CalculatingTrueCounts.htm.

 

[QFIT]

I don't understand why Powersim gets such a bad rap. It's completely free, and does exactly what it purports to do. The simulation results have been thoroughly verified as being just as accurate as any other sim.

I'm sorry, this is simply not true. After I pointed out inadequacies and made helpful suggestions, including donating code, ETF wrote a long "article" claiming this. The "article" mentions my name 17 times. It is filled with misquotes, fabrications, slurs, and basic misunderstandings of both Blackjack and technology. He then placed this article in a place where I am not allowed to respond. He then repeatedly claimed I was allowed to respond, although I am barred. He has also, repeatedly, claimed that CVData lacks features it has had for many years where I cannot respond.

If the code is ugly, it's only a big deal if you wanted to modify it for your own purposes. But I really didn't think it was that horrible, as far as Basic goes.

Forget modern techniques like Object Oriented Programming. PS is not even procedural. Proc code was invented in the early 60's. That's over half a century ago. This is very important. The point of "Open Source" is ease in modification. Non-procedural code is notoriously difficult to modify without introduction of errors. Code written in the 50s was all rewritten due to these problems. PS itself does almost nothing. Almost none of the CVData features at http://www.qfit.com/blackjack-simulator.htm exist. That means it must be modified to come up with any answers we don't already have - the purported purpose of PS. The claim is that you can quickly modify it. For example, ETF claims it would be trivial to add index generation. No, you would have to start from scratch. Index gen is far more difficult than basic sims and the code is very different.

I posted the code for a sample, bare-bones simulator using procedural code years ago on the old bj21 free boards. (But those boards were destroyed by one of ETF's friends.) Afraid I have since lost the code.

 

[PrinceDragon]

Pelerus

Here's the sim,
for 5K BR,ROR is 13.9%, it reduce to 1.9% with 10K.
P.D.

 

[ohbehave]

I really don't remember what IRC I used. Has no effect on the SCORE.

What I was trying to ask is does your software correctly determine the OPP count. As you know in OPP the count has to be reset following each round for the number of players in the round. Does your system do that?