BJ Simulations
  • Now, this is kind of serious question. How the outcome of a split in which you lose one and win the other should be recorded? Is there a winning hand and a losing hand or it should be recorded as a regular push hand? How about if you lose both hands? Do you have two loses or a double loss, like when you double and lose? .. How about you win both splits? - Do you have two winning hands or a double win? .., Or, how about you got 16 vs. A and the Tc=5. You bet max and also get the insurance. The dealer’s got the ten in the hole so you lose the bet but win the insurance. How this one should be recorded? A total push or a win on one action and a lost on the other - like WL?

    Well, it depends how you trace these things because it impacts the final percentages of how many winning/losing hands you get or how many push .. etc ..

    Why I’m asking this is because I have an old DOS simulator that it shows a higher percentage of pushing hands when I use the 10 count for insurance. I get like 9.87% for push. So I figure probably it trace the winning insurance vs. losing the main bet as a win-loss = push

    Is that possible?
  • AlexD30 said:
    .....

    Why I’m asking this is because I have an old DOS simulator that it shows a higher percentage of pushing hands when I use the 10 count for insurance. I get like 9.87% for push. So I figure probably it trace the winning insurance vs. losing the main bet as a win-loss = push

    Is that possible?

    i can't answer your question..... but i have a question.
    what is the name of your old DOS simulator?

    best regards,
    mr fr0g
  • AlexD30 said:
    Now, this is kind of serious question. How the outcome of a split in which you lose one and win the other should be recorded? Is there a winning hand and a losing hand or it should be recorded as a regular push hand? How about if you lose both hands? Do you have two loses or a double loss, like when you double and lose? .. How about you win both splits? - Do you have two winning hands or a double win? .., Or, how about you got 16 vs. A and the Tc=5. You bet max and also get the insurance. The dealer’s got the ten in the hole so you lose the bet but win the insurance. How this one should be recorded? A total push or a win on one action and a lost on the other - like WL?

    Well, it depends how you trace these things because it impacts the final percentages of how many winning/losing hands you get or how many push .. etc ..

    Why I’m asking this is because I have an old DOS simulator that it shows a higher percentage of pushing hands when I use the 10 count for insurance. I get like 9.87% for push. So I figure probably it trace the winning insurance vs. losing the main bet as a win-loss = push

    Is that possible?



    I don't think there is any "standard" to apply here. So far as I know, the ISO hasn't delved into blackjack simulations yet. That aside, my own simulations have treated splits as separate hands. That's for convenience since each bet becomes a separate entity with separate playing decisions and the like.

    But if you are trying to produce a B/S chart, the initial "pair" that you split leads to two hands, but the result of that "split" decision has to be the total of what happens to both hands, or it won't be meaningful.

    My only interest in simulations has been to answer the question "should I split or not at this particular true count?" and that has to based on the total EV for that round, which includes the sum of the EVs for each hand resulting from the split.

    In short, the first question has to be "what are you trying to determine with your simulation?" If you are trying to match your DOS results with something more sophisticated like CVData, I wouldn't even waste the time...
  • stainless steel rat said:
    .......In short, the first question has to be "what are you trying to determine with your simulation?" If you are trying to match your DOS results with something more sophisticated like CVData, I wouldn't even waste the time...


    The question I asked in my post was how push hands should be recorded?
    For example: If I have A9 vs. A while the index strategy require insurance based on the TEN count. The ratio of NT/T < 2. OK, let’s say the dealer got the ten in the hole this time but I’m not losing any money! - I won the insurance but lost the original hand. How do I record this fact? – Do I have a lost of the original hand but at the same time I have a win on the insurance bet? Do I record a L, and a W or just P for a push into the statistics?

    Why I’m asking this is because I noticed that when I play the Ten count, NT/T, that has 100% insurance correlation I’m getting a larger percentage of push hands. Also, playing 12 vs. any up card it seems to be more correlated with the ten-count than with Hi-Lo or any other count for that matter. I honestly, suspect that by playing the ten-count with flat bet you simply can bring the house edge to zero just by 100% insurance correlation and having a more accurate play on your 12s. The ten-count tracks the density of T vs. all other cards from the deck.
  • insurance is irrelevant. It is a "side bet". the simulation should be recording two separate values, one for what happens when playing the hand itself, one for what happens on the insurance side-bet.

    The results for your playing against an A-7 or 7-A for the dealer must be the same overall or the results will not be meaningful.. yet for the A-7 you would have an insurance decision while against the 7-A you would not (first card is up).

    Insurance is important for your overall EV. But it should be added in differently than normal hands. When the sim is done, EV = EV of playing decisions + EV of insurance decisions. This latter comes from the winning percentage of insurance bets times the probability of the dealer having an A on the up-card.
  • For the purposes of overall calculations, all bets and results in one round, including multiple hands, Insurance, Splits, DD's, and side bets should be recorded as one play. This is required for the correct calculation of standard deviation, and therefore risk.

Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!