True edge for complete zen count

[FrankieT]

I read the section on zen count true edge in blackbelt in blackjack. If I read it correctly, it states that your edge goes up .25% per each +1 true count(i.e. your true count is +4 you have increased your chances of winning by 1%).


Ok, so if you play a 6 deck game where the casino edge was .6% if you just played basic strategy, you're true count would have to be +3 at least for you to have an edge against the casino, and that edge would only be .15% ; if it was +4 true count, your actual edge against this particular casino for that hand would be .4% or close to it). Does this sound correct or did I misread it somehow?

I've always known how to use zen count but I learned it originally in "play blackjack like the pros" and he just gives a typical betting system for zen count (+1 true count=$5 +2=$20 +3=$30 +4=$40.....). He doesn't mention anything about the actual edge gained per true count point in the zen system. Kind of pisses me off because i've been playing games where the the casino edge was .5% or greater (I knew these kind of games sucked but I used huge bet spreads to overcome) and was raising my bets on +2 counts which I shouldn't have, I should have raised my bets starting at +3 true count because that's where any sort of edge starts)

 

[supercoolmancool]

I read the section on zen count true edge in blackbelt in blackjack. If I read it correctly, it states that your edge goes up .25% per each +1 true count(i.e. your true count is +4 you have increased your chances of winning by 1%).


Ok, so if you play a 6 deck game where the casino edge was .6% if you just played basic strategy, you're true count would have to be +3 at least for you to have an edge against the casino, and that edge would only be .15% ; if it was +4 true count, your actual edge against this particular casino for that hand would be .4% or close to it). Does this sound correct or did I misread it somehow?

I've always known how to use zen count but I learned it originally in "play blackjack like the pros" and he just gives a typical betting system for zen count (+1 true count=$5 +2=$20 +3=$30 +4=$40.....). He doesn't mention anything about the actual edge gained per true count point in the zen system. Kind of pisses me off because i've been playing games where the the casino edge was .5% or greater (I knew these kind of games sucked but I used huge bet spreads to overcome) and was raising my bets on +2 counts which I shouldn't have, I should have raised my bets starting at +3 true count because that's where any sort of edge starts)

You just shouldn't deal with true counts at all and just calculate true edge. I don't think it is really any harder and all the indexes are for true edge in Blackbelt. And yes you have to subtract the house edge off the top from your true edge calculation.

 

[Sonny]

If I read it correctly, it states that your edge goes up .25% per each +1 true count

It should be closer to 0.5% per TC using the standard TC method, which Blackwood uses. Snyder’s numbers are different because the True Edge system does not calibrate by full decks as most systems do but instead uses quarter-decks. If you are using the True Count method then your advantage will start at around +2, but using the True Edge method your advantage will not come until around +4.

Don Schlesinger discusses some of the shortfalls of the True Edge system here:

http://www.advantageplayer.com/blackjack/forums/bj-main/webbbs.cgi?read=21413 (Archive copy)

-Sonny-

 

[QFIT]

You may be confusing Zen Counts. If you divide by remaining full decks (which is likely if you learned from Play BJ Like the Pros or the 1980 version of Blackbelt), then the ramp you were using is closer to correct. If you learned from BiB, then you would be dividing by quarter decks remaining and your ramp would be much too slow. However, the ramp according to True Edge is too slow even if done correctly and reduces SCORE.

Assuming that you divide by full decks, a good Zen ramp for the game you are describing would be:

+1 $5
+2 $15
+3 $30
+4 $40
+5 $60

Close, I gather, to what you were using. This will provide a better SCORE than using True Edge.

 

[FrankieT]

Ok, the game i'm eyeing has the following rules:

2 decks
S17
No double after split
Double on anything

Given these rules the basic strategy engine says the house edge is .33%

I play 2 deck games mainly and use a system where you multiply by a conversion factor, 0 cards seen=.5, 18 cards seen=.6, 30 cards seen=.7, 39 cards seen=.8 47 cards seen=.9, 52+ cards =1...It is something like that...


Lets say that roughly 34 cards have been dealt in this hand, therefore I would multiply my running count (attained using the zen count system) of 9 by .7 and get a true count of 6 (6.3 rounded down). I would have an edge of about 1.16% (6 * .25 + -.33). Or is that not my edge? Anybody know what my edge is if that is not correct.

 

[QFIT]

You are mixing metaphors. The method you are using is perfectly valid. TE is a clever method of estimating advantage and then estimating the optimal bet from that. One estimate on top of another. What you are doing now is more accurate.

However, your posts seem a bit contradictory as I can't figure out whether you are playing 2 or 6 decks.

 

[zengrifter]

It should be closer to 0.5% per TC using the standard TC method, which Blackwood uses.

Blackwood uses the 1/2D calibration, BUT Wizard is correct that for 1D calibration the edge rises .25% per TC +1. zg

 

[FrankieT]

Ok, so I would have a 1.16% edge on that particular hand in the example I just gave. Here is another hypothetical situation..

Say I am playing a game with the following rules:

6 deck
H17
Double after split
Double on anything

According to the basic strategy engine the house would have an edge of
.66% with these rules. Lets say that about 4.5 decks are remaining and I have a running count of 22 (attained using the zen count system). I would divide 22/5 and get a true count of 4. (I always round the number of decks up and round the true count down just for safety's sake). Ok so wouldn't I have an edge of .34% (4*.25 + -.66%) on this particular hand?

 

[QFIT]

Assuming a full set of indexes, your edge would be 0.79%, your standard deviation 1.146 and your bet 7.11 units. This is close to Kevin's 8 units.

 

[FrankieT]

I still don't understand how my way of calculating the advantage for the first example would be correct but my way of calculating the advantage for the second example would not be correct.

When you divide by decks (12/5, 14/4, 22/6) it is the same as using the conversion factors that I used in the first example only that the conversion factors are so tight when you have more then 2 decks (12/3=12 * .333,
14/4= 14 * .25, 22/5= 22 * .2, 11/6 = 11 * .166. So if my advantage in the first example is 1.16% (for that particular hand) shouldn't my advantage in the second example be .34% (for that particular hand)?

 

[QFIT]

True edge is an estimate on top of an estimate. You estimate the edge at TC 0, then add an estmiate for edge per TC. You then use another estimation to calculate the bet. I'm using actual numbers. In your second example, you use a BS advantage of -.66. But, Basic Strategy advantage is not identical to a counter's TC 0 advantage. The correct number for the example is -.58% assuming Zen full indexes, 4.5/6 penetration, and flooring. You use an advantage per TC of .25%. The actual numbers are .43% from TC0 to TC1, .31% from TC1 to TC2, .31% from TC2 to TC3, .32% from TC3 to TC4, .29% from TC4 to TC5, .36% from TC5 to TC6. Using these numbers I get an advantage of .79% at TC4. The optimal Kelly Bank in this example is 1,183 units assuming a 1-12 spread. The optimal bet is the Kelly Bank * the EV / Std Dev squared. 1183*.79/1.146^2 is 7.11. So, the optimal bet is 7.11 units. 7.11 * $5 is $35.6, close to the $40 in Kevin's book.

Which is a long and boring way of saying that the ramp in Kevin's book is good. In double deck, all the numbers are different.

(Kelly Bank is HPH * std dev ^ 2 / win rate. In this example, 100*4.69^2/1.86=1183)

 

[FrankieT]

Oh ok, I see things are a little more mathematically complicated then I originally had thought. I got to pick up that Kelly Book

Can I use your skilled blackjack mind for one more thing? Wondering if you could calculate the average dollar amount earned per 100 hands under the following conditions.....

2 decks
S17
No double after split
Double on anything
66% penetration
Using the zen count with this bet spread: 0=5, 1=5, 2=20, 3=30, 4=40, 5=50........ 10+=100
Using the complete zen count indices as given on this website: (Dead link: http://www.blackjackforumonline.com/content/zen_count_indices.htm) (-10 through +10, not the really crazy ones)

Also for these conditions......

6 decks
H17
Double after split
Double on anything
Split 4 times
Resplit Aces
90% deck penetration
(Same bet spread, count, and indices as above)


Wondering if it's possible to find out the risk of ruin for the two games mentioned above using the same bet spread and count (with a $10,000 bankroll) as well?

 

[QFIT]

Sorry, don't have the time at the moment.

 

[ortango]

Wong follows Thorps estimates, which are approx .5% per + TC. To be more exact, he calculates +.56% per count, with .22% deviation using Hi-Lo. Where did you guys get those other low numbers for edge?
:cow:

 

[QFIT]

Wong follows Thorps estimates, which are approx .5% per + TC. To be more exact, he calculates +.56% per count, with .22% deviation using Hi-Lo. Where did you guys get those other low numbers for edge?
:cow:

This isn't HiLo. EV by TC is affected by the count used, indexes, rules, penetration, and TC calculation method. Optimal bets are affected by additional factors.

 

[zengrifter]

Wondering if it's possible to find out the risk of ruin for the two games mentioned above using the same bet spread and count (with a $10,000 bankroll) as well?

Mental guestimate: $35/100 - I could be way off. Get a simulator! zg

 

[FrankieT]

True edge is an estimate on top of an estimate. You estimate the edge at TC 0, then add an estmiate for edge per TC. You then use another estimation to calculate the bet. I'm using actual numbers. In your second example, you use a BS advantage of -.66. But, Basic Strategy advantage is not identical to a counter's TC 0 advantage. The correct number for the example is -.58% assuming Zen full indexes, 4.5/6 penetration, and flooring. You use an advantage per TC of .25%. The actual numbers are .43% from TC0 to TC1, .31% from TC1 to TC2, .31% from TC2 to TC3, .32% from TC3 to TC4, .29% from TC4 to TC5, .36% from TC5 to TC6. Using these numbers I get an advantage of .79% at TC4.

I never knew this, very useful information. Could you possibly give me the starting edge (that is the edge when the true count is 0 when you're using zen count with full indexes) for other deck penetrations for 6 deck and 2 deck, and give me the edge increase per each point in the true count (at least 1-8)....do you have more information like this on hand? Is there a book I can buy that has detailed charts with more examples (of the starting edge and edge increase for each true count point using zen full indexes)

 

[QFIT]

I never knew this, very useful information. Could you possibly give me the starting edge (that is the edge when the true count is 0 when you're using zen count with full indexes) for other deck penetrations for 6 deck and 2 deck, and give me the edge increase per each point in the true count (at least 1-8)....do you have more information like this on hand? Is there a book I can buy that has detailed charts with more examples (of the starting edge and edge increase for each true count point using zen full indexes)

Blackjack Attack has the info plus optimal bets in Chapter 10. But only for HiLo. CVCX has all this data. However, if you have CVCX, you don't need the info CVCX tells you the actual optimal bets so you don't have to bother with True Edge estimations.

 

[FrankieT]

Seriously? I already bought casino verite 4.0, could you just provide the numbers for a 2 deck game 60% penetration, that's all I ask

 

[QFIT]

Zen Full indexes, divide by full decks remaining, H17, DAS, 60% penetration has a starting edge of -.32% and edge per TC varies from .3% to .45%.