4 best systems for a shoe..

#21
bjcount said:
Are you kidding? 1 & 3 are so close in shoe games that if you make just 1 mistake using ZEN per 20 hrs play time they'd be dead even.

Not taking anything away from AM but BF is a simplified & rounded version using RPC tags, it doesn't come close to 1 & 3 when properly played.

Flash you really need to invest in cvdata and find out for yourself.

BJC
Whoa, wait. In modern shoe games when you're using huge spreads and backcounting, the lower BC of the Zen count is a detriment. Zen somewhat underperforms High-Low if your spread is big enough. The effect is worse in games with any of the most desirable rules (S17,LS,RSA) because the EOR of the ace is boosted by these rules. Zen is a great count for tougher games like H17, D10, DD because the advantage of the ace is significantly diminished by those rules and the half-rank on the ace is a lot closer to reality.

Being the most important playing index is right at 0 and most of the other good ones happen to be very close to 5 and 10, the rounding doesn't really cost you anything. The thing about indices is they represent the point where it doesn't matter if you do the play or not. So if the TC is +8 it doesn't matter at all that an index that is exactly +4 has been rounded to +5. And when the TC is +4, it hardly matters at all, the decision is close to a coin-flip.

All of it amounts to angels dancing on the head of a pin; we all should live so long as to expect to realize a difference between all these counts!
 

boneuphtoner

Well-Known Member
#22
I'll suggest an excellent strong count that no one has mentioned....our very own UBZII Open source. My own highly rounded and compromised 25 index version (described on page 2-3 of the thread) definitely beats RPC and Mentor with the uncompromised Sweet 16, Fab 4 indices. The fully uncompromised indexed versions of RPC and Mentor perform comparatively to UBZII open source in a 6 deck shoe...see here for the SCORE charts:

http://www.blackjackinfo.com/bb/showthread.php?t=6996&page=10

If you use CVCX and take a look at the standard UBZII sims, you will see that it underperforms. I think I have a reasonable explanation. One factor I can attribute this to are those indices are not risk averse....an obvious one would be doubling 10 v. 10. My risk averse index for this one is 15 above the pivot. The canned sim used an index of 1 above the pivot! Splitting 10's? The canned version has you doing it 3 above the pivot, mine has you doing it 10 above the pivot. Another factor are the actual indices chosen...many of the Illustrious 18 aren't even in there! I think both of these factors play in role in explaining why the canned version of UBZII underperforms in shoes.

But my own sims show, with the right indices, it is actually an extremely powerful count for shoes.
 

bjcount

Well-Known Member
#23
boneuphtoner said:
I'll suggest an excellent strong count that no one has mentioned....our very own UBZII Open source. My own highly rounded and compromised 25 index version (described on page 2-3 of the thread) definitely beats RPC and Mentor with the uncompromised Sweet 16, Fab 4 indices. The fully uncompromised indexed versions of RPC and Mentor perform comparatively to UBZII open source in a 6 deck shoe...see here for the SCORE charts:

http://www.blackjackinfo.com/bb/showthread.php?t=6996&page=10

If you use CVCX and take a look at the standard UBZII sims, you will see that it underperforms. I think I have a reasonable explanation. One factor I can attribute this to are those indices are not risk averse....an obvious one would be doubling 10 v. 10. My risk averse index for this one is 15 above the pivot. The canned sim used an index of 1 above the pivot! Splitting 10's? The canned version has you doing it 3 above the pivot, mine has you doing it 10 above the pivot. Another factor are the actual indices chosen...many of the Illustrious 18 aren't even in there! I think both of these factors play in role in explaining why the canned version of UBZII underperforms in shoes.

But my own sims show, with the right indices, it is actually an extremely powerful count for shoes.
Neat work up but I have a few comments... just my opinion, I never have run a UBZII sim so just some bits I'm throwing out:

1) Your results indicate a 1-8 spread, forgeting that a 1-8 spread will barely beat the average 6d game, were the spreads "optimum" or the generic spread provided with cvdata? The generic spreads will not provide comparative results since each strategy you compared uses a different TC divisor which causes the ramps to be significantly different.

2) Were you wonging? If so did you confirm that you were skipping a similar quantity of hands? Because each strategy produces a different percentage of TC frequencies, for example, RPC may have 37% hands skipped at TC-1, Mentor may have 46% hands skipped at TC-1, and I couldn't guess what UBZII skips.

It's not as simple as you think to just run a few sims using different strategies with similar pieces (bet ramp, # of indices etc. ) and expect to give you read ready comparisons. There's more to it then that.

3) Same issues apply to the DD regarding comparisons using the same bet generic bet spreads.

4) Your comparing only SCORE results. This does not mean that the win rates are comparitive, what about the SD? How far apart are they? Have you looked at the total action per hour? While UBZII may suggestively indicate it is better are you betting $3,500/hr vs RPC/Mentor $2,900/hr and winning the same per hr? If thats the case why are you betting so much more to win so much less? For all we know the reverse may be true, your betting less and winning more, but that's not the information your graphs are supplying.

You have to look at all the data, not just one aspect.

BJC
 
Last edited:

bjcount

Well-Known Member
#25
Grisly Dreams said:
Color me confused. Isn't this the precise disparity that SCORE was intended to correct?
SCORE wasn't intended to correct anything. Sorry my error in posting, the higher SCORE should indicate that the win rate and action per hour is higher then a lower SCORE.

It's still tough to compare if all things are not "apples to apples".

BJC
 

FLASH1296

Well-Known Member
#26
I endorse what boneuphtoner has had to say.

UBZII performs most admirably in shoe games; and if I was advising a
newbie or casual player I would have no problem with suggesting UBZII,
if the prospective Card Counter wishes to avoid True Count calculations.

One thing is clear to me. For an unbalanced count, UBZII is the way to go.
 

boneuphtoner

Well-Known Member
#27
Neat work up but I have a few comments... just my opinion, I never have run a UBZII sim so just some bits I'm throwing out:

1) Your results indicate a 1-8 spread, forgeting that a 1-8 spread will barely beat the average 6d game, were the spreads "optimum" or the generic spread provided with cvdata? The generic spreads will not provide comparative results since each strategy you compared uses a different TC divisor which causes the ramps to be significantly different.

2) Were you wonging? If so did you confirm that you were skipping a similar quantity of hands? Because each strategy produces a different percentage of TC frequencies, for example, RPC may have 37% hands skipped at TC-1, Mentor may have 46% hands skipped at TC-1, and I couldn't guess what UBZII skips.

It's not as simple as you think to just run a few sims using different strategies with similar pieces (bet ramp, # of indices etc. ) and expect to give you read ready comparisons. There's more to it then that.

3) Same issues apply to the DD regarding comparisons using the same bet generic bet spreads.

4) Your comparing only SCORE results. This does not mean that the win rates are comparitive, what about the SD? How far apart are they? Have you looked at the total action per hour? While UBZII may suggestively indicate it is better are you betting $3,500/hr vs RPC/Mentor $2,900/hr and winning the same per hr? If thats the case why are you betting so much more to win so much less? For all we know the reverse may be true, your betting less and winning more, but that's not the information your graphs are supplying.
First off, here is a comparison done with a more traditional 1-12 spread...not a lot of difference from before. Again, this compromised strategy was put up against full indexed versions with uncompromised indices. Not bad performance IMHO.

http://s219.photobucket.com/user/boneuphtoner/media/UBZ_6D_full1to12.jpg.html

You want comparative win rates for this system and this spread? You got it!

http://s219.photobucket.com/user/boneuphtoner/media/UBZWinRates.jpg.html

No wonging was done for these comparisons, as such this is a play-all comparison. Also, to answer your question, optimal betting ramps for each strategy were generated with CVCX and these charts reflect that. You can show the effect of Wonging on this new strategy but I'd have to run new sims on the other systems, as it isn't obvious to me how you would do it with the CVCX canned sims. I will tell you this...because of its pivot at a HiLo TC of +2, I think it would be very effective, certainly not much worse, if at all.
 

bjcount

Well-Known Member
#28
boneuphtoner said:
First off, here is a comparison done with a more traditional 1-12 spread...not a lot of difference from before. Again, this compromised strategy was put up against full indexed versions with uncompromised indices. Not bad performance IMHO.

http://s219.photobucket.com/user/boneuphtoner/media/UBZ_6D_full1to12.jpg.html

You want comparative win rates for this system and this spread? You got it!

http://s219.photobucket.com/user/boneuphtoner/media/UBZWinRates.jpg.html

No wonging was done for these comparisons, as such this is a play-all comparison. Also, to answer your question, optimal betting ramps for each strategy were generated with CVCX and these charts reflect that. You can show the effect of Wonging on this new strategy but I'd have to run new sims on the other systems, as it isn't obvious to me how you would do it with the CVCX canned sims. I will tell you this...because of its pivot at a HiLo TC of +2, I think it would be very effective, certainly not much worse, if at all.
Good job with the graphs I see how you developed your theory...... were these run in cvcx or cvdata? I have to presume your looking at cvcx graphs because your quoting cvcx above ...and ...RPC indices are not in cvdata. So if you didn't create a new RPC strategy to run in cvcx and if I'm correct about this then the results are not play all since all the canned sims in cvcx are wong out at TC-1. Which means your comparing wong out sims vs the one you set up, UBZII with 25 indices. or i'm just not getting what you did..

I'm not knocking ubz2,vfrom what I've read it a very good strategy, I'm just wondering how you came up with those results.

BJC
 
Last edited:

beating vegas

Well-Known Member
#29
Kosiuker said:
Hello guys
I am new at blackjack’s world, and newer in this forum. I’ve been spending the last 30 hours on the internet in trying to find “the system” for me.. Where I live (somewhere in Argentina) there are a few casinos, and rules are not the best ones.. for example, 6 decks, double down only with 10, 11 or when your first card dealt is an Ace (is that usual?)..
I see here there is a lot of people that know a lot about it, and I would really thank you help me with this.. I know the most recommended system is Hi-lo, but I really want another thing..
So what I have found is that I would prefer a balanced count instead of and unbalanced one, because I heard (from here) it is more accurate.. and I would be comfortable with a level 2 and maybe 3 count too.. but no with side counts.. and as I have seen, for a SHOE game, I still cannot choose between:
- Mentor
- Wong Halves
- Revere point count
- Uston SS
(yes yes, I know the last one is not balanced, but still, it seems to be very effective)
Other things I can tell you is that I won’t spend a LOT of hours at the tables because of the distance but I plan to be an "green chipper" (with a large spread) if I ever get good at it..
Help me decide for one of them please.. (any kind of information would be appreciated)
And sorry people for my poor English..
If your new to blackjack first make sure you know all of basic stragey and can add up to 21 with out thinking about it. when you pick a count i suggest you start with a level 1 qfit website has a stragey advisor best of luck
 

boneuphtoner

Well-Known Member
#30
Not a CVCX expert, but here is my take....

Good job with the graphs I see how you developed your theory...... were these run in cvcx or cvdata? I have to presume your looking at cvcx graphs because your quoting cvcx above ...and ...RPC indices are not in cvdata. So if you didn't create a new RPC strategy to run in cvcx and if I'm correct about this then the results are not play all since all the canned sims in cvcx are wong out at TC-1. Which means your comparing wong out sims vs the one you set up, UBZII with 25 indices. or i'm just not getting what you did..
These comparisons were done with the canned sims in CVCX. For my UBZ2 strategy, I generated 4 deck compromise risk averse indices with SBA, and then put this into CVCX. I know what you mean about the -1 wonging issue with the canned sims, but from the way I read it, this only comes into play if you check "back counting" under the betting tab. Otherwise, it is considered play all. When you set up these sims in CVCX, on the final sim page gives you the option of "lowest count tracked." The default there is -1. According to CVCX help, it states: "Normally data is not collected for TCs below -1. This is not normally needed for balanced strategies." So, my take on this is that the -1 issue is NOT a wonging phenomenon. I ran a sim once on a modified Zen with tracking down to -5. This allowed me to customize a betting ramp with a "semi-wonging" approach, wonging out at a -2 TC. Not surprisingly, that strategy did better than any of the canned sims, but you can't compare those because of the way they are set up.

In any case, the "backcounting" tab most definitely is not clicked for UBZ2-Ultimate.

I think I figured out how to potentially compare this...it isn't exact, but each one wongs in at approximately +2 (6 Deck, DAS RSA, 78% pen, wong in at +2):

Stategy SCORE Win Rate N0 ROR
RPC Full 52.37 $16.98 19096 1.8%
Mentor Full 50.90 $23.63 19646 6.1%
UBZ2-Ult 51.49 $18.68 19421 2.6%
 
Last edited:

bjcount

Well-Known Member
#31
boneuphtoner said:
These comparisons were done with the canned sims in CVCX. For my UBZ2 strategy, I generated 4 deck compromise risk averse indices with SBA, and then put this into CVCX. I know what you mean about the -1 wonging issue with the canned sims, but from the way I read it, this only comes into play if you check "back counting" under the betting tab. Otherwise, it is considered play all. When you set up these sims in CVCX, on the final sim page gives you the option of "lowest count tracked." The default there is -1. According to CVCX help, it states: "Normally data is not collected for TCs below -1. This is not normally needed for balanced strategies." So, my take on this is that the -1 issue is NOT a wonging phenomenon. I ran a sim once on a modified Zen with tracking down to -5. This allowed me to customize a betting ramp with a "semi-wonging" approach, wonging out at a -2 TC. Not surprisingly, that strategy did better than any of the canned sims, but you can't compare those because of the way they are set up.

In any case, the "backcounting" tab most definitely is not clicked for UBZ2-Ultimate.

I think I figured out how to potentially compare this...it isn't exact, but each one wongs in at approximately +2 (6 Deck, DAS RSA, 78% pen, wong in at +2):

Stategy SCORE Win Rate N0 ROR
RPC Full 52.37 $16.98 19096 1.8%
Mentor Full 50.90 $23.63 19646 6.1%
UBZ2-Ult 51.07 $18.68 19582 2.8%
You can double check with Qfit, or he may chime in later on but the TC-1 whether the backcount is checked or not makes no difference, your sims are only calculated down to -1. As you saw when you ran Zen down to -5 there were percentages of hands at those low counts. If you look at the canned sims at -1 the total hand frequencies add up to 100%. Go back to your archives and look at your Zen sim and you will see a % of those hands were spread down to -5 but still add up to 100%.

So basically what your not seeing in the canned sims @ TC<-1 is the program skipping those hands. CVData provides you with hands skipped so you can balance the amount of hands skipped when you compare strategies. This is not perfect either but much closer.

Now looking at your last sim result at +2 Wong in, RPC beat the other two. Probably because the 1/2dtc conversion usually has the average TC span from TC-6 to TC +6 above and below those are very rare in 6d.
So if your wonging in at +2 your probably skipping 65% of all the hands. This is just my take on it from the many sims I ran since I use RPC.

BJC
 

boneuphtoner

Well-Known Member
#32
I confirmed with Norm that if you look at canned sims, without backcounting checked, means play all. The fact that it stops at -1 means that you can't calculate stats or betting ramps below that. If you don't want to calculate an optimal betting ramp for a semi-wonging approach (like I use), that may be all you need.

But you are correct in that you can't see the RPC indices in CVCX, nor can you see how the TC was calculated. I can only guess that this is a result of copyright issues. As to the interpretation of the results, yeah, RPC full performed a touch better with a lower ROR with the wong-in of +2. You saw the full indexed versions nudging in front of my UBZ-II Ultimate at the deepest penetrations as well. Still, I think a decent argument can be made that this is a strong system for shoes when employed in running count mode only. Certainly vastly better than the canned sims of UBZ.
 
#33
boneuphtoner said:
I confirmed with Norm that if you look at canned sims, without backcounting checked, means play all. The fact that it stops at -1 means that you can't calculate stats or betting ramps below that. If you don't want to calculate an optimal betting ramp for a semi-wonging approach (like I use), that may be all you need.

But you are correct in that you can't see the RPC indices in CVCX, nor can you see how the TC was calculated. I can only guess that this is a result of copyright issues. As to the interpretation of the results, yeah, RPC full performed a touch better with a lower ROR with the wong-in of +2. You saw the full indexed versions nudging in front of my UBZ-II Ultimate at the deepest penetrations as well. Still, I think a decent argument can be made that this is a strong system for shoes when employed in running count mode only. Certainly vastly better than the canned sims of UBZ.
RPC is such a good system for shoes and UBZ such a good system for pitch, it's not a bad idea to learn both, being the only difference between the two is the way the ace is counted. I've been using both for years.
 
Top