Wizard
Joined: Oct 14, 2009
• Posts: 16474
June 15th, 2017 at 2:38:09 PM permalink
Good stuff! May I quote this strategy on my site? If you grant permission, let me know who you would like to be known, as I always give credit where credit is due.
It's not whether you win or lose; it's whether or not you had a good bet.
Ayecarumba
Joined: Nov 17, 2009
• Posts: 5393
June 15th, 2017 at 2:52:45 PM permalink
Quote: gordonm888

Okay, we have simulated the Discard Inferno game using a rough (non-optimum) player strategy -and using the Blind Bet payout table that we had been previously unaware of.

We calculate a House Edge of 1.624 % and I fully expect we would get the 1.600% House Edge if we put in a more complex Basic Strategy for the player.

Basic Strategy in Calculation of HE=1.624%

Always BET 3X with 3 of a Kind and Better

For other player decisions, it depends upon whether the dealer upcard is low (2-8) or high (9-A)

vs 9-A
Bet 2X with a 33 pair or higher
Bet 1X with any K-High or higher
Fold with Q-High or lower

vs 2-8
Bet 2X with JJ-AKQ or higher
Bet 1X with any 44 pair or higher vs 3-8
Bet 1X with any 22 pair or higher vs 2
Otherwise FOLD

The above strategy for when to BET 1X vs 2-8 can be improved significantly and would probably reduce the House Edge to something very close to the value of 1.600% that is said to be in the math report.

Why would 22 vs. a 2 be okay to raise, but not 33 vs. a Dealer 3?

My initial thinking would be to raise 1X vs. 2-8 if your pair is the same or higher than the Dealer's upcard (and maybe even if you have an Ace+face card and one of the other three of the dealer's upcards as a singleton).

How does a 44 vs. a dealer 8 generate more winners than losers? You will definitely face a 7 card hand in that situation.
America is all about speed. Hot, nasty, bad-ass speed. - Eleanor Roosevelt, 1936
CrystalMath

Joined: May 10, 2011
• Posts: 1491
June 15th, 2017 at 7:18:52 PM permalink
Quote: Ayecarumba

How does a 44 vs. a dealer 8 generate more winners than losers? You will definitely face a 7 card hand in that situation.

It only needs to win 20% of the time to make it worth playing vs folding. If you fold, you are folding 2 units, so your EV must be better than -2.

Assuming a 20% chance of winning, EV=0.2*2 - 0.8*3 = -2.
I heart Crystal Math.
gordonm888
Joined: Feb 18, 2015
• Posts: 606
Thanks for this post from:
June 16th, 2017 at 7:57:26 AM permalink
Quote: Ayecarumba

Why would 22 vs. a 2 be okay to raise, but not 33 vs. a Dealer 3?

My initial thinking would be to raise 1X vs. 2-8 if your pair is the same or higher than the Dealer's upcard (and maybe even if you have an Ace+face card and one of the other three of the dealer's upcards as a singleton).

How does a 44 vs. a dealer 8 generate more winners than losers? You will definitely face a 7 card hand in that situation.

You are definitely correct on 33 vs a 3 - and I was going to change that for the strategy that I would suggest that the Wizard post. At the time of the calculation I quoted, we were just using an approximate strategy while we debugged our simulator code and we were lazy about how many rules we put in the code. The rules for 22 vs 2 and 33 vs 3, etc. can only improve your EV by a small amount because those hands are relatively infrequent. But you are definietly correct, I think it is better to hit 33 vs 3, than to fold it.

Ignoring ties, you can think of the the criteria for BET 1X as: player hand must win 20% of the time. Given a fresh deck, The dealer's 7 card hand will be No Pair/High Card about 16.8% of the time, and with an 8 upcard the player needs a pair of 4s to be able to beat the dealer at least 20% of the time..

The criteria to BET 2X is that the player must win 50% of the time (ignoring ties) because that makes the Play bet a positive EV bet.
Last edited by: gordonm888 on Jun 16, 2017
gordonm888
Joined: Feb 18, 2015
• Posts: 606
June 16th, 2017 at 8:11:04 AM permalink
You may quote the strategy and use my real name: Gordon Michaels. It wiould be an honor to be credited for something on the WOO site -sort of a bucket list goal for me, lol.

Per the above post, you might wish to consider changing the BET 1X threshold vs 3 to be a pair of 3s, rather than a pair of 4s. We have simulated Discard Inferno with that change in strategy over 10 million trials and get about the HE that I quoted previously and a microscopically higher HE with Fold 33 vs 3.
Wizard
Joined: Oct 14, 2009