CVDATA "REAL SHUFFLE" vs "STANDARD SHUFFLE"

[James989]

Why is the EV value simulated by CVDATA with "REAL SHUFFLE" always lower than the EV value of "STANDARD SHUFFLE" ?

[iCountNTrack]

Why is the EV value simulated by CVDATA with "REAL SHUFFLE" always lower than the EV value of "STANDARD SHUFFLE" ?

Well it depends on the shuffle. What shuffle are using? Did you define a good complex shuffle. If I remember correctly it had a simple riffle restack. Poor shuffles have a pronounced effect. Please look at one of the threads I started

[Norm]

Yes, the sample shuffle is a simple R&R which has an effect on EV.

[Archvaldor]

Well it depends on the shuffle. What shuffle are using? Did you define a good complex shuffle. If I remember correctly it had a simple riffle restack. Poor shuffles have a pronounced effect. Please look at one of the threads I started

Yeah this.

A specific shuffle with a specific initial deck order will have a completely deterministic output unless some randomness is added. This is the basis of a lot of magic tricks.

This means very little in practice since most actual casino shuffles are complex. If you find a truly terrible shuffle then you can destroy it with sequencing.

[MJ1]

I was always under the impression that the shuffle has no bearing on EV. Isn't the player advantage unaffected by the shuffle?

MJ

[Norm]

Generally not. The one pass R&R I have as a sample is basically non-existent these days. Every several years someone tells me they have seen it. Of course there are many levels of randomness included in the CVData shuffles which can be detailed by the user if you have the CVData/ST add-in. The sample is just a sample.

[iCountNTrack]

I was always under the impression that the shuffle has no bearing on EV. Isn't the player advantage unaffected by the shuffle?

MJ

Please see the below, it appears that the less random the shuffle the worst is the effect on EV


https://www.blackjacktheforum.com/sh...trategy-player

[iCountNTrack]

@Norm, one idea I have in mind is whether the initial state of the shoe has an impact on the EV where you are more or less "stuck" in a bad state because of the non-randomness of the shuffle and you are basically "oscillating" between similar states.

One idea to counter this, is to randomize the shoe after a certain number of real shuffles (say 1000 for example) this in fact would mimic more real life scenarios.

Not sure if this possible in CVData real shuffles or there is a quick way for you to implement it

[Norm]

CVData does randomize periodically like a casino table getting new cards. Don't remember how often. If you have multi-threading on, it is basically playing at multiple tables with different shoes and aggregating results. There are a bunch of parts to randomization customizable by step. Not likely to have an oscillating effect -- like a magician's shuffle:

sh55.jpg