Like I said, I think that's the simplest and best way to go if we're going to switch to a model where I get a percentage of bets rather than of profits.

I don't think it's particularly useful to run a trial.  It's clear enough that in weeks where the site keeps more than 1% of the total wagered, I do better taking 10% of profits, and when the site keeps less than 1% of the total wagered, I di better taking 0.1% of the total wagered...

It strikes me that implementation could be a problem, only because of all the investors whose aren't at peak profit, and so who are due some commission-free profits.

a) I guess the simplest way to fix that is to just refund them any excess they've paid so far.  So the guy who invested 100 BTC, got up to 140 BTC, paid 4 BTC commission, then saw his investment shrink to 130 BTC is 10 BTC below his peak charged profit, has over-paid by 1 BTC, and gets a 1 BTC credit to his balance when we switch to the new system.

Is that fair?  There are lots of investors who pulled out after large losses to nakowa and never came back.  They would get a nice windfall from this change, at a relatively large cost to myself.  What about the guy who invested 100 BTC, got up to 140 BTC then lost to nakowa down to 80 BTC?  He's at a net loss.  Does he get all of his commission refunded?

b) At the other extreme we can just say that overpaid commission is lost, and we're switching to a per-bet charge from now on.  I doubt that's acceptable to anyone (especially to the guy who paid 10 BTC or something for the old mechs account with a view to saving 20 BTC in commission over the coming months or years)

c) There are probably schemes between these two.  I'm thinking of charging 0.1% per bet, then refunding investors who have overpaid commission at the end of each week/day/hour until they have no longer overpaid.

Also worth considering are the pros and cons of switching, assuming we can find a fair way of handling investors who already paid too much commission.

pros:

* If I know I'm getting 0.1% of every bet, I can more easily offer referral schemes.  I can tell people they'll get 0.05% of all volume they refer to me for instance and know that I won't be running at a loss, even if the site has a bad run.

* Under the current system if the site has a few losing weeks in a row, some investors could get worried that since I'm not earning any commission I might be tempted to acquire an income by cheating the investors.  Switching guarantees me a more stable income from the site and removes any such temptation.

cons:

* It could be seen that since my income no longer depends on whether the site wins or loses, I have no incentive to make sure the players aren't somehow cheating.

* When the site has a losing week, the investors not only lose coins to the players, but also to me, making their losses even worse for the week.

Anything else?

Yeah, damn me and my mathematical rigour.


A site is fair or it isn't.  Just-Dice is.

I've never ignored anyone on this forum ever.  You're close to becoming the first.

zipmaster: yet again I find myself without time to give you the response you deserve, but I'm sad to say I'm starting to see it your way, I think.

I think you're wrong that investors are -EV for all bets under 90% chance, as you can clearly see if you imagine every bet was 50%, with roughly half of them winning and half losing.  Investors would make a profit in the long run, of around 1% of the amount wagered, and would pay just 10% of that profit to me.  That still leaves it +EV.

If, on the other hand, you're looking at the short term, then perhaps you have a point.  In the extreme situation where I do a commission run every hour, or even after every bet, then every investor is charged commission at the peak of their random (but +EV) walk, and so it likely at any given point in time to have paid more than 10% of their current profit in commission (since most of the time they won't be at peak profit).

Is that your point?  That whenever an investor is at 10% of their peak (commission processed) profit, they're actually at a loss, because the commission offsets the profit?

If so, then some kind of flat rate (proportional to amount wagered on-site, not amount won) commission structure does seem fairer to investors, even if it does result in them bearing 100% of the variance.

Did I get your point now?

I think the only way it makes sense is to rake 0.1% of every bet from the bankroll, not from the players.

That way instead of getting 10% of profits, I get 10% of expected profits, paid by the investors in real time.

It leaves the edge exactly the same for the players, leaves the investors expecting to pay 10% of their profits, but gives me a much more steady income.  I no longer have good weeks and bad weeks.  All the variance is shifted to the investors rather than me taking 10% of it upon myself.

I don't mind either way.  I don't feel the need to win when the investors are losing, and don't mind that some weeks my commission is very close to zero, because the good weeks make up for that.  But if people think it's fairer for me to take 0.1% of every bet rather than 10% of new profits, I'm happy to do so.

We'd have to find a way to make this work with the fact that some investors have over-paid commission already in that they're still making up coins they lost to nakowa.  Maybe a weekly rebate until they're back where they should be or some such.

That's not true.  I bid on it, but I was outbid by BayAreaCoins.  He pretty quickly sold it on for something like twice what he paid for it, too.

The account is now earning commission-free profits for its new owner.

I'll address this one first, since it's easy...

What if someone becomes 80% of the bankroll and never divests?  He keeps making more and more profit, and I never get to see any of it.  I need to take commission regularly because otherwise I'm incentivising people to avoid paying commission by never divesting.  Maybe one day a big whale comes and wipes out their profits, in which case I get nothing.  Or maybe that never happens, they stay invested forever and I still get nothing.

I didn't take screenshots, but I can query the bet database.

Here are all the bets of 100 BTC or more on account 1337 yesterday:

You again?

The game is probably provably fair as I keep telling you.

Just because I've blocked you from trolling the support system doesn't mean you have to come back and troll the forums.

Edit: sick typo!

Edit2: Also, I'm not encouraging anyone to invest.  I don't care if people invest or not.  The bankroll is arguably too big already, and I wouldn't mind at all being responsible for a smaller number of other people's coins.

It's quite incredible isn't it, but I can assure you that it really does happen.

I've heard that he has a massive mining setup and mines huge numbers of coins.


I hope you stuck around for a few days...


The same guy who won the 602 BTC lost 2190 BTC yesterday, into a 37k bankroll, giving everyone invested an almost instant 6% return on investment.  After losing, he said:

(1337) <LiKaShing> nevermind it's just my playing wallet LOL

ie. this wasn't a serious amount for him - it's just play money.

I hear he's collecting "investments" from people to stage an attempt at a comeback.  Whether to stay in the bankroll for it or not is up to you.

The results are supposed to match up, yes.

Probably the reason they don't is that you're using the wrong roll generation algorithm.

On Just-Dice we used to use a simple algorithm which only used the player's "nonce" value once per roll calculation.  Some investors complained that maybe that wasn't good enough (nakowa had just won a lot and investors were losing confidence) so I gave in to pressure and announced that from bet 145,000,000 we would be using a more complex algorithm to generate rolls.  (We were on bet 144 million or so at the time).

When I made Doge-Dice.com, I copied the Just-Dice code, so that too uses the simple algorithm up to bet 145M, and the other one afterwards.  The two algorithms are both described on the 'Fair?' tab, as is the fact that it switches at bet 145M.

So for doge-dice, which has seen less than 145M bets so far, you need to use the old algorithm.  The verifier was designed with Just-Dice in mind, after the change, but if you read the small grey writing, you'll see that you can put the bet ID into the 2nd box, and if it's less than 145M, it will use the old algorithm.

I just tested it out on my most recent Doge-Dice rolls.  First pull up details on a bet:



Copy the values into the verifier.  Make sure you put a number less that 145M into the betid field.  It doesn't matter what number you use.  I put '1':



Compare with your actual rolls.  The numbers should be identical:

Well, I'm not invested in it much at all any more even though I AM the owner.  So probably not!

Although I was invested with both BS&T (Pirate's ponzi) and Starfish (Patrick Harnett's fund) for a while...

How much is this one worth, do you think?

http://www.reddit.com/r/Bitcoin/comments/1w397k/justdicecom_a_bitcoin_gambling_site_is_very/

Are you thinking that I take 10% of every winning bet?  I don't.  Only new net profits.

Does that change your math?

Suppose every bet was at 50%, paying 1.98x for a win.

After 2 million bets of 1 BTC each, we expect players to have lost 1,000,000 and won 980,000 for a net loss of 20,000 BTC (1% of wagered), and for me to have taken 2k BTC as commission.  That's less than 90%, yet is +EV for investors.

Doesn't that prove your math wrong, since a series of bets less than 90% are still +EV for investors?

If it ain't fixed don't break it.

Assuming I'm already taking the correct amount of commission, and we just want to find a way to make it steady across the weeks instead of being 'feast and famine' style, we can fine-tune a fee, charged to investors per bet rather than on their new profits.  Ignoring the fact that that will annoy the investors who are still running at a loss and so currently not having to pay any commission, the investors end up paying me the same in the long run as they would with the current system.

The only difference is that when the site has a bad week and the investors take a loss, they have to pay out my fee on top of that loss, and so lose even more than they 'should'.

Current system:

1000 BTC wagered.
Bad: Site loses 10 BTC.  Investors lose 10 BTC.
Regular: Site wins 10 BTC.  Investors pay 1 BTC, keep 9 BTC
Good: Site wins 30 BTC.  Investors pay 2 BTC (10 BTC of the profit was recovering bad week's losses), keep 28 BTC.

New system:
1000 BTC wagered.
Bad: Site loses 10 BTC.  Investors pay 1 BTC, net loss 11 BTC.
Regular: Site wins 10 BTC.  Investors pay 1 BTC, keep 9 BTC
Good: Site wins 30 BTC.  Investors pay 1 BTC, keep 29 BTC.

So which is better?  In both cases my total take is 3 BTC.

Current system: -10, 9, 28 (sum = 27)
New system: -11, 9, 29 (sum = 27)

The new system exaggerates the variance of the site by passing it all to the investors, and none to me.  Bigger losses (-11 instead of -10) and bigger gains (29 instead of 28).

Is that a good thing?  It's not clear to me that it is.  It's good for me if I want a steady income from the site.  But I don't care about that.  It's not like I'm dependant on next week's commission for anything.  I can wait for profits, just like the investors can.

It gives me even less reason to care about the profitability of the site than I have now.  Currently if the site loses, I don't get paid.  With the new system I don't care whether the site wins or loses; I get paid anyway.

I was approached by the coinroll guys to hold the coinroll raffle secret before I had plans to start Just-Dice.  I doubt if they would have asked me as the owner of a competing site to help with their raffle, since it just confuses things.  But since I agreed, I followed through and released my secret when requested.  It's nothing to do with Just-Dice though.  We've never had any such raffle or paid for forum signatures.  Other than Klye's, but he's a special case.  Somehow.

He posted his story here:

https://bitcointalk.org/?topic=429331

Posting the link since it's relevant to this thread.

Thanks for writing up your story KoS.  I'll make sure Deb posts it to the Just-Dice blog.  Good luck staying off the gambling sites.  If you want your IP or account blocked on Just-Dice, drop me an email at the address at the end of the FAQ tab.


This shouldn't be a matter of opinion.  With the advent of provably fair casinos you can verify for yourself whether they are cheating you or not.  Of course, they all have a house edge, and so in that sense they aren't "fair".  Most only pay out 2x for a bet that wins 49.5% of the time.  That's a losing proposition long term.


There is.  Email the address at the end of the FAQ asking to have your account or your IP address banned from play, and it will happen.  Many people have already done this.  I never reverse a ban either, no matter how much you beg, or insist that "my friend did it as a joke".  There are links to gambling addiction resources in the FAQ too.

I just made the following chart:



It shows that some 98% of all bets are less than 0.02 BTC, and that the median bet size is around 0.000001xy BTC.

Edit: compare with this one, which plots the percentage of the amount wagered on the y-axis, instead of the percentage of the number of bets:



Half of the volume on the site comes from bets over around 90 BTC.

Edit 2: the same charts but with linear scale x-axes:

Firstly, I don't think it's fair to calculate how much I would have earned if I was charging a fee per bet by looking at the number of bets that were made when it was free to bet.  As you say yourself, charging a fee would cut down on the number of bets, and so would cut down on the amount I earned.

Also, by charging a fee, Just-Dice is no longer on a par with the other reputable dice sites with a 1% house edge - we're now charging 1% edge plus a flat rate.  Why would anyone continue to play at Just-Dice when the other places are 1% house edge with no fee?

By charging a flat fee in addition to the house edge we would be unfairly penalising small players, probably displacing them to the competition, where they may grow up to be big players.

I didn't do any research into it, but I expect it's not unusual for players to start small to try the site out, and increase their bet size as they come to trust the site more.  We don't want to push the small players away to the competition by being less competitive for them.

Finally, switching from a "percentage of investor profit" model of commission to a "flat rate, win or lose" makes for a much more steady income for me at the expense of the investors.  Currently if the site has a losing week, most investors pay absolutely no commission.  If we move to a flat-rate percentage-of-volume model (as some have suggested, but maybe not you) then the investors are paying me each week whether the site wins or not.  That doesn't feel right to me, but as someone pointed out, your stock broker charges a fee whether your trades are profitable for you or not, so maybe I'm wrong on that.

I'm happy with the current system.  I designed it to be as fair to investors as possible, and assumed they were happy with it to.  If there's a compelling reason to change things, then I'm willing to listen to it.  It's true that the investors are "taking all the risk" but I don't see how the change you propose changes that at all.  They still take all the risk, don't they?

It was brought to my attention today that someone has been sending spam advertisements advertising Just-Dice, pretending to be me.

See https://i.imgur.com/5q92ChC.png for an example.

I have nothing to do with this (though the words look like mine - I guess they're copy/pasted from something Deb or I wrote).

While I'm fine with investors spreading the word about Just-Dice, email spam isn't the way to do it, and neither is trying to impersonate Deb or me.