ssho88
• Posts: 669
Joined: Oct 16, 2011
July 1st, 2019 at 5:18:17 AM permalink
Hi All,

My program can generate EORs and can simulate games once user input the "Human" counting system. I want to make my simulation program become fully automatic, less data input and generate "Human" counting system automatically.

I am thinking to convert generated EORs to a "Human" counting system with an additional sub-routine program. For example, the generated EOR from Ace to KING is ( 0.115, 0.105, 0.240, 0.230, 0.131, 0.121, 0.011, -0.701, -0.556, 0.103, 0.103, 0.103, 0.103), once we read it with naked eye, we know that the good counting system(with high correlation) is (+1, +1, +2, +2, +1, +1, 0, -7, -5, +1, +1, +1, +1), but how can we write a program codes to generate such a high correlation human counting system ?

My suggestions :-

1) Convert all EORs to absolute value(ABS) and identify the maximum EOR absolute value, ABS(EORmax).
2) Find the ratio of ABS(EORmax) / ABS(EORn), where n = 1,2, 3, . . . . 13.
3) Elimiate EORn if ratio ABS(EORmax) / ABS(EORn) > 10, OR ratio ABS(EORmax) / ABS(EORn) > 12 ?
4) After elimination process, find a multiplier factor :-

a) if ABS(EORmax) >= 0.1, multiplier factor = 10.
b) if 0.01 =< ABS(EORmax) < 0.1, multiplier factor = 100
c) if 0.001 =< ABS(EORmax) < 0.01, multiplier factor = 1000
d) . . . . .

5) Then apply this factor to all original generated EORs, and rounding up all new EORs (after apply multiplier factor)

6) Modifications . . . to find a more human friendly system. Instead of (+1, +1, +2, +2, +1, +1, 0, -7, -5, +1, +1, +1, +1), why don't we use (+1, +1, +2, +2, +1, +1, 0, -6, -6, +1, +1, +1, +1) ? Which I think it is more user friendly and still maintain high betting correlation.

The above procedures may good for common set of EORs but unable to cover certain awkward EORs set. Please give your comments and help to give ideas to improve above methodology so that it can cover 99% of all possible EORs sets. LOL

James
gordonm888
• Posts: 5121
Joined: Feb 18, 2015
July 1st, 2019 at 11:20:14 AM permalink
James, are you calculating EORs for individual decisions in BJ like 16 vs 10 and 3-3 vs 8? Or EORs for BJ variant games or for other games?
Last edited by: gordonm888 on Jul 1, 2019
So many better men, a few of them friends, are dead. And a thousand thousand slimy things live on, and so do I.
ssho88
• Posts: 669
Joined: Oct 16, 2011
July 1st, 2019 at 3:22:28 PM permalink
Quote: gordonm888

James, are you calculating EORs for individual decisions in BJ like 16 vs 10 and 3-3 vs 8? Or EORs for BJ variant games or for other games?

I have a program to calculate ev for any given BJ hand, for example 16 vs T, the total 16 can be consisted any no of cards( up to maximum 10 cards), provided that you can peovide correct strategy. For example (2, 4,, A, 9) vs T, the program can calculate(simulate) the ev if you specify to HIT(or STAND, DOUBLE...). So if you repeat the same process by removing a card "2" from the shoe, re-calculate the ev2, then you can get EOR2 = ev2 - ev.

6 deck, PEEK rule, 300 million trials(simulation) :-

1) (2, 4,, A, 9) vs T, STAND, ev = -54.028%
2) (2, 4,, A, 9) vs T, STAND, removed a "2" from the shoe, ev2 = -54.144
3) (2, 4,, A, 9) vs T, STAND, removed a "3" from the shoe, ev3 = -54.133
4) . . . . .

EOR2 = ev2 - ev = -0.116%
EOR3 = ev3 - ev = -0.105%
. . . . . .

The program can calculate EORs for BJ variant games, it cover special rules such as Dealer Push 22, Double after split AA, Hit after split AA.....

AND

I have other programs to calculate EORs for baccarat as well.
Last edited by: ssho88 on Jul 2, 2019
ssho88
• Posts: 669
Joined: Oct 16, 2011
July 2nd, 2019 at 6:40:20 AM permalink
Quote: gordonm888

James, are you calculating EORs for individual decisions in BJ like 16 vs 10 and 3-3 vs 8? Or EORs for BJ variant games or for other games?

Another example, 6 deck, peek, DAS, SPL1 :-

a) 3,3, vs 8, HIT, ev = -21.91%
b) 3,3, vs 8, SPLIT, ev = -22.98%
c) 3,3, vs 8, remove 4 no of cards(2,2,3,T) from the shoe, HIT, ev = -22.06%
d) 3,3, vs 8, remove 4 no of cards(2,2,3,T) from the shoe, SPLIT, ev = -21.84%

From above, HIT is a better option if you don't have any extra info, However. if other players holding (2,2) and (3,T), then you should SPLIT !
gordonm888
• Posts: 5121
Joined: Feb 18, 2015
July 2nd, 2019 at 8:30:00 AM permalink
Okay, that is what I thought you were referring to. As I have mentioned in WOV posts, I have a similar capability.

Here is a graphic showing the marginal EV, in %, for decisions of all close calls in BJ. This is for 8 decks, H17, DOA, DAS, no surrender. The marginal EV is calculated while making no assumptions about what cards the player holds, their removal will be counted in whatever counting system you use. The closest decisions, 0-1.3% are in yellow-orange, decisions that are 1.31-3.2% are in red and 3.2-5% in blue . (Edit: corrected colors in image)

1. Other than Insurance, the most important decision by far is 16vsT - it is ultra-close and occurs more frequently than any other close decision. And the EORs are relatively large. A side count of 5=-1; 6=+1 is very simple and captures a lot of the additional EV.

2. Judging by closeness and frequency, the next three most important are 12v4 and 13v2 and 9vs2. 12v4 is all about the 8s and 9s; and 13v2 is all about the 9s.

3, Hands like A3v4 are no good, because you basically have the same hand as dealer and so any cards that are going to help you will also help dealer. In other words, EORs are small.
- similarly, EORs on 12vs2 will be small.

4. Soft hands and pairs don't occur very frequently. A7vs 2 and 33vs2 are interesting.

5. When dealer has a 2, the EOR of 9 will always be large -and there are a relatively large number of player hands that are close decisions vs a dealer 2. It may be that a generic '9 count' might have some value in telling you whether dealer's 2 is weak or strong.
Last edited by: gordonm888 on Jul 2, 2019
So many better men, a few of them friends, are dead. And a thousand thousand slimy things live on, and so do I.