charliepatrick
charliepatrick
Joined: Jun 17, 2011
  • Threads: 30
  • Posts: 1781
Thanks for this post from:
MrCasinoGames
August 12th, 2019 at 1:25:02 PM permalink
I'm still getting a slightly different answer but haven't got the hang of catering for peeking. For instance I guess you know there's a slightly higher chance of a picture when the player is playing their hand if the dealer with an Ace has already peeked. I'm getting 1.98% so there's probably a small bug in my program or logic somewhere.

Anyway this is the playing strategy it gives (so please accept there might be some small errors in it) but it must be fairly close to normal single deck strategy.
Soft 19Double A8 vs 6
Soft 18Double A7 vs 3-6Hit AA24 vs 9
Hit AAA5 vs A 9 10
Hit Axx vs A 9 10
Hit A7 vs A 9 10
Soft 17Double A6 vs 2-6
Hard 16Some hit, some don't
Soft 16Double A5 vs 4-6
Hard 15Stand a few vs 10
Soft 15Double A4 vs 4-6
Hard 14Stand 77 vs 10
Soft 14Double A3 vs 4-6
Hard 13Hit A2T, 49, 3T vs 2
Soft 13Double A2 vs 4-6
Hard 12Stand some vs 3
Hit 2T vs 4
Soft 12Double AA vs 5-6
Hard 11Double
Hard 10Double vs 2-9
Hard 9Double vs 2-6
Hard 8Double 44 or 35 vs 5-6
btw one advantage of giving the players their five cards is their hands cannot be affected by other players standing or hitting. An easier way is to deal the cards out as you go, but face down if the player has already stood.
CharlesMousseau
CharlesMousseau
Joined: Nov 1, 2016
  • Threads: 3
  • Posts: 26
Thanks for this post from:
GialmereDJTeddyBearMrCasinoGamescharliepatrickbeachbumbabstringlomane
August 12th, 2019 at 4:37:25 PM permalink
Quote: mrsuit31

It may have been Charles, based on a linkedin post I just saw.



I did the math for the original game (I got 1.9148% house edge based on an exact calculation) and a number of sidebets. One of them was a straight poker-hand-on-a-paytable one very similar to the Bonus bet (most of the payouts are the same) but the "Twisted Stud" thing and the 16-point thing was a complete after-market addition and I have had absolutely no influence on that or any of the other confusion that seems to be around here. The end game from my analysis was certainly much cleaner, but that was at about 7 years ago.
Wizard
Administrator
Wizard
Joined: Oct 14, 2009
  • Threads: 1248
  • Posts: 20591
Thanks for this post from:
tringlomane
August 12th, 2019 at 7:53:52 PM permalink
Quote: CharlesMousseau

I did the math for the original game (I got 1.9148% house edge based on an exact calculation) and a number of sidebets. One of them was a straight poker-hand-on-a-paytable one very similar to the Bonus bet (most of the payouts are the same) but the "Twisted Stud" thing and the 16-point thing was a complete after-market addition and I have had absolutely no influence on that or any of the other confusion that seems to be around here. The end game from my analysis was certainly much cleaner, but that was at about 7 years ago.



Thanks for chiming in Charles. Glad to hear you don't deserve any blame for the Twisted Stud fiasco.

Too bad Dan Lubin isn't around. I would have loved to see his comments on this game.
It's not whether you win or lose; it's whether or not you had a good bet.
Wizard
Administrator
Wizard
Joined: Oct 14, 2009
  • Threads: 1248
  • Posts: 20591
Thanks for this post from:
charliepatrick
August 12th, 2019 at 7:59:59 PM permalink
Quote: charliepatrick

I'm still getting a slightly different answer but haven't got the hang of catering for peeking. For instance I guess you know there's a slightly higher chance of a picture when the player is playing their hand if the dealer with an Ace has already peeked. I'm getting 1.98% so there's probably a small bug in my program or logic somewhere.



I'd be happy to discuss the peeking math with you. Maybe I'll make a separate thread for it. I pretty much use a Bayesian approach: EV(hand | no dealer BJ) = EV(hand and no dealer BJ)/Prob(no dealer BJ). However, its easier said than done.

I'm pretty confident in my figure. I hope he won't mind me saying, but CrystalMath agreed with it and Charles is 0.01% off (or we're off from him).


Quote:

Anyway this is the playing strategy it gives (so please accept there might be some small errors in it) but it must be fairly close to normal single deck strategy.

Soft 19Double A8 vs 6
Soft 18Double A7 vs 3-6Hit AA24 vs 9
Hit AAA5 vs A 9 10
Hit Axx vs A 9 10
Hit A7 vs A 9 10
Soft 17Double A6 vs 2-6
Hard 16Some hit, some don't
Soft 16Double A5 vs 4-6
Hard 15Stand a few vs 10
Soft 15Double A4 vs 4-6
Hard 14Stand 77 vs 10
Soft 14Double A3 vs 4-6
Hard 13Hit A2T, 49, 3T vs 2
Soft 13Double A2 vs 4-6
Hard 12Stand some vs 3
Hit 2T vs 4
Soft 12Double AA vs 5-6
Hard 11Double
Hard 10Double vs 2-9
Hard 9Double vs 2-6
Hard 8Double 44 or 35 vs 5-6
btw one advantage of giving the players their five cards is their hands cannot be affected by other players standing or hitting. An easier way is to deal the cards out as you go, but face down if the player has already stood.



Thanks. I wrote a combinatorial program, so can't comment, other than your strategy looks very reasonable out of its similarity to single deck H17 blackjack.
It's not whether you win or lose; it's whether or not you had a good bet.
CharlesMousseau
CharlesMousseau
Joined: Nov 1, 2016
  • Threads: 3
  • Posts: 26
Thanks for this post from:
tringlomane
August 12th, 2019 at 8:09:37 PM permalink
Quote: Wizard

Thanks for chiming in Charles. Glad to hear you don't deserve any blame for the Twisted Stud fiasco.

Too bad Dan Lubin isn't around. I would have loved to see his comments on this game.



Heck, I half expected you mentioning the potential of a player edge or the "don't correct the dealer if they're mispaying in the player's favor" comment to bring him back from the grave :)

I miss that little lowering-the-head gesture when he got on a roll about one of his pet crusades.

Incidentally, if you've ever watch the GLOW TV show on Netflix, there's a character (Sam Sylvia) that is an absolute dead ringer for him. Like I watched the show with the good lady and when I saw him she had to pause the show cause I was almost on my knees with laughter.
charliepatrick
charliepatrick
Joined: Jun 17, 2011
  • Threads: 30
  • Posts: 1781
August 12th, 2019 at 11:38:02 PM permalink
Quote: Wizard

I'd be happy to discuss the peeking math with you. Maybe I'll make a separate thread for it. I pretty much use a Bayesian approach: EV(hand | no dealer BJ) = EV(hand and no dealer BJ)/Prob(no dealer BJ). However, its easier said than done....

Thanks - I agree a thread would be good, it's fun trying to develop a method of working it out.

It's a new program for me and I haven't yet checked the 999 decks against an infinite deck spreadsheet, but it's probably my method to cater for BJs that's wrong. Personally I did use something like EV(player|no DBJ)-Pr(DBJ)+Pr(DBJ)*Pr(PBJ)|Pr(DBJ) but suspect there's a small bug somewhere.

Interestingly there was a -.008% difference taking account in the cards left in the deck because of peeking. For instance opposite and Ace my original method when picking the chances of player's hands was to assume the deck had 3 Aces, 4 of everything else and 16 tens (minus what the player already had), but actually shouldn't it be 3-1/9 Aces, 4-1/9 everything else and 16 tens as you know the dealer has a non-10.

As you say it's complicated and getting what seems a viable strategy implies I'm on the right course, although I was surprised you hit 3-card soft-18 vs A.
superbri
superbri
Joined: Apr 27, 2011
  • Threads: 3
  • Posts: 10
August 16th, 2019 at 5:37:02 AM permalink
Thank you for your excellent work. However I do not know how a strategy can be developed until the dealer and pit bosses are being consistent. eg the earlier example re 3:1 or twisted stud @ 20:1.
Wizard
Administrator
Wizard
Joined: Oct 14, 2009
  • Threads: 1248
  • Posts: 20591
August 16th, 2019 at 6:15:49 AM permalink
Quote: superbri

Thank you for your excellent work. However I do not know how a strategy can be developed until the dealer and pit bosses are being consistent. eg the earlier example re 3:1 or twisted stud @ 20:1.



The strategy is for the base game. If the player always gets to use all five cards for the Stud Bonus, that bet has no affect on the strategy.
It's not whether you win or lose; it's whether or not you had a good bet.
charliepatrick
charliepatrick
Joined: Jun 17, 2011
  • Threads: 30
  • Posts: 1781
August 16th, 2019 at 7:08:26 AM permalink
Quote: CrystalMath

....I am now at 1.925% with 1 deck....

Quote: Wizard

...After correcting for that, we agree....

At last I've got 1.9248% but it needed to cater for peeking by adjustments such as 3/35 4/35 when working out which hands the player can make by hitting/doubling.
Interestingly the sim got 1.974%-2.005% for single deck shuffling after every hand, slightly higher if you deal multiple hands from each deck, up to 2.72% for 999 decks.
skyscanner
skyscanner
Joined: Aug 12, 2019
  • Threads: 1
  • Posts: 5
September 21st, 2019 at 6:48:03 PM permalink
I obeserved the game last week and another weird situation came up.

Player got dealt 10 and 2 against the dealer's 5, so he stayed. The dealer only had small cards and his total from the 5 cards was 16, meaning that he did not hit soft 17. They decided that the player lost the hand, as 12<16, and took his money.

  • Jump to: