Adjusting the Halves Count to Improve Score

[Midwest Player]

Vegas Casino Talk is not really a blackjack advantage player site. There are very few folks over there who know about advantage play. However, one ex-member of this site is posting on Vegas Talk that if you adjust the Halves count you can improve SCORE. He suggests changing the tag for the 2 to 0 and increasing the tag from .5 to 1 for the 7.

I want to know if there is any truth to his claim. If I remember correctly Stanford Wong based his halves count on the value of each card. Wouldn't he have rated the 7 as 1 and the 2 as 0 if that was the case.

The only information I can find on effect of card removal is from these two links. I believe it is also in Wong's book, but I'm too lazy to look it up.
https://wizardofodds.com/games/blackjack/appendix/7/
https://www.qfit.com/book/ModernBlackjackPage68.htm
Both these links give a higher rating to the 2 than the 7.

[seriousplayer]

Vegas Casino Talk is not really a blackjack advantage player site. There are very few folks over there who know about advantage play. However, one ex-member of this site is posting on Vegas Talk that if you adjust the Halves count you can improve SCORE. He suggests changing the tag for the 2 to 0 and increasing the tag from .5 to 1 for the 7.

I want to know if there is any truth to his claim. If I remember correctly Stanford Wong based his halves count on the value of each card. Wouldn't he have rated the 7 as 1 and the 2 as 0 if that was the case.

The only information I can find on effect of card removal is from these two links. I believe it is also in Wong's book, but I'm too lazy to look it up.
https://wizardofodds.com/games/blackjack/appendix/7/
https://www.qfit.com/book/ModernBlackjackPage68.htm
Both these links give a higher rating to the 2 than the 7.

Firstly, if you are going to change the tags on a Halves it is no longer Halves. Secondly, changing the 2 to 0 and .5 to 1 for the 7. It will decrease the betting correlation to .96 compare to original Halves with .99 BC. It also decreases the insurance correlation.

A-(-1)
2-0
3-1
4-1
5-1.5
6-1
7-1
8-0
9-(-.5)
T-(-1)

PE = .56 BC =.96 IC =.70

Halves

A-(-1)
2-.5
3-1
4-1
5-1.5
6-1
7-.5
8-0
9-(-.5)
T-(-1)

PE = .56 BC =.99 IC =.72

[Three]

I want to know if there is any truth to his claim. If I remember correctly Stanford Wong based his halves count on the value of each card. Wouldn't he have rated the 7 as 1 and the 2 as 0 if that was the case.

There was a poster here that got banned named Moses. he made this claim repeatedly. He played SD Reno rules exclusively. All his research was based on this rule set and his perceived restrictions on play due to heat and preferential shuffling if you made certain plays. Unless you are playing SD you should do my own research rather than follow Moses' advice. Maybe your research will confirm his claims for the game you play. Performance in real life can defy BC, PE and IC stats. Particularly BC stats. BC is how well the count tags correlate to betting EoRs. It doesn't tell you anything about betting accuracy, which is the range of actual optimal bet for the deck compositions around the betting bin average that you base all your bets on. If you keep RoR, spread, and BR constant to generate optimal bets for each approach, the one that bets more aggressively probably has better betting accuracy; although, PE and IC would also influence optimal bets. To be exact, plot the bell curve of actual advantage around each average advantage for each TC the one with the narrower bell curve bets more accurately. A high BC can have a very wide bell curve around each TC advantage. Just look at a Hilo advantage bell curve around any advantage TC compared to a count that bets more accurately.

https://www.blackjacktheforum.com/sh...-at-tc-1/page2

See post number 17 for Hilo (BC 97) at TC +1 and post number 30 for the equivalent advantage Hiopt2/ASC (BC 98). Note the scale on each graph. You see Hilo with a range of 3% around its TC average advantage and Hiopt2/ASC with a range of 1.5% around its TC average advantage.

[Midwest Player]

Firstly, if you are going to change the tags on a Halves it is no longer Halves. Secondly, changing the 2 to 0 and .5 to 1 for the 7. It will decrease the betting correlation to .96 compare to original Halves with .99 BC. It also decreases the insurance correlation.

A-(-1)
2-0
3-1
4-1
5-1.5
6-1
7-1
8-0
9-(-.5)
T-(-1)

PE = .56 BC =.96 IC =.70

Halves

A-(-1)
2-.5
3-1
4-1
5-1.5
6-1
7-.5
8-0
9-(-.5)
T-(-1)

PE = .56 BC =.99 IC =.72

Thanks, I just confirmed your numbers using this blackjack efficiency calculator. How could SCORE possibly be better with a BC of .96 vs .99?
https://www.blackjacktheforum.com/re...php?do=effpage

[Three]

Thanks, I just confirmed your numbers using this blackjack efficiency calculator. How could SCORE possibly be better with a BC of .96 vs .99?
https://www.blackjacktheforum.com/re...php?do=effpage

BC and betting accuracy are two different things. SCORE has to do with the square of EV to variance ratio. You need to do the sims to see whether the BC/PE/IC stats translates into higher SCORE. Don would only trust Doghand's sim results. Perhaps DH would run this sim and say how they compare for different games and spreads with the same BR, RoR etc.

[Dog Hand]

BC and betting accuracy are two different things. SCORE has to do with the square of EV to variance ratio. You need to do the sims to see whether the BC/PE/IC stats translates into higher SCORE. Don would only trust Doghand's sim results. Perhaps DH would run this sim and say how they compare for different games and spreads with the same BR, RoR etc.

Three,

Well, I'd first have to generate indexes for the "Halves Not" tags. Then I could run the sims to get the SCOREs.

What parameters would you recommend for the index generation and comparison sims?

Dog Hand

[Three]

Use the same parameters that were used to generate the Wong Halves indices. That would keep everything equal.

[Midwest Player]

I would be interested in a double deck game. h17, split 4 hands, doa, das, no surrender, one card on aces.

[Dog Hand]

Use the same parameters that were used to generate the Wong Halves indices. That would keep everything equal.

Three,

Excellent suggestion! What are they?

Dog Hand

[Dog Hand]

I would be interested in a double deck game. h17, split 4 hands, doa, das, no surrender, one card on aces.

Midwest Player,

What penetration? What bet spread? How many players?

Dog Hand

[Midwest Player]

Midwest Player,

What penetration? What bet spread? How many players?

Dog Hand

How about 50% pen, 1 to 8 spread on DD, and 2 players.

[Three]

Three,

Excellent suggestion! What are they?

Dog Hand

I never studied Halves. Maybe Freightman will tell you. He uses or at least learned to use the pure version of Halves.

[Freightman]

I never studied Halves. Maybe Freightman will tell you. He uses or at least learned to use the pure version of Halves.

My sims include all indices, since I want to measure myself against the best possible play, even if don’t use the index, such as splitting 10’s. So, halves, full indices. Though my primary game is ES10, I would go, for the masses, with LS.

Standard game for me is 6d, ES10, RSA,H17,DA2,DAS,3-2. Common penetrations are 4/6,4.5/6,5.0/6. I can’t, or don’t know how to sim my current ramp, which is dual, and includes switching back and forth within the same shoe. I did get a suggestion that I could sim, simply by switching ramps per shoe. I don’t want to do that though. I also play a couple of other nuances which again, I don’t know how to sim.

For those familiar with concept, let me know if too sensitive and I will delete.