Right, and all I'm saying is that if you or I "turn evil", we can still run off with player funds.  Being player-vs-player doesn't change that.

I thought of all the above already before you wrote it, which makes me regret offering the bet even more.  Even if I win, I lose credibility, since someone's bound to think I cheated, and there's no way I can prove I didn't.

I started work on processing the giant bet list to find long streaks.  There's a bit of a problem, in that I don't record whether the bets win or lose, only the profit or loss.  For zero stake bets, I just record that the profit was zero, and can't tell easily whether it was a win or a loss.  To work that out I have to recalculate the 'lucky number' and see if they guessed hi/lo correctly.

But the plan is to find a 28 streak of non-zero bets.  If I can do that, then the zero bets are moot.  Mostly people don't bet zero anyway.

I've extracted a list of all the 49.5% bets, and am currently splitting them into one file per userid.  To check the script was working, I did a search for the "longest streak so far" after processing the first few million bets.  It found a 24:


This guy actually martingaled from a single satoshi, lost 24 in a row, and still ended up winning!

So I'll leave the script to finish splitting the bets into multiple files, and not search for streaks until the 7 days are up.

I'm really enjoying this bet.  I especially liked the bit where I said:


ffs.

He's asking about the total number, not the 27 streak.

But I don't know the answer. I didn't do any analysis on the bets yet.

So here's the bet for anyone who wants to take me up on it: Edit4: betting is over - the 50 BTC limit was reached

I bet there has been a streak of 28 or more losses in a row at 49.5% by a single account on JD in the first 1,216,045,813 bets (that's how many there have been up to now).  I'll ignore bets at different chances - so if they lost 15 times at 49.5%, then win one at 90%, then lose another 15 at 49.5% that counts as a losing streak of 30 at 49.5% - the 90% bet is ignored.

I don't know how many of the 1216M bets at JD were at 49.5%, but I guess it's around half of them and so I think I have a decent chance of winning this.  The offer is open to anyone who wants to take me up.  Send your bets to 1CNPpAxC5Si15qcVBQn7jEoXmBKTn7fSqg and if I lose I'll return twice as much to the first sending address of each transaction, satoshidice style.  Offer open until exactly 7 days from the time of this post.

Edit: I guess I should set a maximum so I don't bankrupt myself.  Once the total reaches 50 BTC I'll return any excess to the sending address.  First come, first served, in blockchain order.

Edit2: just to be clear, the player doesn't need to be martingale betting, and the stake doesn't matter at all; losing bets of 0 BTC are still losing bets.

Edit3: I'm talking about sequences of 28 or more bets at 49.5% (after filtering out non-49.5% bets) where the player lost.  Bets where the house lost aren't counted as "losing bets" for the purposes of this bet.

Still not right, sorry.

23 losses = 0.08388607 BTC - not even a tenth of 1 BTC.

Maybe you're right, but I doubt it.  I've never seen more than 27 in a row, but with over a billion bets at Just-Dice I would be surprised if there wasn't at least a 28 streak at one point.  2^20 is about a million, 2^8 is 128, so it would take about 128 million bets on average to see a 28 loss streak at 50%.  And quite a few less at 49.5%.

People keep asking me "what's the longest streak on JD?".  I didn't look.  It will be a pain in the ass to go through the data and work it out.  But if you want to make a little wager on the existence of a 28 loss streak at 49.5%, that would make it worth my time.

So how about it?  Promise I won't cheat and look it up ahead of time.

Here's the 27 streak:

https://i.imgur.com/S5zjWv8.png - photo of most of it
https://i.imgur.com/AeIqshZ.png - screenshot of all of it

It is unlikely, but not super unlikely for a lucky player to destroy a site's bankroll.  I tried warning the ED admin that his maximum bet was too high to be safe.  I think they had a house edge of 0.99% and a max bet of 1% of their bankroll.  That's asking for trouble.  Bear in mind that nakowa's biggest play was after we had reduced the max profit from 1% of bankroll to 0.5%.  It didn't stop him from winning.

Even now at Just-Dice.com, with the max profit being just 0.5% of the bankroll we still often see players 100x their initial deposit by being lucky.  Most of them give it back eventually, but occasionally we'll get a player who knows when to stop withdrawing a massive profit.

Oh, that would be a good conspiracy theory!

dooglus set up ED and made it fail to show people that you should only trust established sites like JD!

Only problem is that ED was too well designed to be a dooglus site. 

From what I saw:

Ruru offered to buy his balance at a huge discount, planning to sell it back to ED at a profit.  1dice1 didn't accept the offer.

Same at ED.  You weren't playing against the site, you were playing against the investors.  The site made a profit from every bet, just like you do on your site.

It's much the same.  You could decide to run off with player funds at any time, since you hold them which the bets are pending.

People sometimes make the mistake of thinking that the on-site chat is for support.  It isn't.  From the FAQ tab on the site:

Virtex has been running longer and has more trust than either of those two.

Also, bitfinex.com and bitstamp.com were already registered by the companies trading under the corresponding names.

Virtex were trading using the cavirtex.com domain only, so virtex.com was available to register by someone looking to cash in on their good name.

As one of the oldest exchanges pretty much everyone involved in Bitcoin has heard of Virtex.  To claim it is only known to Canadians is just wrong.

So in order to get more check marks we need to make the site work less well so you have a reason to contact support?

That seems a little messed up. 

It's the other way around.

I left him negative trust because he's running this scam.

I don't see any ethical problem with Ruru buying 1dice1's ED balance if both parties are happy with the deal.

The offer was maybe a little low, but it's starting to look like it may have been a little high, too.  For those who think otherwise, I'm owed 1 BTC by pirateat40, and am willing to sell the debt for 1% of its face value.

I don't plot any of that kind of data except that the slope of the green line tells you the betting volume since expected profit is exactly 1% of total volume.

https://bitcoinproject.net/justdice.php has live charts of the kinds of things you're asking about.

I'm just trying to get the lines to start close to each other at the beginning of the chart.

How about I post them here rather than in the blog?

Here's a chart showing the profit and expected profit since the big nakowa/mechs mess at the end of September 2013.  You can see the profit has followed the expected profit pretty closely, and in recent weeks has been rising a little faster than expected (click the charts for bigger versions):



And here are the last 3 weeks:

The QR code is nothing more than an encoding of the 16 character key, and a descriptive text "just-dice 123456" where 123456 is your account number, and which google-auth shows next to the 6 digit code so you know which site each code in your app is for.

You can set up a new device with 2FA either by scanning the QR code or by manually entering the 16 character key, so you only need to keep a copy of one of them.

Depending on your paranoia levels, you might want to consider whether your printer is keeping a copy of the QR code when you print it.  Some have long-term memories apparently.  Noting the code on paper and storing it somewhere safe is probably the best option.