Topic: Just-Dice.com : Invest in 1% House Edge Dice Game - page 50. (Read 435357 times)

I would argue that the manager and staff are likely paid the same each week rather than having their wages depend on the ups and downs of the house profit for the week, so your example kind of supports the new system more than the old one.


I don't think it's reasonable to think that I need to expect to be able to legitimately earn as much as I could potentially steal to stop me from stealing it.  Does a security guard expect to be paid the value of the thing he's guarding?  There are other reasons preventing me from walking off with the bankroll, such as it being wrong.  I want to be able to feel good about myself, and stealing from a bunch of people who trusted me wouldn't let me do that.


I was just pointing out that some investors still aren't paying commission, and so wouldn't be happy if we just forgot that they have overpaid commission in the past and started charging them per bet now.


That chart starts part-way through the site's history.  He started collecting data at some point in time, and that's where the charts start.


No, you're wrong here.  The 1% and 0.25% that people are talking about is site profit expressed as a percentage of the total amount wagered, and we can totally do expect it to be 1%.  That is the expected profit.  The size of the bankroll is irrelevant to that calculation.  That it isn't 1% is entirely due to nakowa getting lucky with huge amounts of very big bets.


I'm not melancholic.  Well, no more than usual.  I'm just looking at a possible way of making things more fair to investors going forward.


If I had been taking 0.1% of every bet as commission from the start, instead of 10% of profit, I would have taken 4931 BTC in commission, since there have been 4931k BTC wagered.  So I would have made 4 times more commission that way than I have made by taking 10% of profits.  That's exactly because we've only made 0.25% profit on turnover instead of the 1% that we "should" have.  You weren't around when nakowa was playing, but he really hit the site hard.  See mid July and the end of September in this profit chart: http://i.imgur.com/0Ikj2PC.png - those are the two times when the red 'expected' line pulled away from the black 'actual' line.  The difference between those two lines represents the difference between how much I would have earned using the two different commission systems.

Of course, before launch there was no way of knowing which line would be higher.  It's quite possible that actual profits are way above the expected profits, in which case taking 10% of profits pays better than taking 0.1% of turnover.  That could be the case going forward - we can't know how the site will perform in the coming months, so we can't say which system would pay me the most.


Yes, exactly that.  I already credit/debit each investor after each bet according to how much they won or lost, so it's no extra work to take off the extra 0.1% at that point too.  What actually happens is I know what percentage of the bankroll each investor has, and just update the bankroll after each bet.  So all I'd have to do is update the bankroll to take the extra 0.1% commission charge into account as well as accounting for the new profit/loss.  In this way each investor could see exactly how much they have invested at any point in time.  With the current system they can't be sure how much they owe me in commission without doing some math.


There are some investors with negative investment profit who are no longer invested.  If they ever come back, the calculations need to still be running for them.

I'm home now, and the hot wallet has been refilled.

Thanks for those who helped keep it from going empty.  Please feel free to withdraw again now if you like.

Oo What? If nakowas happening, which are big betters, making fluktuating the profit in a mathematically optimised border would make someone think dooglus isnt credible then for sure the investor isnt right in his head. Otherwise your fortune telling threat would mean dooglus would have to lower max profit even more which in fact means a lower kelly criterion.

And i think the argument that the new system has to come because dooglus could steal the house isnt valid since he already became a USD-Millionaire with JD. Seeing him in the forum, seeing what he writes about and thinks let me think that he didnt change livestyle the way that 1 Million USD would be too worse than 4 Million USD or maybe even 36Million when stealing the house. If there would be a hint that dooglus's character is that way then investing is the highest risk. Fees wouldnt safe it.

Dooglus:
I think it shows a nice character that you care for the investors that lost but investment is risk and you dont have to make sure everyone wins. Otherwise you would have a unsolveable task doing that for gamblers too.
Investors know the risk, investors should know the mathematics behind and because of that they should know that unusual high wins of players have to go back sooner or later. Staying invested is the way to solve it.
You made the decision to lower the risk for all investors though that means on the other hand that investors that believe in the math behind and want to take higher risks are punished now. Again, i dont think that you have to ensure that every investor makes profit.

Dooglus
I believe he meant the chart at https://bitcoinproject.net/justdice.php that calculates the theoretical profit coming from the house edge. Thats why i dont see why you think you would have earned more fees with the new system. Is it because of the 0.2% profit? Im not sure but i believe its natural when the house isnt as high as it is today from the start.
The JD-Profit cant possibly be at 1% because the house started with low amounts. Those low amounts made their 1% profit each. But when you add more and more thousands to the house then it looks like the profit isnt matching anymore. The reason is only that the house wasnt from the start at 36000BTC. I think thats the reason. And no i really dont think that it looks bad. It looks GREAT. Im angry at myself that i didnt check out JD some months ago. The monthly profit is great. Way more than the 24% Asicminer shares brought and 24% annual is already great. So i think you have a great website and i dont see the reason for your melancholy.

Because of that i dont see why you should have gotten more fees with the new system. I believe you would be exactly at the point you are now. Otherwise... think about it... 4 times higher fee means the investors would have made only 60% profit anymore while you get 40%. Thats nothing that will work very well i think.

Dooglus
When you charge per bet then when would you charge the investors? Surely not really at each bet. So at the end you would charge each investors investment once a week, daily or so in order to not kill your server. And when they divest. So thats the time when you take investment losses into account. It even would work together with individual risk settings per investor when you divide the investment value into the investment value AND the max risked amount of bitcoins like i proposed as a way for implementing individual kelly values.

The investment losses would die with the new system over time though because everyone pays a practically fixed fee and the theoretical profit is NEVER dropping down. So those calculations can stop at some point because everyone pays always.

Dooglus
Nice... Its not hurting. Remaining investors have a slightly higher risk but higher reward too because the house is smaller. If the max profit at some point really is too low then there are options to defend...

It would appear that "many" of the investors that chimed in have shown a certain regard towards the priority that they give to Doog "being on the same boat as them".

Sort of like saying: "We expect a certain standard of security with regards to the site's functioning." As we have been accustomed to expect due to Doog's diligence. A diligence to which we have deposited our trust in exchange for a 1% edged stake in a bitcoin casino.

The taking of commission is not the only example where our expected 0.01% EV is skewed in the short to medium term.

The stability in the fluctuations of the bankroll is also a matter which should seriously be taken under consideration due to its massive decline over both the Nakowa event and the recent months. Whereas we cannot agreeably presume to limit the amount of coins that get invested into the site, we should establish a set of principles to somewhat control the rate of divestment in the short term. The con of this is that there would certainly be limits to the amounts we can withdraw from our accounts all in one go.

The pro on the other hand is that we could sit back, forget about the functioning of the site and learn to at least trust a rigorous, statistically justified edge in a gambling casino. The trollbox might be often animated with spirited scammers, spammers and trolls. The investor sphere of the site however should be taylored for investors for whom the trust in Doog has value. These must be the long term investors and their interests should underline the principles of site governance.

So what is a long term investor exactly? I, personally would define him as one willing to trust Doog and ready to take the mathematically well defined risk of investing in JD over a LONG amount of time. This presumes that the risks are at least quantifiable and, luckily for us, they are. Someone willing not to chicken out after a loss such as the Nakowa event was, should be proportionally missing out on the same percentage of profit as Doog has.

Lets look at the numbers then. Doog has made 1.2k BTC as opposed to the 4.9k BTC (75% less) he could have made had he been charging a steady 0.1% fee. What did the "long term" investor miss out on? For a rough estimate, lets realize that the bankroll size shrank from a maximum of ~70k to the current ~35k that we have today. Some of it was due to the site loss (~10k) but the rest was divested out of fear (~25k). The scared are to be considered weeded out. Those left behind willingly digested a 14% risk for a longterm profit that is proportionately higher than Doog's (~50-60% less than what they expected to make) due to the fact that the bankroll shrank so dramatically in size.

Our investor risk adjusted profit would be more in line with Doog's if the bankroll didn't vary too much over time. In the case of the Nakowa event just described, the investors profited off the risk more than Doog did. However, if the site concedes a large loss on a timescale similar to a large divestment wave, the "long term" investors would catch the short end of the stick.

A "long term" investor must therefore demand a reasonably stabilized bank roll and we should discuss how to enable such a thing. Ultimately, aligning the risk-adjusted profits of both long term investors and Doog should be the proper way to proceed if we are to give Doog's trust objective value at least in defining how the site operates.

As a fun little aside one could ask if Doog is being remunerated properly for his services, particularly when compared to the amount of money he can run off with. Judging from the current site stats, Doog could choose to run off with 35k now or expect to make ~5k over the course of a year, eventually making more than he could steal over a 7 year period.

Is this convenient for Doog? Legally, how long can we realistically expect the site to exist for? If that number is less than Doog's expected turnover on stealing bank, then we should be worried. Under the fee scheme, his turnover would be more like two years, possibly giving us more peace of mind.

Current system is so natural, and mimics the classic roulette system that's been in place for a few hundred years. Players on roulette perfectly understand that 0 is the house edge, everything else is eventually returned to them by laws of probability and everybody is OK with that. There are no goddamned commissions for every turn of the wheel. At Just-dice, 1% edge replaces 0 on roulette, and everything else follows this model. Dooglus keeps the operation go smoothly and collects his % of the winnings, just like the management and staff are paid to keep the real casino operational. Releasing the management (dooglus) of obligation that everything runs with profits is not a very wise move for the owners of the casino/JD. By the new proposed system dooglus would not have to watch if the site is profitable or not, because he would collect his commissions regardless of that. I'm just a small investor, but if new system would be eventually put in place, I would divest immediately.

Not only, someone could hack your server without you knowing it and slowly cheating by making winning bets. I am sure you would notice it faster if you start losing money as well
Also there are all the other possibilities that neither you and I are thinking about.

I'm an investor that understands the change. I wouldn't object to keeping it the same or implementing the new system. If it eases doog's ability to expand the site via referrals, then that sounds good to me.

Well, unless the reason is that I am nakowa/LiKaShing/TheNextWhale and the reason the site is losing is that I'm cheating.

See http://forumserver.twoplustwo.com/25/probability/bitcoin-dice-site-called-just-dice-com-rigged-not-1410413/ for more on this subject...

Oh, one other thing.

I've been out of the house for 24 hours or so, and left the cold wallet at home.

I brought 1500 BTC with me on my trip to refill the hot wallet, thinking that would be plenty, but I've run out already.

It seems I timed it badly.  Lots of investors have been divesting and withdrawing recently.  I'm guessing a bunch finally got back into profit after suffering large losses to nakowa and have been waiting for that point to quit the site.

Anyway, I'll not be home for a few hours still, so the site's hot wallet will likely be empty for a while unless anyone has some coins available and can help me refill it.  There's no need to invest them (but you can if you like); just depositing to your deposit address for a few hours will help.  I know investors mostly understand about the hot wallet, and balancing security with convenience, but it doesn't look good to players when they're unable to withdraw their winnings.

Once coins start appearing in 122BNoyhmuUt9G9mdEm3mN4nb73c1UgNKt it means the hot wallet is full enough.  That's the overflow, and I have access to that wallet from here.

Well, in the old system then we still have a credit of 10btc so over the long run it evens out and doesnt make any difference in terms of return. And as pointed out the major advantage (I believe why everyone agree with me) is that you are in the same boat and if for whatever reason the site loses you will be much more proactive in making it stop.

I'm a little disappointed that after finally changing my mind about the best way to collect commission, none of you investors like it and want to keep things the same.

The proposed new system looks to me like it is slightly better for investors if anything, as shown in this (exaggerated) example:

--
Suppose: one week nakowa bets 5k and loses 200 BTC (150 more than expected), and the next week he bets 5k and wins 100 BTC (150 more than expected).

old system: investors pay 20 BTC at the end of the first week, 0 at the end of the 2nd week, and have a net profit of 100-20 = 80

new system: investors pay 5 BTC each week and have a net profit of 100-10 = 90
--

In other news, I just upgraded the Just-Dice server so hopefully it won't get bogged down and laggy like it used to.

I vote for the current system.

I got negative with Nakowa and just sunday got finally even (wow!). It helped a lot (psychologically) to have a lower variance when I was at a loss as an investor and the fact the Doog is "part of the game" is somehow reassuring.

+1 For the current/old system, as an investor on doge-dice.com.

+1 as well
(I'm an investor too, obviously)

+1. I like the fact that doog is along for the ride.

As an investor on the site, and although i do understand the points in favor of the change, I do prefer the current system as Doog has interests more in line with investors and will investigate if some players win "too much" ...

Wow! Great plot Peter.
Doog can this be added to the site stats?
We should have better site stats. Things like:

Bet size probability distribution
Bet percentage distribution
A nakowa information tab so that whenever the question "why are we so far below the true 1% line" arises, we can direct people to the relevant information

Once all this information. Is freely available, it would be nice to have full probability projections for the site's profit based on the user's general betting patterns. A plot like the one Peter just posted related to the investor's profit would add a serious touch of extra "legitness" to JD. I'd be willing to write a python code to model all the data myself if no one else wants to do it.

Last but not least, I apologize for my liberal use of the "1% line" term. Generally I'm referring to the post-Nakowa projected profits as opposed to the profits we should have from the beginning.

Nakowa's winnings weren't that improbable.  In September, I calculated his profit and superposed it over the associated probability density function (see below).  Nakowa's win sequence was no better than a 2-sigma event, and this was prior to his losses a few months later.  Also remember that Mechs lost 7000 BTC in what was probably a 1 sigma event in the other direction (good for the house).  

Yes, but...  some people divested all their coins after losing to nakowa.  They could reinvest at any time and have months of commission-free profit.  Others divest most of their coins after nakowa and are still slowly recovering, making semi-regular profit and paying no commission.  There's also the effect caused by nakowa being invested whenever he wasn't playing, but divesting while he played.  That meant that everyone except him bore the brunt of his wins, so just getting back to pre-nakowa site profit levels isn't enough for investors to get back to peak profit.

Also, what is this "fall back to the 1% line" thing?  We've been around 0.2 to 0.3% profit for a couple of months now, it seems to me.  It will take a long time to get back near 1%, if it ever happens.  I guess I misunderstood what you're suggesting we wait for.


However I collect commission is unrelated to whether we ever have another 'nakowa' or not.  I can't control whether big players win or not.  If another big player comes and decimates the bankroll it's going to look bad either way I think.  I mean, it already looks pretty bad...


I'm really not concerned about the kind of bankroll volatility we've been seeing.  180 or so is a massive amount to be able to win per roll.  If it drops to 100 BTC max profit per roll would anyone really be upset?  It's still so much bigger than the nearest competition.


I don't like to change things too much, since I don't have a good communication channel with the investors, and so can't easily canvas their opinion.  I guess since I've reserved the right to change the rate of commission with a full week's notice, that a small change like this is also acceptable.  I could put the commission rate up to 90% with a week's notice, so I guess any investor who is at all interested is already watching for such announcements.

And I don't feel like I have a good solution for how to treat those investors who are due a bunch of commission-free profits.  Like the guy who bought the old mechs account, which had some 200 BTC of investment losses, and so is able to earn 200 BTC of investment profit before paying any more commission.

Doog, I thought we already surpassed the pre Nakowa win max site profit. If so just wait for site profit to fall back on the current 1% projected profits line and then switch to charging per bet.

I don't think this scheme deters you necessarily from caring about the investors. If anything it puts you on the spot if site profit ever deviates significantly again from the 1% line. In other words you're gonna lose all credibility if another nakowa event takes place.

Switching to Sebastian's point, keeping the scheme as is could be justified by saying that it deters short term investors. By exposing short term investors to greater downside variance, they should be incentivized to keep their coins invested thus leading to a more stable bankroll.

Charging per bet does make the whole presentation of the site more elegant. You are officially brokeraging for gamblers and taking an honest fee for it. I agree with the monetization of your own personal profit for referrals. Furthermore you could invest in a constant stream of advertising once you have a stable income. It works out better for investors when whales play as they don't have to worry about paying a commission on an unlikely transient house win which will both make them feel like they're making more money and protect them from wild up/down variance in the medium term. In the long term, assuming 1% line is followed, investor profits should be identical to the current commission scheme.

Overall I agree that you should make the final decision. It's your site and no matter what you decide, people will undoubtedly keep on playing and investing. I thank you again for the great service you provide and look forward to contribute to discussions of this kind.