Statistic

Hello!

Id like ask:

How many approximately hands/shoes enough for make some conclusion about statistic events.

Thanks so much!

Quote: donik

Hello!

Id like ask:

How many approximately hands/shoes enough for make some conclusion about statistic events.

Thanks so much!

Hi, donik. I'll throw out 50,000 hands because that's what I do - throw things out. That should do it but there are oh so many variables. The N0 formula deals with this and should be easy to look up. It's pronounced N-Zero. Enjoy!

Thank so much!
My opinion 500 shoes enough for make some conlusion.

Best regards.

Quote: 1BB

Hi, donik. I'll throw out 50,000 hands because that's what I do - throw things out. That should do it but there are oh so many variables. The N0 formula deals with this and should be easy to look up. It's pronounced N-Zero. Enjoy!

I believe it's pronounced N-Nought (or N-Naught).



As for determining how many rounds you need to figure it out if your results are statistically significant....that's largely dependent on the game you play. If you're playing a game like Video Poker, 50,000 hands is nothing. You could hit 0 Royal Flushes (which are usually worth about 2% of your return) or you could hit a bunch. There's high variance in video poker. If you're playing a game like blackjack, 50,000 hands should be statistically significant -- but that doesn't mean there isn't any variance. If you're playing some hypothetical game (like say, a coin-flip: heads or tails), you don't need very many rounds at all.

If you're wanting to figure out if a game is "fair", it also depends on how far away your results are from expectation. For example, if you're doing a coin flip and after 100 flips there are 45 tails and 55 heads, you are not that far away from normal expectation (50 heads 50 tails). If you've done 10,000 flips and you have 45,000 heads and 55,000 tails.....now it's looking a little more suspicious, even though the percentage or ratio is the same.

But say after 100 flips, you have 20 heads and 80 tails....even though you only have a small sample (100 flips) that is far away from normal expectation and variance.

If you're playing blackjack (flat-betting the same amount) and you're down 20 units after 100 hands....that's relatively normal and within expectation/variance. If you're down 20,000 units after 100,000 hands....that's very un-normal and not within expectation. Same percentage/ratio in both cases, but have to take into account the # of hands.


So, it's a combination of # of rounds AND how far away you are from expectation.



PS: You can't ever make a definitive conclusion about your results and what you can expect in the future. If you're trying to "invent a system" and are looking for patterns in your results -- you're going to find patterns.

I could flip a coin a thousand times and record each result. If I'm looking for a pattern, I might realize some pattern such as:

Whenever there are three HEADS in a row, 90% of the time the 8'th flip from the third head is going to be a tails.


That just means there is a pattern in previous results. It has no indication on future results.

The original question referred to hands/shoes so I took it as a blackjack question. It's been N-zero for as long as I can remember.

For blackjack, I have always sort of informally used 50,000 as well. In one of the books about the MIT team, there was a reference to "law of large numbers", and there was a loose definition of 50,000 rounds. This number has always worked for me as I play 75,000-80,000 rounds per year and my numbers have always been fairly close to or above expectation. Last year was the biggest difference, when my actual results came in at 40 grand or just about 50% above expectation. That was a good problem to have and I didn't think too much about it. Lol.

Now this year, and I have just gone above 80,000 rounds for the year in the last few days, I am 42 grand or just more than 50% BELOW expectation. And for the first time, I am thinking that 50,000 doesn't really work all that well for my situation. It's probably more like 100,000+ or even 150,000. That would basically be 2 years worth of play for me.

I think my situation is a little unique because I play a variety of different games with different spreads and max bets (based on what I perceive as the tolerance or comfort level of each individual place). So in reality my 80,000 rounds a year is broken up into smaller segments. I might have 8,000 at a double deck game spreading up to $400, 10,000 rounds at double deck only spreading to $300, 10,000 a 6 deck spreading $50-$600 and 26,000 rounds at 6 deck spreading $25-400. That is a total of 54,000 rounds, but it really isn't. It's really 4 different situations with totals ranging from 8000-26,000.

The fact that I wasn't using a large enough N0 (and it is N-zero, as 1bb stated) was one of the big things I have learned this year and learned to deal with. Even in my 11th year of supporting myself from majority card counting AP, every year is a learning experience and opportunity. :)

What about baccarat shoes?
500 shoes is enough?

Appreciate you!

Quote: donik

What about baccarat shoes?
500 shoes is enough?

Appreciate you!

No matter how many baccarat shoes you analyze, it's never enough.

^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

LOL!!!