2 billion round simulation? Is that enough of rounds?

[iCountNTrack]

[blackjack avenger]

Joking

If one were simming rare indicies 2 billion may not be enough?

[aslan]

My Excel spreadsheet simulator has been running for 2 months, 14 days, and a little over 3 hours now, but I'm sure it will soon give me a statistically relevant answer to my last query, that is, if I entered the data right. It's a little tricky the way I have it set up. Someone suggested I give in and get Qfit's blackjack simulator. What!? And miss all the fun?!

[k_c]

If I understand correctly you are enumerating the number of possible card combinations that could possibly lead to an outcome for any player strategy for a round of blackjack for 1 player versus dealer.

For what it's worth there are 3072 possible unbusted player hands of 2 or more cards. If dealer stands on soft 17 there are 1677 possible dealer hands. If dealer hits soft 17 there are 1740 possible dealer hands. As number of decks becomes more limited there are fewer possibilities than above. Therefore for a single unsplit round there are up to 3072*1677 = 5,151,744 possibilities for s17 and up to 3072*1740 = 5,345,280 possibilities for h17.

If I am not wrong there are 15,450 player hands of 2 or more cards if busts are included.
If 2 hands are dealt back to back to account for pair splits there are (15,450)^2 = 238,702,500 hand combinations.
If the condition of both hands totaling > 21 is eliminated, that would leave 85,827,635 hand combinations to account for a single split.
Therefore in order to account for 1 possible split there are up to 85,827,635*1677 = 143,932,943,895 possibilities for s17 and up to 85,827,635*1740 = 149,340,084,900 possiblities for h17.

Adding more splits would increase number of possibilities exponentially.

fwiw that's how I would answer the question and I don't have any opinion or knowledge of what the statistical significance may be. I am sure of the numbers for a single hand and less sure of the numbers for a single split. Also I am sure of the number of possible dealer hands.

[sagefr0g]

Originally Posted by aslan

My Excel spreadsheet simulator has been running for 2 months, 14 days, and a little over 3 hours now, but I'm sure it will soon give me a statistically relevant answer to my last query, that is, if I entered the data right. It's a little tricky the way I have it set up. Someone suggested I give in and get Qfit's blackjack simulator. What!? And miss all the fun?!

you really have an excel simulator? i gave up trying to figure out how to heck to handle splits, with the one i tried to make.
but yeah, the other thing was, if it ever was created it would have to run for ever, also i just didn't think it could contain all the data.

[zengrifter]

Originally Posted by iCountNTrack

This where algebraic approximated indices come in handy!

[QFIT]

Originally Posted by blackjack avenger

If one were simming rare indicies 2 billion may not be enough?

Creating indexes is a very different problem. And no, 2 billion would not be enough.

[zengrifter]

Originally Posted by aslan

My Excel spreadsheet simulator has been running for 2 months, 14 days, and a little over 3 hours now, but I'm sure it will soon give me a statistically relevant answer to my last query, that is, if I entered the data right. It's a little tricky the way I have it set up. Someone suggested I give in and get Qfit's blackjack simulator. What!? And miss all the fun?!

Just re-build the Excel for algabraic approximation using the Griffin formula.
Excels over-heat running sims, could start a fire. zg

Here >>

Jensen Algebraic Index Calc (ZIP D/L EXCEL)
http://www.bjmath.com/bjmath/tcindex/Generator.zip

[zengrifter]

Originally Posted by QFIT

Creating indexes is a very different problem. And no, 2 billion would not be enough.

How many sim hands would it take to show a difference between extensively simmed indices and those approximated by algebra? zg

[QFIT]

Originally Posted by zengrifter

How many sim hands would it take to show a difference between extensively simmed indices and those approximated by algebra? zg

As many as it would take to calculate correct indexes, instead of algebraic estimates. Which varies enormously depending on the index, tags, methodology and other variables. Unlike BJ simulation, index generators do not run a preset number of hands, which is why you see some indexes pop up quickly and others grind away for long periods.