The best level one system $1000 contest

[scroogemcduck]

Let's settle the "what's the best Blackjack counting system" debate once and for all with a contest. I'll pay $1000 in bitcoin to the submission with the highest SCORE in CVCX, subject to the following rules:
Stations DD rules (2D H17 DAS 75%pen RS4 noRSA noSurrender 1burncard)
Single-parameter Level1 count
6 to 1 spread
one hand only
Play All
<= 50 indices (counting insurance, not counting A9vX indices if they're a copy of TTvX)
No depth-based or composition-based indices.
No suit-aware tags.
If you do true count conversion, Quarter Deck is the limit on deck estimation. No Exact Card.

If there is a tie, whoever posted it first gets the prize.
Winning submissions will be independently verified in CVCX so don't go messing around with wonging/bonuses.

To be eligible for the prize, you have to beat my system by 1%: https://imgur.com/a/0pTNEFs
That means you need a score of 80.76 in a CVCX sim of at least 10 billion rounds.
Submit your entry by replying to this thread with a screenshot of the sim and all the index tables (just like my imgur link above). Upload it somewhere instead of posting as an attachment because BJTF will make the resolution too low to read.

The contest ends Nov 30 at midnight.

[scroogemcduck]

For some context, on these rules, CVCX's built in sim for Wong's complete hi-lo system (213 indices) does 4.5% worse than my 50-index KO. Halves Full using all the indices between -5 and +10 (way over 50) does a fraction of a percent worse than my 50-index KO (and as a level 3 count, halves wouldn't qualify). I don't expect to have to pay out, but if I do, it'll be worth $1000 to get an extra 1% EV over a lifetime.

[Freightman]

For some context, on these rules, CVCX's built in sim for Wong's complete hi-lo system (213 indices) does 4.5% worse than my 50-index KO. Halves Full using all the indices between -5 and +10 (way over 50) does a fraction of a percent worse than my 50-index KO (and as a level 3 count, halves wouldn't qualify). I don't expect to have to pay out, but if I do, it'll be worth $1000 to get an extra 1% EV over a lifetime.

I surmise your collaboration with bjanalyst? It may well be true what you claim regarding halves, maybe. However, Halves FBM ASC Basic would then handily defeat you. Halves FBM ASC Advanced would therefore - Crush you

[Norm]

Of course, you would need an exact definition of level 1. Some people believe Red7 is level 1 and some don't. By single-parameter, I assume you mean no side counts of any kind.

[scroogemcduck]

Of course, you would need an exact definition of level 1. Some people believe Red7 is level 1 and some don't.

I'll define level one as all the tags are -1, 0, or 1. So Red7 is eligible.

By single-parameter, I assume you mean no side counts of any kind.

Correct.

Scroogemgduck and BZjAnalyst, they be identical twins?

I don't know BJAnalyst and I developed this myself using CVData RA BTD with some manual tweaks to the Hard DD indices to move them down from RA to almost MaxEV levels to conserve cards in positive counts.

[Norm]

Ahh, Red7 is level I. That means this is also.

Double deck:

	Club	Diamond	Heart	Spade
Ace	-1	-1	0	0
2	1	1	1	0
3	1	1	1	0
4	1	1	1	1
5	1	1	1	1
6	1	1	1	1
7	1	1	0	0
8	0	0	0	0
9	-1	-1	0	0
Ten	-1	-1	-1	-1

Six Deck:

	Club	Diamond	Heart	Spade
Ace	-1	-1	-1	-1
2	1	1	1	0
3	1	1	1	1
4	1	1	1	1
5	1	1	1	1
6	1	1	1	1
7	1	1	0	0
8	0	0	0	0
9	-1	0	0	0
Ten	-1	-1	-1	-1

Actually, this could be further refined into something that no human could handle -- but still be level I under that definition.

[scroogemcduck]

Ahh, Red7 is level I. That means this is also.
...
Actually, this could be further refined into something that no human could handle -- but still be level I under that definition.

Yes, but they'd have no chance of winning anyway, because suit-aware counts use irrelevant info that just adds noise to the RC. Red7 is much worse than either KO or Hi-Lo.

Since CVCX doesn't support simming strategies with different values for red and black, suit aware counts are implicitly banned.

[scroogemcduck]

The screenshot should be based on a minimum of 10 billion rounds in CVCX.

[bjanalyst]

Yes, but they'd have no chance of winning anyway, because suit-aware counts use irrelevant info that just adds noise to the RC. Red7 is much worse than either KO or Hi-Lo.
.

Actually Red 7 performs better than HL but worse than KO.

The Red 7 is unbalanced at 2 per deck and KO is unbalanced at 4 per deck so these in calculating true counts.

balanced(Red 7) = Red 7 - 2*dp and balanced(KO) = KO - 4*dp and these balanced counts are used to calculate the true counts which is balanced(Red 7) / dr or balanced(KO) / dr where dp = decks played and dr = decks remaining.

If the truce counts are calculated then the Red 7 outperforms HL but underperforms KO.

See attached PDF.
Primary counts infinite deck CC comparision.pdf

[Norm]

Actually Red 7 performs better than HL but worse than KO.

Please don't make general statements like this. It's nonsense. And, when you find an infinite deck game, maybe -- just maybe -- your attachment might make sense. Seriously, there have been stats based upon actual conditions that are far better than the stuff you keep posting.

[DSchles]

Please don't make general statements like this. It's nonsense. And, when you find an infinite deck game, maybe -- just maybe -- your attachment might make sense. Seriously, there have been stats based upon actual conditions that are far better than the stuff you keep posting.

P. 176, BJA3. :-)

Don

[bjanalyst]

P. 176, BJA3. :-)

Don

Thanks for the reference to the sims you did for HL, Red 7 and KO

The results of your sims are at odds with my CC projections that Red 7 beats HL and that KO beats Red 7. And as you know my CC technique has been proven again and again to be wildly successful. It is hard to argue with success.

In a previous post I had listed the KO indices that I derived using the LSL technique and suggested that you re-run your sims using the true count formula and the KO indices I gave you where tc(KO) = 4 + (KO - 4*n)/dr where tc(KO) = KO true count, KO = KO running count, n = number of decks and dr = decks remaining.

It should further be noted that the KO indices are very close to the HL indices so there really just a few situations where KO indices differs from HL indices that need to be learned.

For the shoe game I suggested KO with my TCRC (Table of Critical Running Counts) so in practice there is no need to use a formula to calculate the true counts for the shoe game.

For the two deck game where true counts jump all over the place, I would not use an unbalanced count but I would suggest HL with Am6c (and 5m9c) or the HL2 with Am6c (and 5m9c) where HL2 is the HL with the 2 counted at +1/2 instead of +1 in the HL and the 7 counted as +1/2 instead of 0 in the HL.

[bjanalyst]

Please don't make general statements like this. It's nonsense. And, when you find an infinite deck game, maybe -- just maybe -- your attachment might make sense. Seriously, there have been stats based upon actual conditions that are far better than the stuff you keep posting.

I was just listing the results of the CC which show that your counts and the Black 2, Red 7 count all have a lower betting CC than HL, Red 7 and KO. That is a fact.

I used the lower betting CC to conclude that B2R7 is the worst count and I believe that is what you are disagreeing with.

Now you argue that the CC do not mean anything which is an entirely different argument and that the count you suggested and that B2R7 are really superior counts to HL, Red 7 and KO, that is an entirely different argument.

But just be aware that if you say that the CC do not mean anything then you are totally ignoring the results of around 20 simulation that showed that whenever the CC increased, even on a case by case basis, such as top 6 strategy changes of HL w 7m9c tested one at a time, the SCORE increased. There was not a single situation where the CC increased and the SCORE decreased.

Anyhow this entire constraint of keeping a level one count with no side counts is artificial and kind a silly.

I use KO with AA89mTc and 5m9c with chips which I can do effortless for hours on end and I play the six deck, five deck dealt game, S17, DAS, LS with Lucky Ladies and Super 4 side count and I do very well. And for LL I use KO + AA89mTc and for Super 4 I use KO - (AA89mTc + 5m9c).

As you are aware the KO with 5m7c and AA89mTc has been shown through Gronbog's sims to beat the pants off of the HO2 w ASC for every single play all and back counting scenario and for both LS and no LS.

And it should be noted that I derived my KO with 5m7c and AA89mTc entirely with infinite deck CC's and the LSL technique to derive indices.

So all I can say is that the CC have been shown to work very well.

I will leave it to the reader to draw their own conclusions.