Quote from: shorena on May 11, 2015, 06:03:05 PM

I hat 22x in a row in the past and IIRC just-dice had a 32 losing streak for martingale.

See https://bitcointalk.org/index.php?topic=610339.msg7060900#msg7060900 for details of that.

In summary:

The longest winning streak at 49.5% (2x) at Just-Dice is 30 in a row.
The longest losing streak at 49.5% (2x) at Just-Dice is 32 in a row.

Obviously the chance of winning changes things drastically. If you're playing with just 1% chance of winning, it's not unusual to see losing streaks of a few hundred bets in a row.

Quote from: shorena on May 11, 2015, 06:03:05 PM

As PenguinFire said there is no maximum. Its very unlikely to lose 22, 32, 42 or 662 times in a row, but eventually it will happen.

I'm not sure about 662. That will almost certainly never happen, due to limitations on the amount of energy in the universe, etc.

Consider that you're about as likely to break into a random Bitcoin address (a 160 bit number) by brute force as you are to roll 160 in a row (at 50%). Then consider that to roll 640 in a row means doing that 4 times in a row without any failures in between. Then consider that 662 is even bigger than 640.

Quote from: XinXan on May 11, 2015, 07:02:13 PM

30 loses in a row is 1 in a billion chance, there are reports of a little bit more but not much more since no one would have enough bankroll to sustain more than 40 loses in a row even if you start with 1 satoshi bet

Even if you flat bet, or bet 0 on each roll, you still won't see much longer than a losing streak of 30 at 49.5%. It's not because people run out of bankroll (they typically just drop their bet right down and keep playing, just to see when they "would have won". It's because very long streaks are incredibly rare.

Quote from: notlist3d on May 11, 2015, 09:22:23 PM

They are free to set their OWN odds. So .... no regulation and they set their own odds. They might be turning it up and down you don't know.

Anyways there is no max you could lose, you could lose it all. And the house always wins, especially when they program their own rules.

All good Bitcoin gambling sites are provably fair. You can check that they aren't cheating. Regulation doesn't stop sites cheating anyway. It just stops them cheating when the regulator is looking.

Quote from: kattekop on May 12, 2015, 06:54:09 AM

One really shouldn't gamble, just ever. You're playing against 'invoke GetTickCount' from Windows (random number generator), and it's really not the way to go.

First off, it's really quite dangerous for the casino to randomize their rolls using something as predictable as a system clock. It's possible someone will be able to exploit it and predict their rolls with enough accuracy to beat the house edge.

And secondly, all decent Bitcoin gambling sites are provably fair. That means they can prove that their rolls are made fairly using techniques like publishing the hash some seed used to generate the roll before you play, so you can check that they committed to your roll before you made it. You'll typically be playing against sha256 or similar rather than a random number generator.

Quote from: tspacepilot on May 15, 2015, 02:59:45 PM

Quote from: a1choi on May 15, 2015, 06:19:21 AM

before JD shut down the first time, Doog did some stats on longest losing and winning streak at 49.5%.

I believe loses was as big as 32 in a row, and wins was 19 or 20.. i'm not 100% sure of the stats. its in the old JD thread somewhere.

Remember that this isnt simple 50/50. at 49.5/50.5, the max loss streaks will be longer than the max win streaks. (thus you can see the house edge in action, sortof)

RIght, that's actually an interesting point, a1choi, the longer the streak you're multiplying out, the more you're going to see that divergence between wins vs. losses multiplied out and getting bigger. Ie, as you say, you can observe the house edge in action.

In the analysis I did (linked at the top of this post), the 5 longest winning streaks at 49.5% were four 27's and a 30, while the 5 longest losing streaks at 49.5% were 32, 27, and three 26's. It's interesting that you don't really see the house edge in operation there, in that the winning streaks were longer than the losing streaks "on average". (I know, you can't average streak lengths like that...):

Quote

W L
30 32
27 27
27 26
27 26
27 26