Zenfighter: Playing correlations for 6dks.

[Zenfighter]

 

			CORRELATIONS FOR 6 DECKS AND S17			 

							 

  Hand	        Hi-Lo	Hi-OptI	  RPC	  Zen	OmegaII	 UAPC	Halves 

							 

Insurance	0.7601	0.8498	0.7799	0.8498	0.8498	0.9014	0.7247 

16 vs. T	0.5499	0.6284	0.5769	0.6150	0.6567	0.7048	0.6432 

16 vs. 9	0.5287	0.5804	0.4459	0.4678	0.5090	0.5035	0.5032 

15 vs. T	0.7758	0.8985	0.8246	0.8908	0.9288	0.8922	0.8615 

13 vs. 3	0.6884	0.7872	0.7528	0.8252	0.8932	0.9247	0.7884 

13 vs. 2	0.6633	0.7605	0.7321	0.8070	0.8576	0.9122	0.7566 

12 vs. 6	0.6627	0.7937	0.6766	0.7651	0.7836	0.8431	0.6228 

12 vs. 5	0.8201	0.8688	0.8321	0.8837	0.8573	0.9048	0.7696 

12 vs. 4	0.7719	0.8720	0.8267	0.8773	0.8498	0.9121	0.7659 

12 vs. 3	0.7095	0.8110	0.7687	0.8423	0.8140	0.8781	0.7128 

12 vs. 2	0.6385	0.7319	0.6984	0.7699	0.7232	0.8078	0.6321 

11 vs. A	0.8260	0.9216	0.8443	0.9157	0.9677	0.9680	0.8460 

10 vs. A	0.9434	0.9031	0.9755	0.9844	0.9396	0.9498	0.9513 

10 vs T	        0.8278	0.6872	0.8915	0.8526	0.7664	0.8139	0.8730 

9 vs. 2	        0.8176	0.7921	0.8472	0.8894	0.8455	0.9059	0.8033 

9 vs. 7	        0.8590	0.7940	0.8274	0.8281	0.8837	0.8254	0.8691 

8 vs. 6	        0.8660	0.8606	0.8284	0.8295	0.8660	0.8564	0.8528 

8 vs. 5	        0.9368	0.9374	0.9134	0.9188	0.9422	0.8991	0.9346 

A8 vs. 6	0.8038	0.9493	0.8633	0.8775	0.9127	0.9279	0.8937 

A8 vs. 5	0.8499	0.9540	0.9153	0.9042	0.9053	0.9004	0.9296 

							 

							 

			THE FABULOUS FOUR				 

							 

15 v 9	        0.8145	0.8742	0.8069	0.8320	0.8464	0.7579	0.8546 

15 v T	        0.8132	0.8737	0.7982	0.8367	0.8729	0.7801	0.8387 

15 v A	        0.6805	0.7492	0.6553	0.6829	0.7452	0.6161	0.7264 

14 v T	        0.6581	0.7191	0.7597	0.8115	0.8507	0.8237	0.7787 

 

 

 

 

 

a) Correlations are generic, and not composition-dependent. Not any card has been removed here (see BJA, page 522 for the Insurance correlations,
after removing dealer?s ace). E.g. T,5 vs. T; 9,6 vs. T and 8,7 vs. T. We do not get the same correlations, as the above printed one for 15 vs. T, but their results will be quiet similar and cluster in the nearby. For Hi-lo and T, 5 vs. T we get: 0.7729. Well in the proximity of the generic.

b) With 11 winners out of 24, UAPC behaves itself, as a flamboyant count, to deal with the particular situations during the play of the hands. How could be otherwise, coming from Ken? God bless his soul!

c) Almost unbelievable is watching Dr. Humble?s Hiopt I, running with 6 winners. Especially for Surrender decisions this count is a killer. I?m beginning to understand now, his affection for Caribbean destinations. My bet is, he winning enough money to pay for his vacation expenses, at least. :-)

d) There is nothing it can be done to handle the 16 vs. 9 with moderate good efficiency. The meager correlations speak for themselves. What about turning your face from the table to avoid the vision of its fatal fate? Isn't there any surrender table handy? A good medicine against despair!

e) Finally, my apologies for the `unbalanced? counters round here. My analyser is ill suited for this. But most likely Cacarulo?s one would not. A ?subtle? hint!

All in all, enjoy your own count.

Regards,

Zf

[Don Schlesinger]

Thank you for this!

In comparing the correlations of the Hi-Lo I18 to my column 3 on p. 62 of BJA3, I was pleased to see not one number off by more than 0.01, and many exact. Since I did my work for four decks, 23 years ago, and with a pocket calculator, needless to say, I'm happy.

Less perfect are the Fab 4, especially two of the four. Unfortunately, Griffin did not have EORs for surrender, and I relied on data from Steve Jacobs, some 11 years later, to generate those values. Obviously, he made some errors. If there is ever a BJA4, I will have to make corrections.

Don

[Zenfighter]

Griffin did not have EORs for surrender, and I relied on data from Steve Jacobs, some 11 years later, to generate those values. Obviously, he made some errors.

Yes, but slightly with the individual EoR?s. Nevertheless, look at the full-deck favorability numbers below and double check your column 5. They are perfect, as printed!

 

15 vs. T m (4) =   0.462768 

14 vs. T m (4) =  -3.3377 

15 vs. 9 m (4) =  -2.86308 

15 vs. A m (4) =  -1.90655 

 

With the probability of occurrence column sealed, too, no small differences in correlations will alter the order of gain and importance of the four indices,
with the 14 vs. T beating every other one in actual gain. In plain English: The Fab four are The Fab four!

If there is ever a BJA4, I will have to make corrections.

Vous ?tes infatigable, monsieur Schlesinger! :-)

Zf

[Cacarulo]

Hi ZF,

First of all let me say: Great work!

Here are some interesting tag' properties that you can use in order to solve the unbalanced problem. Take into account that this is only useful for true-counted systems:

1) We can add/subtract any number to/from the system' tags and still get an equivalent system with exactly the same correlations.

2) We can also divide/multiply by any number and still get an equivalent system.

Let's work out an example like TKO:

TKO: -1 1 1 1 1 1 1 0 0 -1

let's calculate the tag's mean as:

m = sum (tag(i)) / 13 for i = 1, 13
m = 1/13

If we now subtract "m" from each tag we get:

ETKO: -14/13 12/13 12/13 12/13 12/13 12/13 12/13 -1/13 -1/13 -14/13

At this point we have a system with the same correlations of TKO but no longer unbalanced! If you calculate the mean now you'll find that it's exactly zero.

Playing a little more with the above properties we can multiply each of these tags by 13 and get another system with exactly the same correlations as TKO. The interesting thing is that we are now dealing with integer tags. I don't know how much "ill suited" is your analyzer

So:

ETKO-1: -14 12 12 12 12 12 12 -1 -1 -14

which is still a balanced system.

we can go a little further and add +1 to each of these tags:

ETKO-2: -13 13 13 13 13 13 13 0 0 -13

but the balance is lost again. Note that if we divide by 13 we'll get TKO!

Now you can complete your work!

Sincerely,
Cac

[Zenfighter]

 

		CORRELATIONS FOR 6 DECKS AND S17	 

						 

  Hand		         TKO		UBZ II		Uston SS 

						 

Insurance		0.7832		0.8448		0.7288 

16 vs. T		0.5715		0.6204		0.6333 

16 vs. 9		0.3811		0.4909		0.5139 

15 vs. T		0.7818		0.8673		0.8275 

13 vs. 3		0.7575		0.7993		0.7683 

13 vs. 2		0.7406		0.7782		0.7374 

12 vs. 6		0.6795		0.7599		0.6280 

12 vs. 5		0.8305		0.8803		0.7758 

12 vs. 4		0.8168		0.8742		0.7481 

12 vs. 3		0.7608		0.8136		0.6930 

12 vs. 2		0.6970		0.7401		0.6146 

11 vs. A		0.8429		0.9125		0.8474 

10 vs. A		0.9746		0.9773		0.9491 

10 vs. T		0.9203		0.8436		0.8693 

9 vs. 2		        0.8709		0.8662		0.8131 

9 vs. 7		        0.8420		0.8510		0.8998 

8 vs. 6		        0.8177		0.8692		0.8746 

8 vs. 5		        0.8836		0.9428		0.9418 

A8 vs. 6		0.8083		0.9064		0.8496 

A8 vs. 5		0.8482		0.9310		0.8781 

						 

						 

						 

			THE FABULOUS FOUR			 

						 

15 vs. 9		0.7248		0.8237		0.8272 

15 vs. T		0.7375		0.8306		0.8251 

15 vs. A		0.5813		0.6837		0.7083 

14 vs. T		0.7975		0.7902		0.7590 

 

 

 

 

 

Systems tags and efficiencies

 

TKO        = -1 1 1 1 1 1 1 0  0 ?1	PE = .5528 

UBZ II     = -1 1 2 2 2 2 1 0  0 ?2	PE = .6160 

Uston SS   = -2 2 2 2 3 2 1 0 -1 -2 	PE = .5548 

 

 

The Uston SS player is a sort of Halves? one. That is, a monster betting correlation, but as a consequence, the efficiency sinks.

The UBZ II outperforms slightly his ?father? in betting correlation, but on the other hand the playing efficiency drops slightly, too.

TKO is clearly better than Hi-lo. The numbers speak for themselves.

At the risk of repeating myself:

Enjoy your own count!

Sincerely,

Zf

P.S. My special thanks to Cac and Brett Harris. Seems like I?ve learned a couple of good things from these two BJ experts, lately!

[Cacarulo]

You're welcome!

One thing that you need to clarify is that when you give efficiencies for UBZII you're in fact talking about True UBZII! UBZII is actually a RC system like K-O.

Sincerely,
Cac

[Zenfighter]

[GeoC]

IRC starts at 0. You can chanage these numbers to conform to your own IRC. For shoe games.

Decks Played TC =>+1 TC => +2 TC => +3

1.0 9 14 19

2.0 12 16 20

3.0 15 18 21
Wong out:

1.0 deck gone: running of -2
2.0 decks gone: +4
3.0 decks gone: +8

This adds a bit of power to the UBZ2 user for the shoe games.

[GeoC]

For a TC of +1 and 1 deck dealt you need a running of +9.
For a TC of +2 and 1 deck dealt you need a running of +14.
For a TC of +3 and 1 deck dealt you need a running of +19.

I think you can figure the rest of it out from the data above.