How much Edge from Optimal Play

[Egon Olsen]

Acording to the basic strategy engine,
with the rules of 6 Decks, S17, DAS, No Surrender, Peek
the house has an edge of 0.44% if we play basic strategy,

but what if we don't play basic strategy,
what if we don't use a counting system,
but insted had a computer take the theoretical correct decision every time.

How much would the EV of this game change,
assuming we flat bet 1 unit every hand, and about 4-to-5 of the 6 decks was dealt before shuffling?

Have you made the calculations/simulations, or do you have an educated guess?

 

[tthree]

You need to specify penetration. Sorry missed the 4 to 5 but that is a big range.

 

[iCountNTrack]

Acording to the basic strategy engine,
with the rules of 6 Decks, S17, DAS, No Surrender, Peek
the house has an edge of 0.44% if we play basic strategy,

but what if we don't play basic strategy,
what if we don't use a counting system,
but insted had a computer take the theoretical correct decision every time.

How much would the EV of this game change,
assuming we flat bet 1 unit every hand, and about 4-to-5 of the 6 decks was dealt before shuffling?

Have you made the calculations/simulations, or do you have an educated guess?

It all depends on the number of decks and the penetration, the fewer the decks and the deeper the penetration the more gain we get from playing strategy deviations. for deeply dealt single deck (might neeed a time machine) the edge is about 5-6% depending on the rules. Also for a single deck at about 14 cards left you will always play be playing at an advantage if you knew the composition of the remaining cards

 

[Egon Olsen]

6 deck game
4.5 decks dealt before shuffling

 

[ericfarmer]

Acording to the basic strategy engine,
with the rules of 6 Decks, S17, DAS, No Surrender, Peek
the house has an edge of 0.44% if we play basic strategy,

but what if we don't play basic strategy,
what if we don't use a counting system,
but insted had a computer take the theoretical correct decision every time.

How much would the EV of this game change,
assuming we flat bet 1 unit every hand, and about 4-to-5 of the 6 decks was dealt before shuffling?

Have you made the calculations/simulations, or do you have an educated guess?

I was running something unrelated to this, but the data go partway toward answering your question. The following plots show expected return versus shoe penetration for both basic and optimal strategy. For each plot, a single simulated player played 500 shoes heads-up. On the left, the player used the same composition-dependent, “full-shoe” basic strategy (CDZ-) for every hand. On the right, the player used a different strategy continuously re-optimized for each hand (but still CDZ-), based on the current depleted shoe.

Each blue point indicates expected return for the corresponding depleted shoe prior to each hand. The red curves simply give some view of density, indicating 10th percentiles of the distribution of expected return.

 

[Egon Olsen]

@ericfarmer
Interresting

I have just started programming my own simulations, and will report back here when/if I succeed