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

[dooglus]

The answer is to allow some investors to choose a fixed percentage return of .75 percent.  For instance, let's say the profit of the site averaged 2 percent for a given week.  The investors that choose the fixed percent would get paid .75 percent and the ones who took the rest of the downside risk would get 2 percent plus .25 for paying the fixed percentage.

In weeks the site has a loss the investors that took the fixed percentage of bets would get paid .75 from the investors that didn't.  In this way some investors can lower their variance by giving up some profit and other investors can profit from that.

This is a +ev decision for the investors taking more risk because it ups their longterm returns to 1.25 percent.  

This has been discussed several times before.  I think Will Phase came up with the idea first, possibly in this very thread.

It doesn't work, and here's why:

Anyone getting a guaranteed return isn't contributing to the site.  If you invest 1 BTC today and I have to give you more than that back tomorrow then what use is your 1 BTC?  I can't use it in the bankroll because I might lose it.  I'm better off just telling you I don't want your 1 BTC, because then I don't have the liability of paying you any interest.  There's no upside for me to having investors who aren't willing to take a loss.

Look at it this way:  suppose I have just 2 investors.  One has a fixed return of 0.5% of 1% of the amount wagered.  The other gets the rest (he expects to get his own 1% of turnover plus the other 0.5% of turnover that isn't going to the other guy).

Suppose both players "invest" 100 BTC, and on an average week we have 2000 BTC wagered, and keep the expected 1% of it, ie. 20 BTC.  Of that 20 BTC profit, the fixed-return guy gets 5%, and the other guy gets 15% of it.  That's cool - the fixed return guy gets a good return, and the regular risk-taking guy gets 3 times as much, rewarding him for taking a risk.  It looks like a good scheme, if that's as far as you look.

But what if we have a bad week, and after 2000 BTC was wagered, the site ends up 20 BTC *down*?  The fixed-return guy still gets his 0.5% of 50% of the 2000 BTC wagered, or 5 BTC, and the risk taker ends up 25 BTC down, since he has to pay for the house loss as well as the fixed-return guy's profit for the week.

Finally, since we considered what happens when the result is 40 BTC worse than expected, what when it's 40 BTC better than expected, and we end up with a profit of 60 BTC from 2000 BTC wagered?  The fixed-return guy gets 5 BTC, as he always does when 2000 BTC is wagered.  So the risk taker gets the other 55 BTC.

In summary, how is each person's return in bad, neutral, and good weeks?

fixed return: 5, 5, 5
risk taker: -25, 15, 55

And if the fixed return guy just wasn't involved, the risk taker would get the whole amount to himself:

single regular investor: -20, 20, 60

We can see that the presence of the fixed-return guy has done nothing but drain 5 BTC per week from the other guy's profits, positive or negative.

So what use it he?  We can't increase the max profit per bet due to the fixed-return guy's investment.  Because he's not risking any of his coins, we can't use them to pay out winners.  He's simply a parasite on the regular investors.

Also, assuming this is you in the chat:

16:45:49 (416053) <whale> I have some people analyzing the site now
16:46:13 (416053) <whale> see the average bet is .02 and 500 million bets
16:46:22 (416053) <whale> doesn't add up even with the big bettors
16:46:33 (416053) <whale> percentage shoudl be at least .60

Your math is wrong.  You can't assume all the bets are the same size.  Here's an exaggerated example which shows why:

Suppose there were a hundred bets and that 99 of them were for 0 BTC and one was for 100 BTC.  The 100 BTC bet won at 49.5% and the site profit is 100%.

The average bet size is 1 BTC.

You run your simulation and find out that there's no way after 100 bets of 1 BTC each that the site profit should be 100%, but it is.

Do you see the problem?  Calculating the "average bet" is meaningless.  The "average" couple has 2.3 children.  There weren't 100 bets of 1 BTC each.  There was a single bet of 100 BTC and 99 bets of zero.

Again, this has been gone over in detail in this thread already, mostly when nakowa was playing heavily.

[dooglus]

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.

[willphase]

The answer is to allow some investors to choose a fixed percentage return of .75 percent.  For instance, let's say the profit of the site averaged 2 percent for a given week.  The investors that choose the fixed percent would get paid .75 percent and the ones who took the rest of the downside risk would get 2 percent plus .25 for paying the fixed percentage.

In weeks the site has a loss the investors that took the fixed percentage of bets would get paid .75 from the investors that didn't.  In this way some investors can lower their variance by giving up some profit and other investors can profit from that.

This is a +ev decision for the investors taking more risk because it ups their longterm returns to 1.25 percent.  

This has been discussed several times before.  I think Will Phase came up with the idea first, possibly in this very thread.

It doesn't work, and here's why:

Doog is right - it doesn't work.  If you want less risk, then just invest less.

Will

[zipmaster]

Dear Dooglus,
I'd lie to reiterate a point discussed in the trollbox tonight so that you may read it over at your convenience and answer it at your leisure. A few users complained that your cut over the site is unfair to the investors due to their taking all the risk. It is true that you run the site and you have done such a spectacular job at it all throughout but it would be nice to come to reconsider the distribution of profits as the site and its usage evolves.

One suggestion that I proposed was to drop the percentage cut from the investors all together and instead charge a flat fee to bet. Such a fee would be reminiscent of a stock brokerage service where you allow bettors and bankrollers to contract a bet on your site.

A flat fee furthermore would weed out the monstruos amount of satoshi/dust bettors which wreak lag on the website and populate the trollbox with... well.... more trolls. 99% of the site's profit must be generated by wagered bets larger than 1 mBTC at least (could we get a fact check on this).

How about you charge a flat fee from now on of 10-200 uBTC (that is micron-btc 10^(-6)) to bet on JD. Based on the amounts wagered up to now, a 250 uBTC flat fee would have earned you the same amount of coins. As the gold price (measured in bitcoins) increases over time the fee can be reduced such that every ~50-100k bets you can expect to have earned an ounce of gold.

This would work well for investors in that now they would reap the full profit of their risk and you sir, for guaranteeing such a consistently phenomenal service. Furthermore you would be incentivized to advertise the site and lure new bettors mid-range bettors.

As a further plus to the site, we would experience less trolling, less spamming, less begging and, more importantly, LESS LAG.

I look forward to hearing your thoughts on this.

[zipmaster]

P.S. you could even make the flat fee completely to your control so that you may explicitly tune the amount of traffic to your site

[dooglus]

I look forward to hearing your thoughts on this.

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?

[dooglus]

Also, assuming this is you in the chat:

16:45:49 (416053) <whale> I have some people analyzing the site now
16:46:13 (416053) <whale> see the average bet is .02 and 500 million bets
16:46:22 (416053) <whale> doesn't add up even with the big bettors
16:46:33 (416053) <whale> percentage shoudl be at least .60

Your math is wrong.  You can't assume all the bets are the same size.

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:

[zipmaster]

I'm not "whale" (was simply in the chatroom hearing him abnoxiously spaz yesterday), but thank you fr the wonderful plots. You should plot the derivative with respect to the betting amount instead of the cumulative distributions you just gave us. It would give us a better idea of the standard range within which the vast majority of bets are placed. For example, if 99.9% of bets were due to amounts wagered between 0.01 and 50 btc, it would be in the investors best interest to concentrate on these bettors and penalize the tails (ultra whale bettors and micro-dust bettors) by both introducing a flat fee and considering about lowering the max bet percentage.

I don't think competition due to the flat fee would be much of an issue granted that the vast majority of bettors bet amounts three orders of magnitude larger than the fee. If you're worried about blocking off new users which might simply want to test the site, give a promotional amount of bets at no, or reduced, fee. Hell, the brockerage fee could be charged to the investors for bet amounts above a certain threshold.

I'm not saying that these changes would allow investors to not take risks (I'm not stupid like "whale" who expected a bond like return from his JD profits). What I'm asking is that they reap the full profit of the risk they take.

The site really does work wonderfully, I just think maybe it could be fine tuned a little bit.

[Timetwister]

I'm not "whale" (was simply in the chatroom hearing him abnoxiously spaz yesterday), but thank you fr the wonderful plots. You should plot the derivative with respect to the betting amount instead of the cumulative distributions you just gave us. It would give us a better idea of the standard range within which the vast majority of bets are placed. For example, if 99.9% of bets were due to amounts wagered between 0.01 and 50 btc, it would be in the investors best interest to concentrate on these bettors and penalize the tails (ultra whale bettors and micro-dust bettors) by both introducing a flat fee and considering about lowering the max bet percentage.

I wish we had many more "ultra whale bettors"... And I don't see any problem with "micro-dust bettors". We want to increase wagered bet, not decrease it, and we shouldn't force players to bet in a specific way, but give them as many options as possible to encourage betting.

[Timetwister]

Shortly after launching Doge-Dice, I made a second account and invested 1 million DOGE, with a view to never touching the investment, just so I would have an easy answer to the common question "if someone invested X on day one, how would they be doing now?"

Here's how:



Edit: in case that isn't clear, it shows a 36% return over a one month period.  That's better than Pirate's BS&T ponzi!

Edit2: Re: "the slice of the pie is getting really small..." I agree.  My slice has shrunk from over 13% to a little over 0.1%.

Doge-Dice is super profitable, it was the only reason why I bought Dogecoins to invest them there when the casino was released. But I finally sold them, as I think its purchasing power will eventually decrease at a higher rate than my investment in terms of Dogecoins would grow. That coin is just a fad.

[zipmaster]

To be honest TimeTwister I agree with you. Ultra whale bettors are what makes this site move up and down. They even allow day trading on JD for some people. I'm just throwing out ideas on how the site could be fine tuned that's all.

Another thing Doog could do is charge a flat fee which is some percentage of the bet amount.
Something like, if the wagered amount is above a certain threshold, then charge 0.1-0.01% of the wagered amount. This way dust bettors could keep on betting freely and Doog is incentivized to market the site for larger bettors.

There are a million different ways in which Doog can keep on getting paid for the spectacular job he does. Furthermore, he deserves a steady income for the services he provides. We investors, on the other hand have got money on the line. If the house edge is 1%, that is the risk I am taking. I should gain profits that match the risk I take.

[Equilux]

If the house edge is 1%, that is the risk I am taking. I should gain profits that match the risk I take.

Aren't you?

[SebastianJu]

Dear Dooglus,
I'd lie to reiterate a point discussed in the trollbox tonight so that you may read it over at your convenience and answer it at your leisure. A few users complained that your cut over the site is unfair to the investors due to their taking all the risk. It is true that you run the site and you have done such a spectacular job at it all throughout but it would be nice to come to reconsider the distribution of profits as the site and its usage evolves.

One suggestion that I proposed was to drop the percentage cut from the investors all together and instead charge a flat fee to bet. Such a fee would be reminiscent of a stock brokerage service where you allow bettors and bankrollers to contract a bet on your site.

A flat fee furthermore would weed out the monstruos amount of satoshi/dust bettors which wreak lag on the website and populate the trollbox with... well.... more trolls. 99% of the site's profit must be generated by wagered bets larger than 1 mBTC at least (could we get a fact check on this).

How about you charge a flat fee from now on of 10-200 uBTC (that is micron-btc 10^(-6)) to bet on JD. Based on the amounts wagered up to now, a 250 uBTC flat fee would have earned you the same amount of coins. As the gold price (measured in bitcoins) increases over time the fee can be reduced such that every ~50-100k bets you can expect to have earned an ounce of gold.

This would work well for investors in that now they would reap the full profit of their risk and you sir, for guaranteeing such a consistently phenomenal service. Furthermore you would be incentivized to advertise the site and lure new bettors mid-range bettors.

As a further plus to the site, we would experience less trolling, less spamming, less begging and, more importantly, LESS LAG.

I look forward to hearing your thoughts on this.

Bad idea... people that see a fee are gone. A house edge is a way better, because normal and hidden "fee".

I think 10% is pretty fair for dooglus. He allows us to invest. Not like other big dice sites that only make profit for single persons. So a fair share is a good trade. And why cut it down? Dooglus would lose his incentive to attract more players.

Its not a good idea trying to take save bitcoins from dooglus loss. At the end you already can keep 90%. Tell me a casino where you can do anything like that. So dont fight the happiness of dooglus.

[Timetwister]

Bad idea... people that see a fee are gone. A house edge is a way better, because normal and hidden "fee".

I think 10% is pretty fair for dooglus. He allows us to invest. Not like other big dice sites that only make profit for single persons. So a fair share is a good trade. And why cut it down? Dooglus would lose his incentive to attract more players.

Its not a good idea trying to take save bitcoins from dooglus loss. At the end you already can keep 90%. Tell me a casino where you can do anything like that. So dont fight the happiness of dooglus.

I also really like the current system and I admire Dooglus for what he has accomplished. If anything, I'd add more options to players and marketing.

[zipmaster]

I'm not trying to take anything a way from Doog. If anything I think he deserves a more consistent stream of income for his services independently on whether the site wins or loses as a whole.

Charging uBTC or 0.01% fees for amounts wagered above 10mBTC should be totally acceptable. Ultimately the investor is receiving a very valuable service. Where else on the internet does one have the opportunity to play such a low house edge against such a large bankroll? There HAS to be value in that. Primedice and Satoshidice offer max profits of 20-40 btc. It's nothing compared to what JD offers.

Look, I'm not saying the amounts I'm proposing are perfect. I'm just trying to point out that there is ample wiggle room to discuss among investors and Doog what the best course of action is. Ultimately, though, the gambling profits should be totally distributed among the people who are risking the coins: a.k.a the investors.

Again, I'm not attacking Doog in any way and I value his work very much. The man is always available and present. He's knowledgeable and very able at what he does. I would him up there along with Friedcat and Mike Caldwell among the most trustworthy forum entrepreneurs. We need to ask ourselves what service is Doog offering and how should he be remunerated.

The way I see it, he gives us an awesome brokerage service to coordinate bets between gamblers and bankrollers. He deserves to be paid for that independently on whether the site wins or loses and it should be worthwhile for him also!

All I ask is that at least an investor referendum is organized so that this can be discussed more openly among all investors so that a consensus can be found.

[itod]

Look, I'm not saying the amounts I'm proposing are perfect. I'm just trying to point out that there is ample wiggle room to discuss among investors and Doog what the best course of action is.

Additional fixed fee for every bet is not a good idea, people are playing martingale like crazy and adding such a fee would totally screw their martingale calculations, bots, etc. The main advantage of JD over SatoshiDice (as players are concerned) is that you can martingale without any fees. Remove that and you would have many unhappy customers. IMHO "the best course of action" is no action.

[dooglus]

IMHO "the best course of action" is no action.

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.

[zipmaster]

Before answering Doog I want to quickly confute the martingale argument. Charging a variable fee that is 0.01-0.1% or, more simply, a flat fee of ~10 uBTC wouldn't alter the martingale strategies at all. It would at most penalize the dust bettors of which:

a) We choose to not care since they are contributing less than 0.1% of total amounts wagered on the site
OR
b) we allow them to play for free as long as they're betting amounts that are lower than, say, 1mBTC

In either case, the most generally played martingales would be hardly affected in case of the flat fee. They would be slightly affected but not by much in case of the variable fee. (I personally prefer flat fee).

Now, on to Doog. You wouldn't be disincentivized by the new profit scheme. With a flat fee being paid, your job becomes that of enabling large amounts of bets which, ultimately, benefit the investors also since the 1% house edge is more and more guaranteed the more bets (not necessarily wagered amounts) are placed on the site. One variant of the flat fee could even be that when the bankroll wins a bet, the investors pay the fee for that bet while when the player wins the player pays.

It is not true that overtime investors would be paying the same amount. The current scheme places friction on the upside of the bankroll while exposing us investors to a larger downside. If some whale comes in and chooses to place subsequent 100+ btc bets, we the investors are risking huge downside. This is especially true since large bets are few and far between and the law of large numbers doesn't really apply to them. In these circumstances the investors are just stuck with the variance of the whale's betting strategy. If we have to pay a 10% fee on any profit we make in these circumstances, the EV for the investors is easily made negative.

Just follow the math. Let P be the probability of a bank win off of a given bet and Q be that of a loss. For convenience let M be the multiplier of the bet. The 1% house edge boils down to a very simple relationship between Q and M, namely:

QxM=0.99

In the absence of a Doog fee, the EV of a bet is (in units of the betting amount):

EV = P-(M-1)xQ = P-[(0.99/Q)-1]xQ = P-0.99+Q = P-0.99+(1-P) = 0.01

Where the normalization condition P+Q=1 was used in the last step. Let's now factor in Doog's cut:

EV = 0.9xP-(M-1)xQ = 0.01-0.1xP

It's easily seen that whenever P is greater than 0.1 (i.e. anything less than 90% bet from the bettor's perspective) the EV for the bankroll is negative. This would confirm the general perception in the trollbox that 90+% betting schemes have been very profitable for the house. Mind you that this doesn't imply anything for the gambler... he's playing the normal house odds with negative EV=-0.01.
Hence, at the current state of affairs, the investors can expect to lose money whenever gamblers bet at less than 90%.

The EV calculation shown above can be generalized for a different "Doog tax" which I will denote as G. Currently, G=0.9. Repeating the calculation results in:

EV=0.01-(1-G)xP

which is negative whenever P>0.01/(1-G).
Clearly, unless G=1, there will always be percentage threshold beyond which the EV for investors is negative. Considering that a lot of gamblers bet at probabilities that are less than 90%, this is something which should be seriously addressed.

Now, in Doog's defense, one might argue that this EV calculation which already considers a large number of bets being made, also assumes that a cut is being taken whenever the house wins. In reality though, Doog takes his cut only once a week thus allowing us to play at a positive EV for a week at a time and, only on sundays, paying the cut to Doog.

This argument can be easily confuted though by compounding all weekly bets into one average bet after which, in case of a bank win, Doog takes his cut. We can then reproduce the same exact calculation for the investor EV as before. The one key parameter that we need is the average betting probability weighted by the bet wager size. DOOG... if you could supply me with this number I'd be very grateful.
In other words, for every bet made on a weekly basis denote P_1,2,...,n the probability at which each bet was placed and K_1,2,...,n the amount wagered. Then the average betting probability is:

<P>=SUM{P_nK_n}/SUM{K_n}

If <P> is less than 90% then we have a serious problem on our hands!

Last but not least, I'd like to tackle the all too easy standard of "if it ain't broke, don't fix it".
Lets stop for a moment and discuss honestly about what it means for something to be broken.

The house profit is a quarter of what it should be. Granted that this was due to a "very lucky" streak by the biggest whale bettor of all times (Nakowa), and that the biggest loss the site incurred happened on a timescale of about a week, the above mathematical EV argument shows that the investors risk to incur a larger burden than they should granted the money that they put up. In some circumstances they even stand to lose with a higher probability than the bettors themselves!

It is for all these reasons that I continue to press for a serious consideration of a flat betting fee in exchange for full profit in the pockets of the investors.

I look forward to receiving the piece of data I requested.

[SebastianJu]

Before answering Doog I want to quickly confute the martingale argument. Charging a variable fee that is 0.01-0.1% or, more simply, a flat fee of ~10 uBTC wouldn't alter the martingale strategies at all. It would at most penalize the dust bettors of which:

a) We choose to not care since they are contributing less than 0.1% of total amounts wagered on the site
OR
b) we allow them to play for free as long as they're betting amounts that are lower than, say, 1mBTC

In either case, the most generally played martingales would be hardly affected in case of the flat fee. They would be slightly affected but not by much in case of the variable fee. (I personally prefer flat fee).

You really underestimate the impact of the word "fee". First... people dont want to pay a fee when they can play for free. They will use the free version elsewhere regardless how high the fee is. Only because there IS a fee.
On top... the player already pay a fee. Its only not a fixed one. Its a mathematically proven fee that enables all casinos worldwide to live. And its a hidden fee that the player dont see. Its the game itself. It would be grotesque if players have to pay a fee to play gambling games. Thats so very much no gambling anymore.

[BayAreaCoins]

If a casino tried to charge me a fee for gambling. I would tell them to fuck off and I would go across the street

ANY "fee"