UBZ2 vs. Uston SS

[ace157]

I'm looking to "graduate" from KO. I was interested in UBZ2, but Uston SS caught my eye. The 99+% BC for SS is pretty tempting and i don't mind the difficulty, just takes time to get used to. However, UBZ2 is not far behind and has a much greater PE. What are everyone's thoughts on a comparison of the two? I don not hav ne of the QFIT software, but i would be interested in seeing how the two score against each other

[zengrifter]

UBZ is a little easier. Also, our very own BJINFO* has its collaborative open-source UBZ composite. zg

*aka WBPITTW

[SD Padres]

Why not just add more indices to KO?

[Mimosine]

learning UBZII is far better than learning more indicies for KO.
You can play a VASTLY simplified UBZ index and still trounce KO. See thread and sims for UBZ "moron" where all index plays except insurance are at RC = 0. It still beats KO Preferred or FULL (can't remember).

[Automatic Monkey]

Originally Posted by ace157

I'm looking to "graduate" from KO. I was interested in UBZ2, but Uston SS caught my eye. The 99+% BC for SS is pretty tempting and i don't mind the difficulty, just takes time to get used to. However, UBZ2 is not far behind and has a much greater PE. What are everyone's thoughts on a comparison of the two? I don not hav ne of the QFIT software, but i would be interested in seeing how the two score against each other

If you're looking to stay with an unbalanced count BRH-1 will probably outperform them both, particularly in shoe games.

System tags {A...10} are {-2,1,2,2,3,2,1,0,0,-2}

[ace157]

Thanks for the input guys.

Thanks for pointing out OS-UBZ ZG, and thanks Mimo for givn' me the quick rundown opposed to readn' the entire thread (even though i did afterwords).

Monkey, i'm havn' a lil trouble finding info on BRH, i can't seem to turn up much on Brett Harris' systems, can u point me in the right direction?

For some reason I had a liking for SS, but i'll do more research before switching so i can get a good feel for all 3 alternative systems, for the time being I'm fairly satisfied using a culmination of KO-Preferred and "The Ultimate Gambit"(Ian Anderson)

[ace157]

So in lieu of more information to make my decision, i was playing around with CVData's Efficiency Calculator. I began taking the first step of Qfit(.com's) steps to creating a card counting system. I don't really want (dunno if i can) create my own, but possibly hybrid a couple systems and see what i come up with. Anyway, here are the results:

(BC/PE/IC)
UBZ2: 98.4/54.1/78
Uston SS: 99.5/55.4/72.8
BRH-1: 98.9/55.7/76.2
SS w/o 9 count: 99/54.7/76.4
UBZ2 (9=-1): 98.9/54.8/74.2
None of this is anything new, other than not counting 9s w/ SS or possibly counting 9s (as -1) with UBZ2. I just thought the comparison was interesting. Yes Monkey, BRH-1 seems to be a nice fit between the two. I did find it interesting that by simply adding a 9 count to UBZ2 it perfomred better and still remained the only Lvl 2 system in my comparison. When i attempted to make a compromised (between UBZ2 and SS) lvl 2 its efficiency failed horribly.

My own trials:
Trial 1: BC = 94.9, PE = 59, IC = 83.3
Tags 2 - A: 1, 1, 2, 3, 1, 1, 0, 0, -3, -2

Trial 2: 90.7, 62.3, 88
Tags: 1, 1, 1, 3, 1, 1, 0, 0, -3, -1

Trial 3: 92.2, 55.4, 82.9
Tags: 1, 1, 2, 2, 1, 0, 0, 0, -3, -2

Trial 4: 98.1, 60.5, 81.3
Tags: 2, 2, 2, 3, 2, 1, 0, -1, -3, -2

Trial 5: 97, 60.3, 83.9
Tags: 1, 2, 2, 3, 2, 1, 0, 0, -3, -2

First off i apologize if i have "stolen" any other systems tags, i was jsut trying numbers based on calculations. I had never seen any other count with a negative three, but i thought i would make sense in a lvl 3 count since the weighted impact of all the 10s in a deck vs any other number is much MUCH higher (bcuz there r 16 rather than 4). I of course claculated what i would call each card value's "weighted impact" taking the card value's percent effect of removal and multiplying it by their weight in the deck (ie: 4/52 for everything except 10s). Based on this number i tried to place fitting card tags on the card values. The only trial that came close to performing next to the others would be trial 4, which sacrifices almost a whole percent in BC, but has a strong PE. Most of the trials had a whopping IC, but that is the leas important (in my opinoin) of the three efficiency stats. They all seemed to lack in the BC department but made up for it with a more than acceptable PE. However, we know that BC is the most important stat in shoe games. The only downside to trial 4 is that the tags are somewhat complicated compared to the supercharged performance of BRH-1, or if ur a stickler to ur system, remove 9s for SS or add them to UBZ2.

Next I'll run some sims and calculate some indexes and hopefully come up with something good. I don't have a problem with lvl 3 systems, and im not arguing with some of the VERY smart ppl who who have recommended UBZ2 over SS, i just want to find out for myself and see if there is a possible compromise to be had... though Brett Harris seems to have found one.

[ace157]

Using standard rules, DAS, S17, no surrender, 15:10 blackjack, 40 million roundsand a 1 - 15 bet spread i compared a few things.

Player 1 - Uston SS
Player 2 - UBZII Composite
Player 3 - Uston SS w/o 9s
Player 4 - UBZII Composite with 9 index of -1
Player 5 - Austin Trial 4 (see previous post)
Player 6 - BRH-1

The bad news is that my trial count system was an epic failure, with a 500 dollar BR, it yielded only $1.42 per hour and lost 118,130 units after 40 mill rounds, as well as a NEGATIVE cSCORE.

The good news, however, while SS-No9 has a NO of 60,418 it yielded a win rate of $16.99/hour and won over 1.4 million units through 40 mill rounds of play. BRH-1, SS, and UBZ2-C all yielded within $0.11/hour of each other. BRH-1 had the best SCORE and NO, followed by UBZ, and then SS in both categories.

THE NEXT STEP:
"Austin Trial 4" looks to be a wash, however simply not counting 9s with an SS system looks promising. Much more so than UBZII w/ a 9 count which performed terribly as well ($1.68/hr, Score 2.12, NO 470,734). I'll keep pushing the SS-No9 theory and explore a few more principles ... perhaps Trial 4 will not have been a complete loss.

[ace157]

tested some new ideas yet again. The control group (for lack of a better word) was the SS system. I kept all of the rules of the game the same EXCEPT i made sure that there were NO insurances, the reason being that i am not far along enough in my own system to declare when the player would have an advantage, so my system would b playing at a disadvantage.

New Systems:
The new ideas were an continuation of my (further) failing Trial 4 (now trial 4.2) which had the tags (2-A) 2, 2, 3, 3, 2, 1, 0, 0, -3, -3. It was also recommended to me by BJI member Jack, Jackson to try (i guess its a) lvl 4 systems that i labeled JJ4 had the tags 2, 2, 3, 4, 3, 2, 0, -1, -3, -3. I also added a new trial of my own (T7) with the tags 2, 2, 3, 3, 3, 1, 0, -1, -3, -3.

Efficiency Results:
SSNo9 - BC 99, PE 54.7, IC 76.4
Trial 4.2 - 98.8, 54.9, 78.5
JJ4 - 99.7, 57.5, 75.2
T7 - 99.3, 55.9, 75.6

Sim Results:
this got sticky.... fast. SS had the best SCORE and NO followed by SSNo9, JJ4, and far, far behind were AT4.2 and T7. However, it was SSNo9 that yielded $7 more per hour than any of the competition.

Conclusion thus far:
Uston knew his s*** no doubt, but by simply removing the nine count, better results were yielded and it seems (to me) it would be a lil easier. If you really wanted to go off the deep end, JJ4 has 4 different tags AND counts 9s; but it also has results to back it up. I don't think the problem for me is having 3 or 4 tags (working w/ a lvl 3 or 4), I feel by not having to keep track of as many numbers helps.

Remaining Questions:
Would it really make a difference removing the 9 count for SS?
Is JJ4 worth a lil extra trouble if ur planning on using a lvl 3 already?
What would come out of a JJ4 w/o a 9 count?
Why to my made from scratch counts suffer when they test to be efficient but suck it up on sims?

[Automatic Monkey]

Just keep one thing in mind now. When you start adding to the unbalance of unbalanced systems you'll need to use it only in true count mode. Otherwise the difference in betting correlation between the beginning and the end of the shoe becomes too large and you end up heaving too much money out there at the end of the shoe with no advantage. But if you are true count adjusting it, you can make the system tags whatever you want.

Your SSno9 surely earned quite a bit of money, but how about the SCORE? Doing things like that raises the standard deviation so when you take into account your risk of ruin, it will actually earn less money than the other existing systems.