Topic: How 999dice.com is stealing your coins, and exactly why you won't believe me - page 6. (Read 41936 times)

That depends if you are into traditional art or modern art which throws out all the rules to wake up the sheeple.  

Thanks for the info.
Haiku seems pretty easy.
I was worried there may have been some kind of silly rule about syllable counts.

haiku is traditional
and short form of
Japanese poetry.

What is a haiku?
I'm just splitting my text up
like Arnab above.

meh its ok.  I just couldnt think of anything funny to say so I thought a pun would have to do.

Arnab went as far as to use a table to lay out his weird post. I've no idea why anyone would do that.

haiku or hai about no?

People win
against
casinos
every
day.

The trick is
to quit while
you're ahead.

If you play
long enough
you will lose
in the end.

Don't be so sure.

On Just-Dice, at 49.5% chance of winning, we have seen a winning streak of length 30, and a losing streak of length 32.

Those are 1-in-3-billion and 1-in-1.5-billion chances respectively.

Billion-to-one shots do happen. (Especially when you have over a billion bets on the site)...

>>> 1 / 0.505**32
3,123,747,346

>>> 1 / 0.495**30
1,451,590,214

Also, are you sure about your math there? If you have only a 1-in-1.28 billion chance of losing, then you can expect to win 1 satoshi around 1.28 billion times before you lose your 11.15 BTC.

There are 100 million satoshis in 1 BTC, so 1.28 billion satoshis is 12.8 BTC, and so your expectation is positive (win 12.8 BTC for every 11.15 BTC you lose).

that's the point..we can't. win aginst them but.they are still addicted to some.techniques who.they belive will work aginst them

As I pointed out when  I did the same math and came up with the  same numbers that assumes their RNG is doing an even distribution.  I also pointed out that I do not believe that it is.  It appears (without doing enough statistical charting to validate my opinion) that there are times where that is much more likely and times where that is much less likely. 

The fact that they double hash also reduces the number of outputs.  If you have all the possible values a  hash will be a subset of that.  A double hash requires the input of the first hash so its outputs are even more limited.  I believe this plays into part of why their RNG appears flawed but I have not done the math to confirm it.  I still think the original RNG is flawed and if they are using MSSQL for rand() as they do for validating the bet (per their site) then its known to be flawed.  Its deterministic based in part of the C call  time(). 

My second post in this thread goes into the numbers much more deeply.

When running simulations, at a 49.95% bet (999dice's 2:1 odds number, given the house edge), I repeatedly saw instances where the roll was failed over 350 times IN A ROW.

He's the simple math of it - if you lower the base bet to a number small enough that you can cover-by-doubling your losses to an "impossible to lose" point, you are going to have to bet enough times to make any significant winnings that the "impossible to lose" odds are suddenly a very real reality.

Lets use 999dice's numbers: Betting 95% win gets you x1.05157 your bet on a win. To even win anything, you'd need to bet at least 20 satoshi, as anything less than 20 satoshi results in winnings less than 1 satoshi.

To recoup your losses on 999dice, you need to bet 19.38x your last bet.
20 sat bets
After 1 loss: 387 satoshi
2? 7500 satoshi
3? 145535 sats
4? .0281 BTC
5? .546 BTC
6? 10.581 BTC

That's all we're willing to do. Not lose more than that.


What are the chances?

To lose once: 5% - 1 in 20
Lose twice in a row: .25% - 1 in 400
Three times: .0125% - 1 in 8000
Four? .000625% chance - 1 in 160,000
Five times in a row? .00003125% chance - 1 in 3.2 million
Six? .0000015625% - 1 in 64 million
And finally... 7... It's a 1 in 1.28 BILLION chance to lose that 7th roll. Impossible. Even the lottery is easier to win.

So, on any given roll, you are risking about 11.15 BTC to win 1 satoshi. Of course your chances of losing are 1 in 1.28 BILLION, so it's "impossible".

But winning 1 satoshi is hardly worth the effort. Whats your goal? Lets be really safe and say we just want .25 BTC. To get that, we'd need to roll slightly over 25 million times. Slightly over, because 5% of all rolls will be a loss, and it's a wasted roll, so we'll need to roll the recovery roll. So...

25 million - 5% is:
1.25mil re-rolls, and 23.75 million wins.
5% of the 1.25 million rerolls will also be losses, statistically speaking. So:
1.25mil 1st rerolls
62500 2nd... oh and 5% of those...
3125 3rd rerolls... and 5% of those...
156 (rounding down) 4th... and 5% of those...
7 (rounding down) 5th rerolls, and 5% of those...
well, rounding down, thank god we don't lose any of the 6th 10.581 BTC rolls.

So, we have 1,315,788 wasted rolls to losses, and 23,684,212 wins. We lost 5.55% of the rolls to losses, so lets up that to  getting us .2368 BTC in winnings. So we need to roll 26,470,432 times to, in theory, get .25 BTC.

Thats 26.47 million chances to roll that 1 in 1.28 billion. Whats that work out to? 1,280,000,000 / 26,470,432?

48. 1 in 48 odds to fail your roll 7 times in a row. 1 in 48 is a 2% chance. Would you risk 11.15 BTC to win .25 BTC?

If you are, save yourself a lot of time, and go bet 11.15 BTC over at 98% odds. This is where part of my example does not add up, as 999dice doesn't allow 98% bets. While Primedice DOES, their edge is 10x as large, making the math not work out. But lets say 999dice DID allow 98% bets. It would pay out 99.9/98 which is 1.01938. So, you'd get 1.938% of your risk back as winnings.

Whats 1.938% of 11.15?

.216.

Damn close to the .25 you were trying to get (the house edge breaks this example a bit - if you had no edge at all, the numbers would be much closer.) And with only ONE chance to lose it. If you did the 25 million bets, while its a 1 in 48 chance to lose it, there's also a very small chance it happens 2 or more times. Using some complicated binomial probability distribution mumbo jumbo, while there's only a 2% chance you lose 11.15 BTC trying to get .25 BTC, there's also a .000456% chance - 1 in 219k chance it happens 2 or more times. Yes, really small, but you could lose even more than 11.15 BTC.

Moral of the story: How many times can a 5% loss chance lose in a row? An infinite number. It just depends on how many chances you give it to happen.

Moral 2: Don't try and beat math. She's a mean mofo.

Neither. The "correct" scenario is:

[...]Site owner responds: "Thanks for pointing out this serious flaw in our provable fairness. We will fix it as soon as possible so that every player can easily verify that their rolls are fair."

Yes, I implied it. Oddly enough, I do not need proof to imply it or state my opinion.


So by default, no one is guilty of anything if it requires any degree of smarts to pull off, because everyone is too lazy and stupid? Thats a bit narrow for a lawyer. Just strikes me odd. While, yes, people are stupid and overlook things, I think that is EXACTLY what "Jake" was hoping for. To be thought of like you are. Massive benefits of the doubt because who would go to that amount of effort?

Well, I think someone who appears to have profited 2,000 BTC in a single year would go to that amount of effort.

Anyone here not willing to work extra hard for a bonus half a million $$$?

And at the time his site went up, that would have been $3 million. Oh I'd be one hard working sneaky sonofabitch for $3,000,000

That is the thing and with the max win per wager as most sites do even a martingale strategy will eventually fail, assuming you do not first run out of your bankroll.  Even with a 95% chance of winning there is still provable statistical odds of a long string of consecutive losses.  The site profit would increase dramatically if people do not have a large enough bankroll.  The odds of 2 consecutive losses would be 0.05*0.05 = 1:400.  4 consecutive losses 1:160,000.  6 consecutive losses 1:64,000,000.  Lotteries in the US are often 1:175,000,000 for the grand prize (powerball, a multi-state lottery is exactly that).

With 95% chance to win you have to have almost 1900% wager increases to win back enough to cover previously lost bets.  If your base wager is 10 satoshi and a 1900% increase it goes 10->200->40000->8000000->16BTC->320 BTC.  Most people will not have that much in their bankroll anyway.  The odds of that happening are only 1:64 million.  Its not hard to make millions of bets. 

If the RNG is flawed as I suspect it is there will be ups and downs in the numbers generated.  This results in the odds skewing so at some points its less likely at others its more likely. 

This does not mean that no cheating is going on, it just means that its highly probable that over time users will lose. 

I read the statement that a 2 year old could connect the dots which you outlined as you lost, verified then won,  then got banned as a string of events indicating that your losses were related to not checking.  That it was implied they were cheating.  Especially coupled with the title of this thread.


I understand why you can arrive at the conclusion  there is something to hide, however I can see the sheer stupidity in that and am more likely to believe in stupid people than cheating absent proof that the person is exceptionally clever.  I do not believe this person to be exceptionally clever.  Occam's Razor and all that.  It could be that he/they are really clever though, it could be that they are cheating, I do leave the possibility open but the implication is that they are cheating that circumstantial evidence is enough to support that claim. 

I think the extortion claims came from your demand that your losses be returned.  I think that upset the admin a bit (thus the banning).  It speaks to a more immature mind and a thin skin (its the internet people need to have a thicker one if they want to have a publicly accessible site).  This also speaks more to not being clever than being clever, or at least bad interpersonal skills.

because they want to win every time and they forget that the casinos were not invented for that. you cannot win against them.

coz its a gambling site....why will people belive u thar a gambling site is steeling coins

Never said it was proof, and connecting dots isnt.

Let me say this for the 94th time in this thread and in others: I HAVE NO PROOF. I NEVER HAD PROOF. ITS WHY I THINK HIS SCHEME IS BRILLIANT, BECAUSE IT IS IMPOSSIBLE TO PROVE.

My god. Stop saying "its not proof". No shit. I NEVER said it was.

I'm not looking to convict him in a court of law by a jury of his peers beyond all reaonable doubt. But that is not the only measure of the probability (or plausibility) of something.

My son was sent home sick from school yesterday. Threw up in the bathroom. No one saw it. He APPEARED sick. He didn't want to eat. I didn't do DNA swabs of the toilet. The obvious signs of it were enough to convince me he was actually sick.

I BELIEVE (without proof, again, 95th time) 999dice's insane profits are due to a system designed to cheat the user, where the user can prove nothing, and the site can always "prove" fairness, due to intentional site design, and the be behavior of the site owner (anonymous proxies, answering every question BUT this specific one, banning, threatening news sites).

Yes, a two year old could connect those dots.

Another news site contacted me yesterday and asked me a lot of questions. One of which, I forget what we were dicussing, prompted this response:

"You tell me which of these two scenarios seems most likely for a legitimate company falsely accused:

User claims site is scam. News story is posted. News outlet contacts site. Site owner makes threats, claims news outlet is an accessory to extortion, attacks the user's character.

OR

User claims site is scam. News story posted. News outlet contacts site. Site owner responds, sorry about user's losses, we are provably fair, we'd be happy to show you."

Which of those two are the actions of someone with something to hide?

Listen, if anyone doesn't wasnt to believe me, fine, I'm at the point defending this position that I don't care anymore. Go gamble there. Have fun. Let me know how it works out. Statistics say you should, after an infinite number of bets, lose only .1% of your total bets. Let me know if that number works out for you.