bustabit – The original crash game

[malevolent]

As a side note, as of a few days ago all bankroll investors have seen the announcement.

And yet the bankroll size barely budged. I think it didn't even move at all on bustadice.

What other services do to successfully draw users away is offer subpar experience or make claims of 'getting hacked'. Maybe more people would divest if it was made clear to them what is their current expected return on investment assuming other variables remain unchanged.

edit: see below; the previous divestment saw some changes, but apparently not much as devans was expecting

[1714009872]

1714009872

[1714009872]

1714009872

[1714009872]

1714009872

[1714009872]

1714009872

[GamblingSiteFinder]

And yet the bankroll size barely budged. I think it didn't even move at all on bustadice.

Unless you're referring to the previous calculation adjustment, I imagine investors have not taken action after reading this news because the proposed changes do not take effect until March 2nd, 2021.

We will probably see large divestments in the days leading up to the March 2nd, 2021 bankroll calculation change.

[JollyGood]

A potential for any form of divestment on any scale that has an impact can only take place if investors know what estimates to expect and how they would be compared to their previous and current calculations. No matter how many investors might divest the website/business is far too big for it to have a negative impact in the long run.
 

And yet the bankroll size barely budged. I think it didn't even move at all on bustadice.

Unless you're referring to the previous calculation adjustment, I imagine investors have not taken action after reading this news because the proposed changes do not take effect until March 2nd, 2021.

We will probably see large divestments in the days leading up to the March 2nd, 2021 bankroll calculation change.

[ironhak]

Does bustabit behave differently when I bet?

Sorry for the common question. I read the "provably fair" section of the site, I also read that bustabit don't behave based on the users bet... Then I have a question.

I have a script, I've let it run for 1 week straight all day 24h/24 in simulation mode and it didn't loose. Then, after that week I ran the same script but with real money, it survived 3 days and then it lost, so I've put more money and the second time it ran for <10hours before loosing.

Now I'm re-running the same script in simulation mode since ~10 days, always 24h/24, and didn't lost so far.  Has anyone experienced this type of situation? Was it a coincidence or not?

[malevolent]

I have a script, I've let it run for 1 week straight all day 24h/24 in simulation mode and it didn't loose. Then, after that week I ran the same script but with real money, it survived 3 days and then it lost, so I've put more money and the second time it ran for <10hours before loosing.

Now I'm re-running the same script in simulation mode since ~10 days, always 24h/24, and didn't lost so far.  Has anyone experienced this type of situation? Was it a coincidence or not?

It's called variance, I don't know what percentage of the bankroll you risk per bet, but the range 10-240 hours really doesn't stand out to me as anomalous.

[ironhak]

I know wht variance is.

But 2 weeks with no loss on simulation (it's two weeks because I stopped the script, not because it lost).
~10hours with real balance.

To me it is a fairly big discrepance

[tagrn1]

I know wht variance is.

But 2 weeks with no loss on simulation (it's two weeks because I stopped the script, not because it lost).
~10hours with real balance.

To me it is a fairly big discrepance

The way to prove it to yourself would be to run the simulation and live scripts at the same time

[RHavar]

I know wht variance is.

But 2 weeks with no loss on simulation (it's two weeks because I stopped the script, not because it lost).
~10hours with real balance.

To me it is a fairly big discrepance

Try run the simulation script again against the same hashes used by the game. The result should be *identical* (assuming no issues like connection loss, etc.)

Then after verifying that, try verify the game hashes to make sure they are fair. One easy enough way is to hash each game hash $gameId times [1] and you should arrive to the same $terminatingHash (from the seeding event).




[1] (Or it might be $gameId+1 or $gameId-1), too lazy to think about it 

[higgg]

I'm an investor who checks occasionally and I just saw this newer change--or likely as not I saw it before and ignored it as I thought it was something about the previous change. Anyway I just noticed it and came here. Some off-the-cuff thoughts:

• Agree that Daniel can do whatever he wants, and I understand the goal of wanting to reduce the bankroll.
• More transparency showing this moving XDR/BTC average and what the commission would be today, would be in the past, and would be in the future if the XDR/BTC rate stays the same is needed.
• It seems like most of the discussion here has been around this XDR/BTC rate, but if I'm not mistaken, this new formula is also the first time the size of the bankroll has been a variable in the commission calculation. To me, setting a "target" size for the bankroll is a bigger change than using this particular rate to set the target size.
• I think the target sets up an equilibrium/expectations problem. The higher the bankroll, the higher the commission, and the lower the rate of return to investors. And, there will be an immediate jump in the commission in the near future. So it may make sense to divest. But I am inclined to think, why not give it a few months and see if other people divest so that the bankroll goes down and my stake as a percentage goes up. But if everyone thinks that, no one will divest, and we'll all be caught in a waiting game.
• Therefore, Daniel I would suggest that if you *want* people to divest, you do something like waive the 2% fee for reinvestments for some period of time. Right now I do not want to divest because if the bankroll subsequently shrinks enough that investing becomes attractive again, I would have to pay the 2% fee again. If I didn't have that worry, I would be more inclined to pull some or all of my stake now--which you want to happen.
• I notice that setting the commission as a proportion of a target size really amplifies the effect of changes in the bankroll. In the past, if the size of the bankroll went up by 1%, my expected return would go down by roughly 1% because of dilution. Now, let's say the commission is at 75% under the new formula, and the size of the bankroll goes up 1%: I still take the 1% hit to dilution, but now investors' share of house profit is reduced from 25% to 24.25%, which is actually a 3% change -- so my expected return is down a whopping 4% from just a 1% change in bankroll. So if I stay in, I'll feel a lot more pressure to check the bankroll size often.

Of course fundamentally the problem is that bustabit does very well, and there are not a ton of other great alternatives out there. You can get 6% yearly with BlockFi, but only up to 2.5 BTC, and above that it's just 3%. Depending on how you assess the risk that bustabit will vanish in a poof of smoke, there's potentially a long way to go before it makes sense to park btc elsewhere.

[edited to remove blank lines from the bottom -- oops!]

[devans]

This CSV file contains some bankroll stats for every day as well as BTCXDR on that day and the 360-day moving average: https://gist.github.com/danielevns/1adbd7297c26988d6b7551a6032bdb4f. Starting in a few days on the 14th, a dataset like this will be available on the website and automatically kept up to date. We'll also start showing the current and expected commission rates in the bankroll overview.

Today the moving average of Bitcoin's price is 10,059.938058 XDR, so if the new formula for the commission were already in effect today the commission rate would be approximately 74.2%. If neither the price nor the size of the bankroll changed at all for 360 days then the commission rate would be about 238% (capped at 100%) as the bankroll is way over the target size at the moment.

For what it's worth the commission rate has been determined by the size of the bankroll since December 2020, but currently it doesn't take into account the value of the bankroll.

Decreasing or suspending the dilution fee is problematic because existing investors paid this fee with the understanding that future investors would in turn pay this fee to them. Although I understand your line of thought, I am not convinced that reducing the barrier to entry would lead to a net divestment. In any case, it won't be necessary as the rising commission rate will create sufficient incentive for bankroll investors to divest eventually. Some investors like yourself will be willing to temporarily tolerate lower expected returns at a commission rate of ~75% in the hopes that other investors divest before them, increasing returns again. But would you still be willing to stay invested at a commission rate of 80%, 85% or even 90%? I am confident that we'll see significant divestments after the new commission rate comes into effect fairly quickly, if not right before.

[devans]

A potential for any form of divestment on any scale that has an impact can only take place if investors know what estimates to expect and how they would be compared to their previous and current calculations. No matter how many investors might divest the website/business is far too big for it to have a negative impact in the long run.

bustabit saw gross gaming revenue of 26,851,012 XDR in 2020. Let's assume that the gross gaming revenue will be the same going forwards. With the current commission rate of 40.55% an estimate for the bankroll's annual return under the current system would be 15,962,926 XDR, which represents a return of 12.17% on its current value of 131,120,101 XDR. If the new commission structure were in place today, the estimated return would only be 6,927,561 XDR or 5.28%.

However, that doesn't account for the commission rate reacting to size of the bankroll or Bitcoin's price, investments or divestments being made, the gross gaming revenue being smaller or larger compared to last year etc. Estimates for the bankroll's expected rate of return can vary wildly depending on what assumptions you make and every investor needs to decide for themselves what assumptions they consider to be reasonable.

[Timetwister]

How much was paid in XDR in dilution fees in the last 365 days?

[devans]

I don't have an easy way to query for the exact number because when someone invests and pays the dilution fee they always receive part of that fee themselves. At most it was 443,743 XDR in the past 365 days.

[RealMalatesta]

Does bustabit behave differently when I bet?

Sorry for the common question. I read the "provably fair" section of the site, I also read that bustabit don't behave based on the users bet... Then I have a question.

I have a script, I've let it run for 1 week straight all day 24h/24 in simulation mode and it didn't loose. Then, after that week I ran the same script but with real money, it survived 3 days and then it lost, so I've put more money and the second time it ran for <10hours before loosing.

Now I'm re-running the same script in simulation mode since ~10 days, always 24h/24, and didn't lost so far.  Has anyone experienced this type of situation? Was it a coincidence or not?

Also I am not entirely sure about the simulation you have could be 100% correct neither. Of course just like everyone said its variance and you should be either losing in 1 hour or winning for 10+ days and it is all possible, remember this is gambling and even though there is a mathematical possibility that you will always lose, there is also chance that you may win a lot longer than you expect and get out before that losing starts, the "you will 100% lose because of house edge" may not hit for more than you gamble, so you may actually get out before it happens.

So, if you gamble on simulation and win, and you gamble on real and lose, that means that mathematical loss happened in the real one, and only "hasn't happened yet" on the simulation, keep it going as long as you can and you will see that it will definitely lose there too because that is how gambling works.

[JollyGood]

Thank you for the explanation.

The assumptions you refer to are essential for any investor to keep in mind and as you rightly stated there is no consensus on exactly what would be considered as reasonable assumption. Everybody has their own opinion.

A potential for any form of divestment on any scale that has an impact can only take place if investors know what estimates to expect and how they would be compared to their previous and current calculations. No matter how many investors might divest the website/business is far too big for it to have a negative impact in the long run.

bustabit saw gross gaming revenue of 26,851,012 XDR in 2020. Let's assume that the gross gaming revenue will be the same going forwards. With the current commission rate of 40.55% an estimate for the bankroll's annual return under the current system would be 15,962,926 XDR, which represents a return of 12.17% on its current value of 131,120,101 XDR. If the new commission structure were in place today, the estimated return would only be 6,927,561 XDR or 5.28%.

However, that doesn't account for the commission rate reacting to size of the bankroll or Bitcoin's price, investments or divestments being made, the gross gaming revenue being smaller or larger compared to last year etc. Estimates for the bankroll's expected rate of return can vary wildly depending on what assumptions you make and every investor needs to decide for themselves what assumptions they consider to be reasonable.

[higgg]

Decreasing or suspending the dilution fee is problematic because existing investors paid this fee with the understanding that future investors would in turn pay this fee to them.

This makes me realize that I never totally understood how this works--so when someone invests say 1 BTC to the bankroll, their portion of the bankroll is 0.98 BTC and the other 0.02 BTC is divided proportionally among all existing investors?

[RHavar]

This makes me realize that I never totally understood how this works--so when someone invests say 1 BTC to the bankroll, their portion of the bankroll is 0.98 BTC and the other 0.02 BTC is divided proportionally among all existing investors?

Imagine the existing bankroll is 1 BTC, and you are going to invest 1 BTC.

You are given a stake of:

($yourInvestment * 0.98) / ($bankroll + ($yourInvestment * 0.98))
= 0.98 / 1.98
= 49.49%

But the bankroll increases by 1 BTC (AKA the new bankroll is 2 BTC)


So your stake is worth 0.98989 BTC.  (AKA you only lost ~1% even though the dilution fee was 2%)


---

The high-level reasoning behind it is that if you didn't recoup some of your own dilution fee, it would incentivize people to "fragment" their large investment into tiny ones (possibly over multiple accounts?).  So the dilution fee is split amongst all investors (including yourself), so that the smart way to invest a lot of money is just simply invest a lot of money.

[higgg]

This makes me realize that I never totally understood how this works--so when someone invests say 1 BTC to the bankroll, their portion of the bankroll is 0.98 BTC and the other 0.02 BTC is divided proportionally among all existing investors?

Imagine the existing bankroll is 1 BTC, and you are going to invest 1 BTC.

You are given a stake of:

($yourInvestment * 0.98) / ($bankroll + ($yourInvestment * 0.98))
= 0.98 / 1.98
= 49.49%

But the bankroll increases by 1 BTC (AKA the new bankroll is 2 BTC)


So your stake is worth 0.98989 BTC.  (AKA you only lost ~1% even though the dilution fee was 2%)


---

The high-level reasoning behind it is that if you didn't recoup some of your own dilution fee, it would incentivize people to "fragment" their large investment into tiny ones (possibly over multiple accounts?).  So the dilution fee is split amongst all investors (including yourself), so that the smart way to invest a lot of money is just simply invest a lot of money.

Ok thanks this is very helpful. I'm still not quite there yet, what I'm not exactly clear on is how the dilution helps the existing investors. In your example, let's say that the bankroll was previously split between two people with 0.5 BTC each, so they each previously had a 50% stake. Without any dilution they would expect to have stakes of 25% after the new investment of 1 BTC.

Are their stakes now instead 0.5/1.98 = 25.2525...%? Since the bankroll is actually 2 BTC, this would give them 0.505050... BTC each, in other words their personal stake increased by 0.005050...BTC upon the new 1 BTC investment.

Indeed 2*0.005050... = 0.0101... which is how much the new investor had deducted from his investment so that works out, so I'm guessing this must be what happens.

[RHavar]

Without any dilution they would expect to have stakes of 25% after the new investment of 1 BTC. Are their stakes now instead 0.5/1.98 = 25.2525%?

Yeah, exactly. I think of it as the investor is first investing 98% of his funds  ... and everyone's stake gets adjusted according to that. And THEN the 2% gets added to the bankroll (without affecting anyones stake)

[higgg]

Without any dilution they would expect to have stakes of 25% after the new investment of 1 BTC. Are their stakes now instead 0.5/1.98 = 25.2525%?

Yeah, exactly. I think of it as the investor is first investing 98% of his funds  ... and everyone's stake gets adjusted according to that. And THEN the 2% gets added to the bankroll (without affecting anyones stake)

Clever. How'd you settle on 2% as the right amount?