Where can you get game analysis tables for the Halves count(like those in Blackjack Attack? I use this count and mostly play multiple decks. What are the betting correlation, playing efficiency and insurance correlations of this count to three decimal places? Is this the best known count for betting?
Unbelievable, I ran a few sims(6 decks) on BJRM and the results are trying to tell me the lame unbalanced KO one level count is superior to the halves 3 level count system ?
I am gaffed.
This suggest ko is the best for you to use,unbelievable!
I wonder if Ken Fuchs programed the software?
See p. 295 of BJA. There isn't a chance that what you're writing is correct. NONE.
Go back and find your mistake.
Don
I plugged in a variety of 6D rules and penetrations, and in every case, Halves outperformed KO, providing the bet units and spreads were equal. I was surprised at how close some of them were, but Halves was on top in every case.
Example:
Game: 6D, S17, DAS, LS, 5/6 pen, 1-12 spread, play all, optimum integer bets, $25 unit.
SCORES:
KO: 48.30
Halves: 49.00
I tried other rules and conditions, but Halves always had the higher SCORE.
BTW, you cannot "run sims" on BJRM. It is not a simulator. You are actually accessing a database of sims that have already been run.
Parker,
Been trying to find the score for TKO for a while now. Can you please compare the score values of TKO to Halves using the Illustrious 18 (not the rounded indices found in the book, but the actual numbers). 6-deck simulation would work well for me seeing i count six and eight decks. I would be interested to how they compare.
Thanks for any help you are willing to provide me
Impala
When you true count an unbalanced count you gain a little bit sometimes by counting the cards that are zero at a slight minus. Blackjack Therapy now on bjrnet.com has a discussion on this and how to estimate the BE and PE of such counts, in the chapter on shuffle tracking. In this case the values ace through ten -1,1,1,1,1,1,1,0,0,-1 get changed effectively to : -14/13,12/13.....-1/13,-1/13,-14/13 which has a slightly higher PE and BE.
You are making the values perform like a near 13 level system. Brett Harris'es count and the T-hopper Bushido counts take advantage of this to push the performance of level 2 values the same way.
You can work this out yourself with Blackjack Therapy to find the long run index of these systems with your game conditions. I don't think you will find many games where TKO actually beats halves though!
While I was "accessing" sims to see what I may have did wrong I realized that your answer is wrong. SCORE is based on a $10,000 bank, using optimal betting at a %13.5 risk of ruin. BJRM shows KO to be the favorite. I believe your results were achieved from raising the minimum bet up to $25, this will increase the win rate but is devastaing to risk of ruin (about 50%.) This is not considered "SCORE".
The SCORE in BJRM is already programmed to a set bank of $10,000.
Plug in the count used>decks>rules>load>calculate>optimal betting at 1-12, and there you have the true SCORE locating in the bottom left boxes. You will see why I find these sims hard to believe when comparing KO to other powerful count systems.
It is not me doing anything wrong,I'm just retreiving the data.
What are you saying? That I believe KO is a stronger count than halves? That I'm wrong about BJRM results?
Please take a few moments and check out the results for yourself.
There appears to be a bug in BJRM with respect to Halves. See the p. 295 BJA charts for the correct data. Halves is clearly superior to K-O, but it would appear that BJRM is not reporting the data properly.
Don
It is not me doing anything wrong,I'm just retreiving the data.
I can't make KO beat halves for any of the 4.5 of 6 or 5 of 6 games, in BJRM 2000, so I don't know what you are talking about.
The KO and halves pre-run sims in BJRM are not apples-to-apples anyway.
Also, as a side note, I believe that counting the "9" is not a net-net good thing for a count to do, especially in multi-deck.
The default settings I use are the basic 4.5,h17,das,optimal 1-12. This retrieves data to show the SCORE at the bottom window,correct?
Results are also showing KO's SCORE to be very close when compared to even the most powerful counting systems. Am I overlooking something or is KO actually that strong of a count?
Thanks John.
S17, DAS, 10K Bank, Spread Opt1 - Max (1-12 spread), "use integer bets" is unchecked. click "calculate optimal bets."
I get the following SCORES:
KO: 19.80
Halves: 19.36
With H17 (everything else the same):
KO: 12.60
Halves: 12.53
If rules are improved (add late surrender) and/or pen is improved, then Halves, jumps out ahead, although not by the margin that I would expect.
This is especially puzzling since I notice that the halves sim uses "many more indices than the I18/Fab 4."
page 295 of BJA gives Halves a huge lead over KO.
Something seems wrong for the current values to be so different from the book values.
Don
I was curious how you arrived at your earlier results when you assumed I was wrong?I tried to get the same results as you did by nipping and tucking a bit. Did you notice how your risk of ruin increased to about 50% in order to make halves out-perform KO? You must have not used all default settings.
Thats where it gets tricky and you must keep an eye on risk of ruin,bankroll,ev,units,etc. by using the calculation punch. When I first purchased BJRM I almost made a terrible mistake while figuring out my bank roll and risk of ruin. Fortunately I caught my mistake in time and had to re-adjust, otherwise I might have went broke assuming I was a much lower risk of ruin than what I actually was.
I normally have BJRM set to a $25 unit and "use integer bets" since this is the level that I play at, and using a unit size of $17.53 (or whatever) is impractical, to say the least.
I guess the size of your unit does matter. :-)
Since your optimal unit bet is $17.53 you have greatly increased your risk of ruin by playing at $25 tables so take pre-cautions.
I would suggest moving down to the $10 tables, this will lower your risk of ruin to a much safer level.
Actually, "$17.53" was just a number I pulled out of the air.
My bankroll is replaceable, so risk of ruin, at least in the classic sense, isn't really applicable.
In the meantime, I want to restate that BJRM 2000 is NOT to be used to compare one system vs the next, unless you KNOW that the underlying sims were carefully setup using the same "everything", including version of SBA, exact penetration, indices, method of TC calc, etc.
I know that the halves and KO do not meet that criteria.
That said, the halves sims in BJRM can certainly be used to set optimum bet placement, and give very reasonable approx of your expected win rates, ror, etc. Keep in mind that you will NEVER be able to play as the computer does, anyway, so don't get too anal about all this. :-)
If you want to see a real, careful comparison, do as Don suggests and look at the SCORE chapter in BJA2. Those sims were setup to be, specifically, a comparison.
-JA
Is that the same as a "flaw" in a computer simulation program?
