Adding AA78mTc side count to High Low

[seriousplayer]

While one must be impressed by his mastery of the math, someone should do an intervention and reacquiant him with the real world.

You might be but I am not. What is bjanalyst doing here when he is so good in math?? He should probably go discover and develop fundamental ideas in mathematics or go solve the Riemann hypothesis. Better way to make 1 million dollars than to play blackjack. Most mathematicians are not that influential in Blackjack.

I am not trying to nit pick, but BC is a correlation not an efficiency. Mathematically there is a big difference.

Yes, Peter Griffin in “The Theory of Blackjack” already said that betting correlation is not the same as betting efficiency but he never gave ways to calculate betting efficiency nor did he give a definition for defining betting efficiency. It is easy to confuse betting efficiency with betting correlation because in betting correlation it measures how effective the count is with respect to your bet sizing.

[bjanalyst]

I generally like to know what the improvement is for each level of extra effort, so I would suggest simulating all of the above.

It's the holiday season and I'm busy spending time with family and friends. The planning of the simulations will begin in the new year.

First I would like to thank you very much for volunteering to do this simulation.

I would like to make this work as simple as possible for you so you do the least amount of work to prove HL with AA78mTc.

First I choose HL with AA78mTc rather than KO with AA89mTc which is what I use. HL with AA78mTc is balanced and you already have HL sim programs and simulating HL with AA78mTc is good enough to show what I did was or was not correct.

Trying to simulate KO with AA89mTc would require too much extra work and the chance for errors since additional work would need to be done because of the unbalanced counts and a different primacy count other than your canned HL program. I want to keep the changes to your canned HL program as simple as possible.

Thus I just want HL with AA78mTc simulated. If my calculations for HL with AA78mTc are shown to be correct by simulations then my calculations of KO with AA89mTc, which were done using the same program and technique as HL with AA78mTc, are also correct.

I would like to simulate the S17, DAS, no LS six deck game, five decks dealt. So no need to do any late surrender calculations and make sure that dealer stands on soft 17 as that is what I did my calculations on.

I would chose the middle option - make changes to your HL program only for the 12 changes in the illustrious 18 that AA78mTc helps with. The other 6 Illustrious 18 changes, AA78mTc does not help with.

I will attach the Illustrious 18 for HL, HL with AA78mTc and HO2 with Adef (deficiency of Aces) to this post.

So I am assuming that you already have canned programs for HL, HO2 and HO2 with ASC for six decks, 5 decks dealt, S17, DAS, no LS game.

If you can simulate it for a back counted game where you leave the table when the betting running count (brc) is negative, that would be even better. For HL, brc = HL. For HO2 with ASC, brc = HO2 - 2*(Adef). Betting correlation of HL for S17, DAS, LS is 96.5% and for HO2 with ASC is 97.9% as shown below so HO2 with ASC definitely has better betting than HL with AA78mTc since AA78mTc does nothing to help with HL betting.

brc = betting running count = HO2 - 2*(Adef) where Adef = deficiency of Aces remaining = Ap - 4*dp = (12/13)*Ap - (1/13)*(23456789Tp) which gives BE = 97.9% for S17, DAS, LS where HO2 = (2367p) + 2*(45p) - 2*(Tp) so brc = HO2 - 2*(Adef) = (1 + 2/13)*(2367p) + (2 + 2/13)*(45p) + (2/13)*(89p) - (2 - 2/13)*Tp - (24/13)*(Ap).

But don't go crazy. The important point is that all variables for all simulations must be identical except for the counts.

So here is what I would like you to do if you can.

(1) Make a copy of your HL program and edit it. Set up a variable to keep track of the AA78mTc. Then use AA78mTc only for the 12 Illustrious 18 strategy changes listed in the attachment to this post. For all other strategy changes and for betting make no changes to the HL program. That is use whatever your original HL program has for indices and whatever routine you already have set up for HL betting. HL betting and HL indices and strategy changes for all situation other than the 12 Illustrious 18 to be changed for use with the AA78mTc needs to be unchanged - that is make changes to the HL program ONLY for the 12 illustrious 18 situations using AA78mTc and leave the rest of your canned HL program unchanged.

(2) Run the original HL program and run your modified HL program with the 12 Illustrious 18 strategy changes and compare.

(3) Run your HO2 program with no ASC and see what the results are.

(4) Now comes HO2 with ASC. Make a copy of that program. Then to keep things even, use ASC only for the Illustrious 18 since you are not making any changes to the HL program for AA78mTc for other than the Illustrious 18. If would be an unfair comparison if I limited the HL with AA78mTc changes to just the Illustrious 18 but you used the HO2 with ASC for all situations. Ceteris Paribus - everything else must be equal except for the variable we are testing which is the various counts.

So here is the first attachment, that I did before, with the HL and HL with AA78mTc for the Illustrious 18.

HL vs HL + k(AA78mTc) SORTED.jpg

So you can see the 12 strategy changes that I would ilke you to make to the HL using the AA78mTc. Make only these 12 strategy changes. Leave all other decisions and betting unchanged using whatever your original HL program had for all other strategy changes and for betting.

So run this modified HL program with AA78mTc for these 12 situations for S17, DAS, no LS six decks, five decks dealt game and produce and document your results.

For HL and HO2 with no side counts, just run the original programs as they are with no changes for S17, DAS, no LS six decks, five decks dealt game.

If you can make the programs to simulate back counting and leaving the table (betting zero) whenever the count is negative that would be good. The important thing is to do the same with all simulations - the only thing that varies between simulations is the count - everything else needs to be identical.

When you email me I can email you my Excel file program that calculates values of k and indices, which Excel file ETFAN reviewed circa 2011.

So attached below is HL with AA78mTc compared to HO2 with ASC for the Illustrious 18. Please note that I calculated the HO2 + k*(Adef) indices and values of k from my LSL program. I did not have access to HO2 and ASC indices and modifications. Your canned HO2 with ASC must have these HO2 indices and values of k. Please review the HO2 indices and values of k that I listed in my Illustrious 18 below. If they agree with what you have then that is another check that I am on the right track.

HL + k(AA78mTc) vs HO2 + k(Adef) SORTED.jpg

Enjoy the holidays and the New Year for now. I can go over more details and you can ask me additional questions about the simulations next year.

Again, I cannot thank you enough for agreeing to do these simulations.

Some users on this site have accused me of being a troll, a charlatan, spreading fairy tales and pedaling useless systems. They are demanding simulations which they are hoping will prove me wrong.

Thanks again. I will be looking for your emails next year.

[Three]

Gronbog will be very thorough. That is how he does things. If you want him interested in the results let him do his thing. You couldn't ask for anyone better to do this for you.

[seriousplayer]

I will attach the Illustrious 18 for HL, HL with AA78mTc and HO2 with Adef (deficiency of Aces) to this post.

You won't get the full power of Hi-OPT II with side count of Aces using only the Illustrious 18 or the 22 indices + fab surrender. The full power is in Hi-OPT II with ASC full indices. Maybe Three would agree with me on this one.

Run your HO2 program with no ASC and see what the results are.
So run this modified HL program with AA78mTc for these 12 situations forS17, DAS, no LS six decks, 5 decks dealt game and produce and document your results.

That would be unfair since Hi-OPT II with no ASC would under perform.

For HL and HO2 with no side counts, just run the original programs you have for S17, DAS, no LS six decks, 5 decks dealt game and produce and document your results.

What is the logic here to compare Hi-OPT II with no side count??? Norm, already did a canned sim for Hi-lo vs Hi-opt II without ASC. This is not including the AA87mTc.

Hi-OPT II with no ASC vs Hi-lo.jpg

Finally lets consider a HO2 with ASC program that you have. I need you to make a copy of HO2 with ASC and modify that program to take out all strategy changes that the ASC is used for except for any strategy changes for the Illustrious 18 and of course use ASC to help with betting.

Same as the post above you are making Hi-OPT II under perform. Hi-OPT II's full power is in ASC and full indices. Hope you are not trying to skew the simulation.

[bjanalyst]

Gronbog will email me next year and we can discuss how to proceed.

I want to make Gronbog's work as simple as possible. I want him to make as few changes to his HL program as possible. That is why I suggested changing only the 12 Illustrious 18 that AA78mTc helps the HL with. But if I limit AA78mTc help to HL only with the Illustrious 18, then to be fair, I must also limit ASC help to HO2 to only the Illustrious 18. This is what I meant by Ceteris Paribus - everything else except the variable being tested must be equal. So if I am considering side count help only with the Illustrious 18 for HL then I must also limit side count help for ASC to HO2 to the Illustrious 18.

I also suggested simulating back counting the game by having the programs betting zero whenever the betting running count is negative.

I do not want to drive Gronberg crazy in many HL indices with the AA78mTc that are not used very often and will have minimal impact especially if I am simulating back counting the six deck, five deck dealt game, S17, DAS, no LS. If I limit strategy changes to the Illustrious 18 to HL with AA78mTc then I must also limit strategy changes to the Illustrious 18 for HO2 with ASC. Doing otherwise would skew the results in favor or the HO2 and since I would like to simulate a back counted game, many of the other HO2 changes with the ASC would not even be done since HO2 brc would be negative and so the bet would be zero.

I will review what Gronberg wants to do next year. We will reach a consensus and whatever parameters he decides to use in his simulations, I am sure he will reveal his parameters and limitations in his results. If he decides to use only the Illustrious 18 for side count help with HL and HO2 then he will mention that in his simulation results.

[bjanalyst]

What is the logic here to compare Hi-OPT II with no side count??? Norm, already did a canned sim for Hi-lo vs Hi-opt II without ASC. This is not including the AA87mTc.

The HL, HO2 and HO2 with ASC programs already exist and are canned. I see that Norm already did a sim on HL vs HO2 with no side counts. But what I would like to see is a comparison of HL with AA78mTc side count against both HO2 with no side counts and HO2 with ASC. HL underperforms HO2 with no side counts. When AA78mTc is added to HL it should then outperform HO2 with no side counts. However comparing with HO2 with ASC is what we are really interested in but throw in HO2 with no side counts just for comparison.

So my guess is that you will see HL at the bottom, then HO2 with no side counts, and the HL with AA78mTc and HO2 with ASC. And I further guess that HL with AA78mTc will be close to HO2 with ASC as I believe that they are both about equal.

Again, I will discuss all of this with Gronberg and then let him decide how to proceed.

[bjanalyst]

Lemme jump in concerning the secondary 4579SC: If one is to do this, then why, oh why, would you use High Low? Simplify your system to a similar system!

High Low[A-T]: {-1, 1, 1, 1, 1, 1, 0, 0, 0, -1}tag
4579SC[A-T]: {0, 0, 0, 1, 1, 0, -1, -1, 0}

Combined[A-T]: {-1, 1 , 1, 2, 2, 1, -1, 0, -1, -1}: 0

While the above is balanced, I would recommend going to a better Level 2 system:

Highly Optimal 2[A-T]: {0, 1, 1, 2, 2, 1, 1, 0 , 0, -2} (use an ASC or a BASC to compute betting)
Zen[A-T]: {-1, 1, 1, 2, 2, 2, 1, 0, 0, -2} (ASC to adjust for Ins and playing in 1/2D games, +3D use Zen only)

Think of brevity and simplicity.

You cannot use 45m79c with the HL for betting. 45m79c is used only with the KO for betting. That is because the KO gives the 7 a tag of +1. The 45m79c give the 7 a tag of -1. So brc = KO + (1/2)*(45m79c) give the 7 a tag of +(1/2) which is exactly what is needed for betting. If you added 45m79c to the HL and used HL + (1/2)*(45m79c) then the 7 would have a tag value of -(1/2) -- a disaster!

Please look at attached KO + (1/2)*(45m79c) file I attached earlier and you will see. And by the way, I changed my mind and suggest keeping just the 5m7c instead of the more complicated 45m79c. KO + (1/2)*(5m7c) had BC of 99.0% which is only 0.6% below the 99.6% BC of KO + (1/2)*(45m79c). 5m7c also has slightly less help with playing strategy changes than the 45m79c does but the value is that 5m7c is much simpler to keep. If you are keeping only one side count, then not a big deal keeping 45m79c. But if you are keeping KO with AA89mTc and want to add a 2nd side count for betting then definitely use the much simpler 5m7c.

Take a look at the last column, brc = KO + (1/2)*(45m79c) in the chart below and you will see that the 7 has a tag value of +(1/2) in that brc.

Betting Efficiency.jpg

[refinery]

I'd like to remind everyone how easy this is.

[Freightman]

I'd like to remind everyone how easy this is.

I lack patience, citing post 160 as an example. Not withstanding, PE could be enhanced by incorporation of the FBM ASC.

[Midwest Player]

Damn, what a long thread. I have to admit, I have not been following this thread very closely. I see a very nice presentation with very nice charts, but the bottom line is after reading the very first post this is way too difficult for actual casino play. If you want something stronger than Hi-Lo then move up to a level 2 or level 3 count. I'm a strong believer in keeping only one count.

I also think stacking chips after each round is a very bad idea. I think this is a sure tell.

PS. I'm curious what kind of results bjanalyst has in actual casino play. For some reason, I don't think he has a lot of experience in actual casino play, however he seems like a very intelligent person.

[lij45o6]

You cannot use 45m79c with the HL for betting. 45m79c is used only with the KO for betting. That is because the KO gives the 7 a tag of +1. The 45m79c give the 7 a tag of -1. So brc = KO + (1/2)*(45m79c) give the 7 a tag of +(1/2) which is exactly what is needed for betting. If you added 45m79c to the HL and used HL + (1/2)*(45m79c) then the 7 would have a tag value of -(1/2) -- a disaster!

Please look at attached KO + (1/2)*(45m79c) file I attached earlier and you will see. And by the way, I changed my mind and suggest keeping just the 5m7c instead of the more complicated 45m79c. KO + (1/2)*(5m7c) had BC of 99.0% which is only 0.6% below the 99.6% BC of KO + (1/2)*(45m79c). 5m7c also has slightly less help with playing strategy changes than the 45m79c does but the value is that 5m7c is much simpler to keep. If you are keeping only one side count, then not a big deal keeping 45m79c. But if you are keeping KO with AA89mTc and want to add a 2nd side count for betting then definitely use the much simpler 5m7c.

Take a look at the last column, brc = KO + (1/2)*(45m79c) in the chart below and you will see that the 7 has a tag value of +(1/2) in that brc.

Betting Efficiency.jpg

Sigh!

You are making this more complex than it needs to be. As another poster stated, you are taking systems that are meant to be simple in execution and complicating it with a convoluted system you though up and making things more difficult for practitioners.

Another issue you have is the verbose application of analysing situations. Three has a better approach than what you are doing. (No one here dare ask me why I know this!!!) Humble already came up with a simplified system that balances BC and PE: HiOpt 2. I have to agree with ZenMasterFlash on using the HOII system. Full disclosure: Flash mentored me on HOII. The application of HOII with ASC and 7SC for "pitch" games and the application of HOII with Three's BASC is all you really need if you want to 'go all out.' All other published systems get the money under the basis of brevity, even Wong's Halves.

You want to go beyond what is offered? See if you can take the Tarzan count, Heath count, or the DHM system and expand on one of those. I will venture to guess that such endeavours will be fruitless!

I will bet that your system would be an improvement of around, oh, 1-2% over High Low Full and 1-3% over Knock Out. That’s me being super fucking liberal with the expectations of your system. With that, take the expected values for HOII w/ASC, Halves, and Zen and see that under each system will see a gain of over 10% with respect to High Low and Knock Out. Rather than use your system (which is 2x the work for 1/2 the gain of other systems,) consider using the above posted.

[Three]

You won't get the full power of Hi-OPT II with side count of Aces using only the Illustrious 18 or the 22 indices + fab surrender. The full power is in Hi-OPT II with ASC full indices. Maybe Three would agree with me on this one.

Yes. Hiopt2/ASC gets its power from counting the 4, 5, and T as a tag magnitude two card, and side counting the ace and factoring it in at tag magnitude 2 for betting. By side counting the ace you can factor it in at whatever strength is optimal for plays either as a high card or low card or keep it neutral. If you are going to use a count you should learn as many indices as you can handle. By keeping the playing count ace neutral and level two the playing count has a high PE. Factoring in the ace appropriately when it is a strong enough card for the play makes these plays that much stronger. The I18 are the most important indices to learn first, but are by no means where you should stop learning indices. Unfortunately a lot of people have the misconception that you should just learn the I18 and fab 4.

Surrender, especially when you have big bets out, is one of your most powerful index plays. That power is about variance reduction a lot more than increased EV. The variance reduction allows you to bet more which is were most of the EV gain comes from. A Hiopt2/ASC 10:1 optimal spread with surrender that is 20 to 200 would have an optimal spread of 16 to 160 without surrender. You lose 20% of your advantage betting EV when you lose surrender. To see the full effect, for the latter optimal bets' EV for Hiopt2/ASC with surrender is $58.40/100 rounds and EV without surrender is $40.77/100 rounds for a total loss of 30.2% of EV when you lose surrender. So 2/3rds of the loss when you give up surrender is from decreased optimal bets and the other 1/3rd is from the EV loss from the surrender plays themselves. This is where RA surrender, especially for your biggest bets, gets its power. You get more EV from the bet change from surrendering than the EV loss when it is slightly negative EV to surrender. Overall variance may increase due to betting more but the variance on your worst matchups that you lose the vast majority of the time goes to 0. Surrender is a huge tool for counters that most don't take to its full potential.

[seriousplayer]

Damn, what a long thread.

PS. I'm curious what kind of results bjanalyst has in actual casino play. For some reason, I don't think he has a lot of experience in actual casino play, however he seems like a very intelligent person.

Yes, because bjanalyst keeps posting the same thing over and over and over and over again. His threads have been repetitive. Like I say if he has a Bachelor’s degree in mathematics, he would be better off working on discovering and developing fundamental ideas in mathematics or solving one of the Millennium problems. This way he would be more influential than attacking the game of blackjack.

You are making this more complex than it needs to be. As another poster stated, you are taking systems that are meant to be simple in execution and complicating it with a convoluted system you though up and making things more difficult for practitioners.

That is correct.