Yeah, I agree it doesn't prove innocence but it does put the numbers back within the realms of normal variance IMO (and hopefully puts to bed any more of the 500M bets at 0.01BTC average talk...).

As for the problem with whales, I've thought about this all night; wondering if there might be some way you could let the individual investors set their own level of risk (via the Kelly fraction, etc), but can't see any way to make it work.

I think the root of the problems you're having is the high Kelly fraction (I see you've reduced it from 100% to 75% so maybe you see this too) and the fact that maximizing the exponential growth of the combined bankroll doesn't necessary go hand in hand with maximizing what's best for the site (in terms of investment, customer numbers, etc).

From an investor perspective: I think they are hoping to see "nice smooth growth" - you only have to take a look at the zulutrade site to see how they all flock to the nice smooth lines of those traders using martingale systems... It's no good if/when the site runs 2 standard deviations below expectation, half your investors think you're cheating them somehow and leave.

You now also have to consider how the player psychology works: yes, there is the opposite problem of what will happen if/when the site runs above expectation, but there are also other subtleties... A drop in the cap / Kelly fraction doesn't necessarily have to mean a drop in profits: these whales may well stake the same amount, but in several smaller bets (thus reducing the variance and the negative effects of it on the investors)... They may even be more likely to stick around longer rather than doing hit-and-runs, thus exposing more of their money to the 1% vig (online poker sites have recently removed high stakes and heads-up tables, I'm guessing for similar reasons).

Perhaps try and get somebody with experience of casino economics to help work out what would be the best changes to make, but I think overall there is a very good case for using a much smaller Kelly fraction or even doing away with it totally and just having a fixed maximum exposure amount.

Juk

Also that's some pretty epic editing you did there.

Yeah, I lost a long post the other day I'd spent ages typing (thanks to one of my cats walking over the kb!) so have fallen into the bad habit of posting asap before reading/correcting it.

Juk

I meant OP - he had a very long post that he edited down to a single sentence or 2.

I think the 1 in 25 number should be put into context because it sounds sort of innocent. I'll make a list that in combination with that number almost prove the site is rigged against investors.

1. With a 1 and 25 chance of happening that means that 25 people could start this website and 96 percent of them would have had a better result than just-dice.com

2. The volume of bets at 532 million should indicate very substantial website traffic. PokerStars, for example, with millions of players has had various billion hand promotions. That's with very aggressive marketing on television. just-dice has had betting volume that would indicate thousands to tens of thousands of players. Where are these players? Does that volume of betting make sense?

3. The site allows bots to be used to bet on the site. How likely is it that the small group of people that have the understanding to automate a bot to make bets would have the bankroll or the willingness to use those bots to place millions of bets? I think it's really unlikely that people would use bots to bet in a -ev game for millions and millions of hands.

4. just-dice.com is kind of a crappy domain name for a site that has millions of dollars sloshing around it.

5. The screen names of the players seem like they are out of an American name book. The owner says he "assigned" those names to people because he thought they were funny because they sound like hillbilly and redneck names. Names like billyjean or marysue etc. His explanation seems reasonable on the surface. It is true that if you do not pick a screen name that it will assign you one from a list. My name on the site defaulted to "whale" because I never picked my screen name manually. But the obvious usefulness of this feature for having bots and fake players run on the site is that they don't need to know English to a pick a screen name. Plus i found it doubtful that so many of the players on the site would go with the randomly assigned name when they could just pick their own name easily.

6. The chat room on the site is very weird and the conversation doesn't make a lot of sense. The conversations seem very scripted and there is a lot of linking happening. I think it's a way to get people to click links that download bitcoin mining programs to peoples computers or for other nefarious reasons.

7. The hot wallet / cold wallet thing. The site supposedly uses one wallet to pay bets and another site to store the majority of the invested money on the site. Multiple times this "hot wallet" was being emptied or changed etc. when a "big bettor" was playing. They have this big story about it that the cold wallet is being kept offline so that the owner of the site doesn't get robbed. This feature too benefits a scam because it can explain away delays in paying out investors and also keeps their money at risk when they are trying to cash it out. They ask everyone to input an emergency adress to send their investments to just in case the site gets shut down. But because this aspect of the cashouts is not automated and has to be done manually by accessing an offline cold wallet it is going to make it easy to steal the money and say it was seized during the "raid" etc. I predict that is exactly what will happen. There would be nothing stopping him from using multiple active wallets and processing all present cash-outs instantly and in real time and have it be just as secure as this supposed hot-wallet/cold-wallet thing.


The bottom line is that I think a lot of thought has been put into this scam. I also believe that a number of people are involved in it. The evidence for that is the post I made on Reditt that initially was getting up-voted by 7-8 people and then to test them I mentioned that I made it in the chat room. Suddenly there was a flock of down-votes and fake "conversations" between people on there.

None of those things constitutes proof and a few of them seriously undermine your position by making it look like you're grasping at straws. Having more points added to your argument is not helpful if those points don't support your thesis.

Really there is nothing here to look at except the actual bets - the other things on your list just do not improve the level of proof.

So, um, having a little trouble keeping your story straight here?

The site's hold on wagers is ~40 million dollars below expectation based on juk's numbers. Your claimed take is 1mm, and it was a hell of a lot less than that pre-explosion. Maybe a bit over 100k. So, uh, yeah. When 40 million goes missing, and my choices are that somebody could have stolen a ton of it unaccountably, or it's all bad luck.. whatever my prior probability of a random trying to steal 8 figures with a good chance of getting away with it, which is pretty high to begin with, is significantly higher now.

How about the fact that the "big winner" on the site just happened to start his own dice site that he used an AMA on Reditt to advertise? The site has many of the same features as the just-dice site. His AMA answers were completely absurd and evasive for the most part.

The crazy part is that I have no reason to post any of this. I didn't invest or play on the site. Also, you don't think website traffic is relevant in this situation? You're talking about 532 million bets. I just think that if those numbers were true that there would be a lot more people talking about the site because they would have been on it.

Another thing to consider is that we are dealing with bitcoin here. It's not just one whale either that has taken the site for a big number.

So in order for the numbers to make sense you need to have whales who happen to be involved in bitcoin who are willing to make bets larger than Phil Ivey makes. Then you need most of these whale to have net winning accounts and have them bet in ways that go against human nature.

I want to add another thing. The owner of the site explicitly admits that the site is not provably fair for investors. That he has the ability to change the results and that you have to trust he isn't. Some of the investors have tried to suggest ways he could make all results provably fair by somehow using a process that is known to create randomness and then having bets on the site interact with that process.

He has given a bunch of "reasons" that the site is just fine the way it is. Also some of the players have been complaining about "lag" on the site. The site intentionally creates different wait times for different bettors.

The cold/hot wallet thing is a feature that can be sold as having a noble purpose for the site but that also happens to be useful in situations where the site was ripping off the investors. When you see a lot of features that work together better in a scam than in a legitimate site it usually turns out to be a scam. Of course he is going to sell the features of the site in legitimate ways. What you're trying to find out after realizing that the site only has a 1 in 25 chance of having the result it has is are there other things about the site that look unrealistic individually or in combination.

If you want to test your theory that it's all a scam then one thing you could try is using this idea on the data file he gave us:

http://en.wikipedia.org/wiki/Benford's_law

Pretty sure there are other tests that you can make if you research "forensic accountancy".

Juk

It bothers me enough that I am going to have someone from the FBI look at it.

Here are the results on the bet sizes for the data file he gave us:



(I excluded the 6000-ish bets of size zero)

Juk

I don't think so. Do you see a contradiction?

Perhaps you're confusing the fact that the site is provably fair for the players (there's no way I can chose what rolls to give them to make their big bets lose more often, for example) with the fact that there's no way I can prove I'm not cheating the investors. I could be one of the big players.

Yes, but almost all of that shortfall is due to the whale not losing his expected 1%. The whale didn't win 40k - he did 3000k of turnover and won 10k instead of losing 30k as expected. So there's not 40k "missing" when 30k of that is just coins that should have been lost but weren't. Either way the amount the whale won is a lot more than I've taken in commission, so your point stands - I have a lot to gain by being dishonest.

Oh my God - it's almost exactly how it should be.

That's suspicious. I must already know about Benford's law and have manipulated the bet sizes so as not to look suspicious.

OP should add this to his list of 'evidence'.

I'm not sure this is worth answering since it's almost entirely off-topic. The only point you made relative to the statistical analysis you requested is that 1-in-25 chances don't happen 96% of the time...

This seems to be basic mathematics.

It's possible for a single player to make 20 or so bets per second on a single account. The site is fast, and allows betting using keyboard shortcuts. Many players seem to use martingale bots, presumably thinking that if they start their betting progression small enough they're never hit their limit. Most of them are right and make a trickle of income from it. Occasionally one hits 20 losses in a row and cries about how we cheated him. 500 million divided by the number of seconds in 7 months is about 27. That's an average of 27 bets per second.

I've barely marketed the site at all. If has a lot of players because it offers the best odds and the highest limits, and because the game is provably fair. We always pay out winners quickly, there's never been a dispute over a cashout. And so on.

The site allows bots in the same way as any site allows bots. Bots look a lot like a regular player, interacting with the site through the same interface. I don't go out of my way to block them because if people want to automate losing their coins that's OK with me. People discover martingale betting every day and think it works. It's boring to do manually, so they bot it. It's dumb, but it happens.

I don't like 'northeastbeast' as a name either, but there's no need to be rude.

I don't see how this matters. There's a list of default names so people don't have to pick their own name, but still appear somewhat distinct in the chat. On the day of launch the default name was 'satoshi', and the chat was full of different satoshis talking to each other. So I added more default names.

This has to be your most crazy point. I'm careful to only 'linkify' things in the chat if they point somewhere safe. I link to blockchain.info if people write bitcoin addresses or transaction IDs. I link to imgur if they post imgur image links, and so on. I don't like to random mining programs! There are moderators in the chat who will mute anyone who tries to convince people to download malware of any kind. Also, you think I've stolen 40 million dollars. Do you think I'm going to bother tricking people into installing mining programs?

If you have spent any time around Bitcoin, you'll have heard of at least one site getting hacked and having all their Bitcoin stolen. It happens a lot more than you might expect. And in almost every case the site had all their Bitcoin stored in a wallet on the hacked server. This is just stupid. Why keep more than you need for regular day to day operation on a web server? I don't - I keep the vast majority of it offline where it can't be stolen.

When a big player comes and deposits 500 BTC, those coins are very quickly moved offsite, so they aren't vulnerable to theft. If the player wins 100 BTC and then tries to withdraw his 600 BTC balance, it's not available. It shouldn't be available. I don't want to be keeping that many coins on a web wallet. Almost all players understand this and appreciate that I'm keeping their coins safe. Having to wait a few hours to cash out half a million dollars isn't such a hardship. Try cashing out from PokerStars sometime and see how long it takes!

The idea that I can achieve a similar level of security by keeping all the coins online but split between multiple wallets, all of which can be accessed from the hacked server, is probably not worth addressing, for obvious reasons.

OR... a lot of thought has been put into this successful website, which you have absolutely no evidence against other than that it lost a bunch to a lucky whale, and that the players in the chat are weird.

So you mentioned your very anti-Just-Dice reddit post in a chat-room full of Just-Dice users and were surprised when a few of them downvoted your post that they disagreed with?

Here's the reddit thread in case you feel like replying to me there.

The site isn't provably fair for the players. That's nonsense. How can you possibly prove that the whale was playing a fair game and wasn't playing with "an edge"? The whole reason you're here is because you couldn't.

JD is provably fair for players, NOT investors. Zero-trust systems work as advertised provided that the house doesn't collude with players to cheat investors. That said, dooglus (the site operator) has a very high trust ranking in the bitcoin community and I personally believe him to be an honorable person.

On the other hand, 38,000 is a LOT of coin! About 32 million at todays prices.

I think this is a question of definitions.

When I say "provably fair for players" I mean that the players can know that the site isn't cheating them.

When I say "not provably fair for investors" I mean that the investors can't know that the players aren't cheating them.

You could argue that if a player is able to cheat the investors then that isn't fair for the player either, but it's unfair in a "good" way (from the cheating player's point of view). That's not the kind of fairness I'm referring to. Maybe I'm misusing the word "fair" but I mean "not being cheated" rather than "not cheating".

I missed this post of yours. I'll respond to it now.

The "big winner" you're referring to called himself "nakowa". He first approached me shortly after the site launched, saying he had a lot of bitcoins to invest, but wanted to buy a share of the site rather than just bankrolling it. That obviously would have been better for him, since the current "investors" aren't really investors at all - they're just providing the bankroll but own none of the site. I told him I wasn't interested, that I didn't need to sell any part of the site, so he ended up bankrolling the site like any other "investor".

I don't know what it was that made him start playing on the site, but he did. He always claimed "I'm not a gambler" yet made huge bets and did very well. I'm sure he thought he had a system that could win - he would bet tiny amounts until he lost a few in a row then bet bigger, falling victim to gamblers' fallacy. Whatever.

That he ended up launching / investing in his own competing site was no surprise to me. It was consistent with his desire to own a dice site rather than just be bankrolling my site. He saw what I created, liked it, wanted part of it, and when I wouldn't sell he made his own version, taking my idea and trying to improve upon it.

As for his AMA, I didn't read most of it. As I understand it he's in China where gambling is illegal, so that could explain his evasiveness. His English language skills aren't great either, which may explain part of the problem too.

If I was draining the site by playing against the house, I don't think it would be very smart to do it as a single huge player making unrealistically large bets and bragging about it on a reddit AMA! This seems like an argument against it being me. Wouldn't it make more sense for me to regularly make new accounts, and to win small amounts on them effectively being "under the radar"? It would do me no good to draw such massive attention to myself.

I don't know much about that. http://www.alexa.com/siteinfo/just-dice.com tells me that Just-Dice is ranked 183,138 in the world. Is that good? Note that the vast majority of the 500 million bets are tiny. Less than a cent's worth. I've no way of knowing, but I would guess most of them were placed by bots too.

nakowa took the site for something around 10k BTC. LiKaShing is still playing regularly, and seems to be up and down large amounts each day. It's hard to know how much he's up or down because he constantly makes new accounts, constantly switches IP addresses, and is less than forthcoming about which accounts are him.

I don't know how familiar you are with Bitcoin. A few years ago you could buy them for less than a dollar each. Today they're worth over $1000 each. A lot of people recognised the potential of Bitcoin early and bought in. It should be no surprise that today there are regular people with hundreds if not thousands of Bitcoin. You make it sound as if it's unlikely that a whale would "just happen to be" involved in Bitcoin. What you miss is that having been involved in Bitcoin for a while is pretty likely to have turned you into a whale, just by the nature of its explosive rise in value.

I don't see why you need most big players to have net winning accounts. It's not unusual to see someone deposit 100 BTC, quickly lose it, and not come back again. Nobody remembers it, because it's not "amazing" like the nakowa and LiKaShing play is amazing.

No, you misunderstand. I can't prove that I'm not playing against the investors. In that respect it isn't provably fair for the investors. But it is provably fair for the players. I *can't* change the results, and that is the point. Here's how it works:

1) the site makes up a secret seed S
2) the site shows you a cryptographic hash of its secret, hash(S)
3) the player picks their own seed C
4) all rolls the player makes are completely determined by S,C, and N, where N is the number of bets the player has made since steps 1 through 3 happened
5) the player hits 'randomize' which reveals S, and goes back to step 1 where a new pair of seeds is picked.

So once step 5 is complete the player known S, can verify that hash(S) matches what he was shown in step 2, and can use S and C to generate the list of numbers he rolled for himself, verifying that the play was fair. Since the player picks C *after* the site picks S there's no way the site can influence the rolls the player gets.

Yes, but they don't understand either. I could do away with the above provably fair system and just get every roll from a trusted source of randomness. random.org for instance. But then the site isn't provably fair at all. You have no way of knowing that I really used random.org and didn't just pick a number that makes you lose. The point is that the roll generation has to be repeatable so that the player can verify it was correct after they played.

It's not that the site is "fine the way it is". I would rather have the site be provably fair for investors as well as players. That wouldn't stop you being suspicious of me, because you wouldn't understand the system, but it would at least prove that I wasn't cheating for those who can follow a proof. The problem is that I don't see any way that the server can generate a sequence of rolls for a player without me also being able to do the same. The server is only running the code that I tell it to run after all. How could it 'know' things that I can't know?

The site is busy, and under constant DDoS attack. It lags sometimes. And suspicious gamblers ofter think that the lag is the site cheating them somehow. But once you understand how the rolls are generated, using a simple HMAC hashing algorithm, you will realise that it's just not possible. Unless the site is able to find collisions in sha256 during the lag (a feat that currently takes billions of years) there's just no way.

I don't see how keeping the majority of the funds safely offline helps anyone scam anyone. If the coins were all online it would be easy for me to say "oops. server got hacked. coins are all gone, sorry". Keeping the coins offline makes them safer, not more at risk. Online coins can be stolen by anyone who can gain access to the wallet. Offline coins can only be stolen by people with physical access to the machine they're on. Both sets include me, of course, but the 2nd set is obviously smaller than the 1st.

Based on Alexa.com findings, just-dice.com has a global traffic rank of 182,257. just-dice.com also reaches roughly 21,506 unique users each month and has 107 backlinks according to Alexa. The estimated monthly ad revenue is $ 66.92.

We estimate the website value of just-dice.com is currently at $ 1,606.00. Traffic and Revenue of just-dice.com is estimated to be increasing within the last 3 months. The domain just-dice.com is a .com domain with the name just-dice preceding it.

The Servers of just-dice.com are powered by the cloudflare-nginx webserver software and the servers are physically located in Europe and United States (California) and use the IP addresses 141.101.124.227 and 108.162.207.227. The average loading time of just-dice.com is 1165 milliseconds - this is faster than 69% of all other websites. At the DMOZ open directory project we found no listing for just-dice.com at this time.

Updated: 01/27/2014 08:27, 0 minutes ago

Worldwide Rank Monthly Users* Monthly Pageviews*
182,257⇑ 21,506⇑
708 per day*
27,528⇑
905 per day*
Website Value* Monthly Earnings* Daily Earnings*
$ 1,606.00⇑ $ 66.92⇑ $ 2.20⇑

How engaged are visitors to just-dice.com?

Bounce Rate
75.90% 10.00%
Daily Pageviews per Visitor
1.28 14.09%
Daily Time on Site
2:38 16.00%

What is a Bounce Rate?

According to Wikipedia:

Bounce rate (sometimes confused with exit rate)[1] is an Internet marketing term used in web traffic analysis. It represents the percentage of visitors who enter the site and "bounce" (leave the site) rather than continue viewing other pages within the same site.

Bounce rate is a measure of the effectiveness of a website in encouraging visitors to continue with their visit. It is expressed as a percentage and represents the proportion of visits that end on the first page of the website that the visitor sees

So 28000 total users a month x 7 months is 196,000 total for both unique and repeat visits combined. A 75 percent bounce rate means 50,000 total people spent an average of 2 hours and 48 minutes on the site including all employees and you. So that is 168 minutes per user multiplied by 50,000 equals 840,000 minutes spent on the site. 532 million divided by 840,000 equals 633 bets per minute per visitor.

Are you ready to tell the truth yet?

Whether you're right or not, you certainly do sound like an idiot.

No idea if the game is fair or whatever, but I tend to side with someone who is actually willing to post all of his/her data.

Who are you saying this to?

You.

In specific, I don't think you understand how alexa ratings work. It's not as if they can really track all the users on a site. Instead, it's sort of like Nielsen ratings. If you are running the Alexa toolbar, then it tracks you (sort of half-assedly) and extrapolates to the web population as a whole. Alexa ratings for all but the most highly trafficed mainstream sites are garbage.

Treating it as an AHA GOTCHA makes you look pretty clueless.

I'd rank it right up there with your argument that it must be rigged because just-dice.com is a crappy domain name.