Just-Dice.com : Invest in 1% House Edge Dice Game

[Mitchell]

Damn. That guy knows how to bet.

[Lohoris]

Of course it is simple.  When Dooglus provides me an updated total betting history of nakowa, we can do a shuffle analysis and see for real.  Last time I did it when he was up 4000 BTC, it was around 7% chance.

5% chance is 1/20.
Many people here will be familiar with dice.
Roll a 20 sided dice, get a 20 --> 5% chance.
You know, it happens.
He critted just-dice.

[Lohoris]

Damn. That guy knows how to bet.

Nope.
He's just lucky.
Or he's cheating.
No other options on the table.

[Peter R]

Some thoughts on the whale's win today (conclusion: he had a 67% chance of pulling it off).

The whale can STOP betting when he chooses, so the probability of what happened today is not as low as it seems.  I haven't dug up the exact numbers, but let's assume:

1. He wanted to win 2500 BTC (after which point he would stop gambling)
2. He was willing to go down as much as 10,000 BTC before giving up
3. He was betting 200 BTC each time and always attempting to 2X his bet (this just simplifies the simulation). 

I ran the following Mathematica code:

In[6]=
bankroll := 10000;
target := 2500;
n := 10000;

bet := If[Random[] > .505, 1, -1]

outcomes = Table[profit = 0;
   While[profit < target && bankroll > -profit, profit += 200 * bet ];
   If[profit > 0, 1., 0], {n}];

100 * (Plus @@ outcomes)/ n

Out[6]= 67.23

So, 67% of the times that he applies this strategy he would earn his 2500BTC target before going bust. 

[Lohoris]

Some thoughts on the whale's win today (conclusion: he had a 67% chance of pulling it off).

The whale can STOP betting when he chooses, so the probability of what happened today is not as low as it seems.  I haven't dug up the exact numbers, but let's assume:

1. He wanted to win 2500 BTC (after which point he would stop gambling)
2. He was willing to go down as much as 10,000 BTC before giving up
3. He was betting 200 BTC each time and always attempting to 2X his bet (this just simplifies the simulation). 

I ran the following Mathematica code:

In[6]=
bankroll := 10000;
target := 2500;
n := 10000;

bet := If[Random[] > .505, 1, -1]

outcomes = Table[profit = 0;
   While[profit < target && bankroll > -profit, profit += 200 * bet ];
   If[profit > 0, 1., 0], {n}];

100 * (Plus @@ outcomes)/ n

Out[6]= 67.23

So, 67% of the times that he applies this strategy he would earn his 2500BTC target before going bust. 

This is a very interesting take.

Unrelated fun fact: while I don't think he's cheating, it is worth noticing that if he was cheating, then pretending to be victim of a gambler's fallacy (like he is doing) is actually a great tactic to go unnoticed.

[chriswen]

So, 67% of the time he wins 2500 btc.  33% of the time he loses 10,000 btc.

[Peter R]

So, 67% of the time he wins 2500 btc.  33% of the time he loses 10,000 btc.

Yes, I believe so.  His expected return is still negative (as required if he is not cheating): 0.67 * 2500 BTC - 0.33 * 10,000 BTC =  -1625 BTC. 

Of course the exact probabilities depend on how much he is willing to lose, and we don't know that for sure yet!

[Peter R]

Some thoughts on the whale's win today (conclusion: he had a 67% chance of pulling it off).

The whale can STOP betting when he chooses, so the probability of what happened today is not as low as it seems.  I haven't dug up the exact numbers, but let's assume:

1. He wanted to win 2500 BTC (after which point he would stop gambling)
2. He was willing to go down as much as 10,000 BTC before giving up
3. He was betting 200 BTC each time and always attempting to 2X his bet (this just simplifies the simulation). 

I ran the following Mathematica code:

In[6]=
bankroll := 10000;
target := 2500;
n := 10000;

bet := If[Random[] > .505, 1, -1]

outcomes = Table[profit = 0;
   While[profit < target && bankroll > -profit, profit += 200 * bet ];
   If[profit > 0, 1., 0], {n}];

100 * (Plus @@ outcomes)/ n

Out[6]= 67.23

So, 67% of the times that he applies this strategy he would earn his 2500BTC target before going bust. 

This is a very interesting take.

Unrelated fun fact: while I don't think he's cheating, it is worth noticing that if he was cheating, then pretending to be victim of a gambler's fallacy (like he is doing) is actually a great tactic to go unnoticed.

I agree.  As long as we can't statistically prove that he is cheating, to within let's say p < 0.01, then he could continue cheating 

[Stunna]

I think this is irresponsible at this point, max bet should be lowered to protect investors in my opinion. Even though j-d is my competition it is distressing to see their profit fall lower and lower every day. My advice is to lower max bet or increase edge or both. Also, doog should explore the possibility that nakowa has found some sort of exploit.

When situations like this occur it is important to go back to the drawing board and figure out how to prevent further losses like this. Changes should have been made a while ago in my opinion.

[addi]

I think this is irresponsible at this point, max bet should be lowered to protect investors in my opinion. Even though j-d is my competition it is distressing to see their profit fall lower and lower every day. My advice is to lower max bet or increase edge or both. Also, doog should explore the possibility that nakowa has found some sort of exploit.

When situations like this occur it is important to go back to the drawing board and figure out how to prevent further losses like this. Changes should have been made a while ago in my opinion.

couldn't agree more

[galbros]

+1 Peter R

This is exactly the type of analysis that explains what is going on.  IIRC he was down 4K coins at least three times today but kept going until he made his target number. 

For individual investors I guess the trick would be to divest when he is down a lot.  Of course, you have to be on when he's playing to do that.

As for the site as a whole I guess it is wait for the big loss, but at this point, he may be willing to lose even more than 10K coins.

[GreenBits]

Of course it is simple.  When Dooglus provides me an updated total betting history of nakowa, we can do a shuffle analysis and see for real.  Last time I did it when he was up 4000 BTC, it was around 7% chance.

5% chance is 1/20.
Many people here will be familiar with dice.
Roll a 20 sided dice, get a 20 --> 5% chance.
You know, it happens.
He critted just-dice.

nah, Just-Dice rolled a natural 1..

[rigel]

Where are you Dooglus?

Any comments on what's going on?

Do you think he found an exploit or is just very lucky?

[dooglus]

For what it's worth, here are the profit charts.  Notice that for almost all of today's session he was down.  Then he quit while ahead, as usual:





As for the calls to "change something", if he's cheating then limiting the maximum bet won't stop him winning, and if he isn't cheating and is just lucky then limiting the maximum bet will only slow his inevitable losses.  I don't see how it helps in either case.

One thing that I can change that will protect investors is whether the site is up or not.  If I take it down they won't lose any more.  But I suspect that most investors wouldn't like that decision.  They want the site to stay up so they can win their coins back again.  The ones who want this to end have the option already of divesting.

I'm quite tempted to just give up on this whole thing.  It looks like the site is simply at nakowa's mercy.  He's promised me in the past that he won't play any more, but never sticks to his word.  So what do I do?  Continue to let the investors take these horrific losses?

[jabetizo]

So what do I do?  Continue to let the investors take these horrific losses?

yes, nobody is forced to invest or stay invested.

[maqifrnswa]

So what do I do?  Continue to let the investors take these horrific losses?

Barring cheating, it is what it is. JD is a high-risk gambling site. Variance increases when someone comes by with a bigger bankroll than JD. I'm down 20%, but I know what I'm getting myself into and that this could happen. Keep on running the business, improve features - try to attract more whales. I'll take my shots.

[TheFuneral]

....
I'm quite tempted to just give up on this whole thing.  It looks like the site is simply at nakowa's mercy.  He's promised me in the past that he won't play any more, but never sticks to his word.  So what do I do?  Continue to let the investors take these horrific losses?

I don't think it's fair to take down the site, but I also don't think it's fair for people to be mad at you when they lose. It's not your fault. Sadly I assume the latter will keep happening.

[Peter R]

I'm quite tempted to just give up on this whole thing.  It looks like the site is simply at nakowa's mercy.  He's promised me in the past that he won't play any more, but never sticks to his word.  So what do I do?  Continue to let the investors take these horrific losses?

Firstly, I want to compliment you on your brilliant ideas for just-dice.com.  I discovered it only a few days ago, and have been thinking about it ever since.   I really appreciated the analysis you did for SatoshiDice and I can tell this has given you a great deal of respect in the quickly-growing bitcoin community.   

In regards to taking the site down, well, you are a statistics expert:  Let’s make the null hypothesis “Nakowa is not cheating” and see what the p-value is when we test it.  Obviously, Nakowa is lucky, but if the p-value is, say, no smaller than p = 0.05, then I don’t think we can conclude that he is cheating.   The investors signed up knowing that there were risks involved, and shutting down the site would not be fair. 

Lastly, while I have no doubt that you and your team are completely honest and trustworthy, I suspect that if you take the site down without strong evidence that there is a real problem, it would sort of make you look guilty as per Lohoris’s insight “while I don't think [Nakowa is] cheating, it is worth noticing that if he was cheating, then pretending to be victim of a gambler's fallacy (like he is doing) is actually a great tactic to go unnoticed.”

Personally, I hope Nakowa comes back again soon...

[rigel]

Don't take me wrong dooglus

I's not your fault. Whatever happens I'm with you.

He is still ddosing the site? This is unfair
Why don't you just take down the site during ddos ?

[stslimited]

Some thoughts on the whale's win today (conclusion: he had a 67% chance of pulling it off).

The whale can STOP betting when he chooses, so the probability of what happened today is not as low as it seems.  I haven't dug up the exact numbers, but let's assume:

1. He wanted to win 2500 BTC (after which point he would stop gambling)
2. He was willing to go down as much as 10,000 BTC before giving up
3. He was betting 200 BTC each time and always attempting to 2X his bet (this just simplifies the simulation). 

I ran the following Mathematica code:

In[6]=
bankroll := 10000;
target := 2500;
n := 10000;

bet := If[Random[] > .505, 1, -1]

outcomes = Table[profit = 0;
   While[profit < target && bankroll > -profit, profit += 200 * bet ];
   If[profit > 0, 1., 0], {n}];

100 * (Plus @@ outcomes)/ n

Out[6]= 67.23

So, 67% of the times that he applies this strategy he would earn his 2500BTC target before going bust. 

This is exactly what I've done using martingales at bitzino's betting limits

Their table limit was 100 btc at the time, denominated in 10000 mbtc. My 67% was based on the fact that I could get martingale'd 6-7 times in a row within 150 turns, at any point of course. This was the chance I was willing to take, gambler's fallacy completely in mind, and in that many turns I could expect to make 30-40% gain.

I made 35 btc and was done.