Shoe Deal vs CSM effect on HE

I've been working on a new game and running it through game simulation software using a shoe dealt setup and have sorted out the probabilities, so far so good. As the game could also be dealt from a CSM I set the game simulation to run as if it was from a CSM expecting a slight drop in HE from the shoe deal. What I have actually seen is a drop of 2 - 2.5% Whilst this is alarming the game is running at 3.93% so this drops it to low to mid 1.x% I at first though I'd done something wrong with the CSM deal but checking it out I am confident this is working and giving a good reflection of a CSM style operation.

Some of the analysis I have been doing revolves around what I would call the Premium Starting Hands, these are where the player should always be playing. With a shoe deal the PSH are dealt 28% of the time with non premium around 72% whilst with the CSM I noticed the PSH rose to 30% and non premium was 70% which was strange as the difference I had seen was around the 2% mark. Considering that with both deals the win rate from a PSH is around 66% I don't think the win rate is a factor if dealt from a shoe or CSM.

So my question:
Am I correct in thinking that the 2% diff I see on Premium Starting Hands between a shoe and CSM deal is accountable for the HE diff I am seeing?

I am surprised to hear that switching from a shoe to a CSM (in a simulation) is causing the frequency of certain dealt hands to increase from 28% to 30%, and lowering HE by 2-2.5%

1. What is the bet structure of the game? Ante followed by optional 1X bet after seeing cards, or something different?
2. How many decks in your shoe and CSM?
3. How many cards does the player receive?
4. Is this game a "poker variant" game or what? (It sounds like a single-deck 3-card poker game.)

If you are saying that your simulator predicts a 2-2.5% decline in house edge then yes, it is possible that a 2% increase in "premium starting hands" for the player could be causing most or all of that decline, depending upon the bet structure and on answers to the questions above.

Hi gordonm888, thanks for the reply and in answer to your questions

1. Ante followed by optional 2x play bet or fold
2. 6 decks
3. Player makes their play or fold decision after seeing 3 cards with 2 cards to follow
4. Yes it is a Poker variant but not a single deck 3 card poker game :)

Does that help?

Yes, with a 2x play bet on the player's premiere hands, I find it quite believable that an increase form 28 to 30% in these premiere hands is causing a 2-2.5% reduction in house edge.

Do you understand why the frequency of these premiere hands has increased when simulating a CSM-dealt game? If so, then you are fine and have no problems. I don't understand it, though.

Poker hands dealt from 6 decks? Yikes, I imagine that flushes and pairs/trips/quads will be somewhat more common.

The game "Texas Shootout" is dealt from a 6 deck shoe. Very good older game.


ZCore13

I would like to say that I do understand exactly why but I am seeing consistency in test results and randomly checking the premium hand % which always comes in at 30%. I am running more tests with csm deal then retest with shoe to re-validate those results. I feel ok with it having found this premium hand variance.

Thanks for your input :)

Hi Zcore13

I was reading another thread yesterday and saw this game mentioned, fortunately it's not similar to my game in the way in plays and glad to see you called it a "very good" older game, perhaps my game will have a chance :)

Hi gordonm888

Quote:

Do you understand why the frequency of these premiere hands has increased when simulating a CSM-dealt game? If so, then you are fine and have no problems. I don't understand it, though.

This got me thinking even more about this and then it occurred to me why we might be seeing the difference.

With a shoe, the cards are shuffled and as the game plays the number of cards available decreases and so the number of cards that can make a premium hand also decrease. I know the cards are in a random sequence and come out in order but creating that random sequence has the decreasing availability.

Now, consider the CSM, what we effectively have is a selection made from all 312 available cards every game, therefore the possibility of having a premium starting hand must increase, thus accounting for the extra 2% I've been seeing through my testing.

What are your thoughts on this being the probable cause for the increase?

Quote: UKMark

Hi gordonm888


This got me thinking even more about this and then it occurred to me why we might be seeing the difference.

With a shoe, the cards are shuffled and as the game plays the number of cards available decreases and so the number of cards that can make a premium hand also decrease. I know the cards are in a random sequence and come out in order but creating that random sequence has the decreasing availability.

Now, consider the CSM, what we effectively have is a selection made from all 312 available cards every game, therefore the possibility of having a premium starting hand must increase, thus accounting for the extra 2% I've been seeing through my testing.

What are your thoughts on this being the probable cause for the increase?

Yes, very good. That may indeed be what it is.

The frequency of paired hands (1 pair, two pairs, trips, etc) is higher with 6 decks than with fewer decks. Averaged over many shoes, I would expect that the frequency of paired hands will decline as you penetrate further and further into a 6 -deck shoe. This effect is relatively insignificant in blackjack but would be much more important in a 5-card poker game.

The same is also true for making flushes, although the effect is not strong as with pairing.

So with a multi-deck shoe, the frequency of pairs will, on average, be highest at the outset and decline with penetration. If the game involves a player vs dealer, then the effect on House Edge would generally be small, because both the dealer and player will be affected equally as pairs become less frequent. But if your game is player vs. a paytable, then the House Edge would, on average, increase as shoe penetration increases.

Does your game have any sort of cut card effect? That could explain some of the difference, that is why blackjack from a CSM is slightly different than from a shoe.

Quote: UKMark

Now, consider the CSM, what we effectively have is a selection made from all 312 available cards every game, therefore the possibility of having a premium starting hand must increase, thus accounting for the extra 2% I've been seeing through my testing.

What are your thoughts on this being the probable cause for the increase?

Mark,
To test this you can sim the game on a CSM with 3 decks. This should give you about the same average number of cards in shoe as a 6 deck shoe dealt game with a deep cut card. If the House edge is very different from your original value you probably have something different happening

Quote: jopke

Does your game have any sort of cut card effect? That could explain some of the difference, that is why blackjack from a CSM is slightly different than from a shoe.

Hi jopke, yes there is a cut card which would, as you said, also affect the HE. A good point as I hadn't factored that aspect into the overall conundrum, thanks.

Hi, UKMark, given the CSM swing it sounds like this game would be beatable by card counting.