Wizard
Joined: Oct 14, 2009
• Posts: 23626
July 23rd, 2020 at 12:11:40 PM permalink
Quote: ThatDonGuy

The RTP is 99.5376324%

Not to say this is wrong, but the following table shows what I'm getting. My EV equates to an RTP of 96.7941%. Any thoughts about something I may not be seeing. I've gone over my code over and over, rewriting it, and I always get the same bottom line.

The only thing I can think of is my program never re-rolls 4 or all 5 dice. Your strategy says not to do that anyway. I'm stumped.

Event Pays Combinations Probability Return
Five of a kind on first roll 9 362,797,056 0.000772 0.006944
Five of a kind 3 18,570,954,240 0.039497 0.118491
Four of a kind 1.5 86,360,256,000 0.183673 0.275509
Straight 1 33,457,950,720 0.071159 0.071159
Full house 0 94,383,221,760 0.200736 0.000000
Loser -1 237,049,804,800 0.504163 -0.504163
Total 470,184,984,576 1.000000 -0.032059
It's not whether you win or lose; it's whether or not you had a good bet.
charliepatrick
Joined: Jun 17, 2011
• Posts: 2495
July 23rd, 2020 at 2:03:29 PM permalink
I agree with the paying figures except I get 39 907 676 160 straights, so it looks as if you've lost some after the 1st roll with hands like KKQJT (i.e. Pair and Easy Straight) 622080 and half the no pair straight draws (AKQJ9 etc) (414720/2).
Wizard
Joined: Oct 14, 2009
• Posts: 23626
Thanks for this post from:
July 23rd, 2020 at 2:42:31 PM permalink
Quote: charliepatrick

I agree with the paying figures except I get 39 907 676 160 straights, so it looks as if you've lost some after the 1st roll with hands like KKQJT (i.e. Pair and Easy Straight) 622080 and half the no pair straight draws (AKQJ9 etc) (414720/2).

Thanks Charlie. Stupid question -- the video shows the game being played with dice numbered 1 to 6. However, you refer to poker dice (9 to A). This would't make any difference would it? Aces high only, right?

My program just loops and makes the best of the 26 plays that re-roll 0 to 3 dice (never 4 or 5). I coded no strategy decisions in it.
It's not whether you win or lose; it's whether or not you had a good bet.
charliepatrick
Joined: Jun 17, 2011
• Posts: 2495
July 23rd, 2020 at 3:00:05 PM permalink
When I saw the puzzle I just thought of poker dice and stuck with the idea, I found it easier to visualise AKQJ9 than 65431.

What I originally used was a simple spreadsheet that just looked at EVs and percentages of various options. It looked at the chances of drawing different types of hands given your current set of dice. But it was easy to work out how many combinations this meant at various stages given the strategy. Luckily, unlike poker with cards, there's no memory of the discards since the dice are rerolled.
Wizard
Joined: Oct 14, 2009
• Posts: 23626
July 23rd, 2020 at 4:12:55 PM permalink
Charlie, not to doubt you, but has anyone else verified your number? I have seen MattUK quoting it, but Matt, can you confirm it?

I would like to starting writing a page on the game and thinking about posting both of our figures for now.
It's not whether you win or lose; it's whether or not you had a good bet.
MattUK
Joined: Jul 25, 2017
• Posts: 138
July 23rd, 2020 at 4:35:02 PM permalink
Gamesys Group says that the RTP is 99.54%. As you know, the British market is heavily regulated with 7 digits penalties for failing to play by the rules. That's why I have a good deal of confidence that Charlie has verified that independently.

***

The expected return is the amount we pay out to players relative to the amount of wagering on the game.

For example, if we take £100 of wagers we will, on average, pay out £99.54 of wins. This return is based on the player making the best choices.
CrystalMath
Joined: May 10, 2011
• Posts: 1907
July 23rd, 2020 at 4:36:50 PM permalink
Quote: Wizard

Charlie, not to doubt you, but has anyone else verified your number? I have seen MattUK quoting it, but Matt, can you confirm it?

I would like to starting writing a page on the game and thinking about posting both of our figures for now.

I get 99.5376324% also.
I heart Crystal Math.
charliepatrick
Joined: Jun 17, 2011
• Posts: 2495
July 23rd, 2020 at 4:38:28 PM permalink
It was quoted earlier, but I think if you can agree with my number for straights, you'll get the same figure. I cannot be certain that I have the correct strategy, but it sounds as if your method would have proved it. Personally I'd put the EVs in EV order with a note about splitting a FH on the first round, which pairs to split from Two Pair, and with a pair always going for an outside straight draw and an inside straight draw on the second round etc.
ThatDonGuy
Joined: Jun 22, 2011
• Posts: 5072
July 23rd, 2020 at 4:41:15 PM permalink
Quote: CrystalMath

I get 99.5376324% also.

As did I.
Wizard