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

[drawingthesun]

Not a welcome sight for investors.  This illustrates how one whale can dominate the action and skew the stats with just a handful of significant bets.

Well the problem is that we only have one real whale, and its this guy.

All these little bets that happen everyday, 1 Nakowa bet just overshadows an entire day of normal betting. Remember about half the amount of bitcoin wagered on this site is Nakowa.

[1715056411]

1715056411

[1715056411]

1715056411

[1715056411]

1715056411

[Rannasha]

I think all the fancy math and assumptions unneeded.  ...

I disagree.  I think we need fancier math. 

If dooglus can make a database-dump with all bets-by-size-and-chance again, we can compute what the expected standard deviation of the profits is and how (un)likely our current situation is. The previous time we had such a dump, the site was at -2 sigma, so the conclusion was that the house was very unlucky, but not unreasonably so. At that time, we had around half the amount wagered, so it's been a while.

How large was the filesize for such a database-dump?

It wasn't a full dump, it was a table with all combinations of bet-size, bet-probability and count. The link should still be in the thread somehwere, but it wasn't that big.

[drawingthesun]

Has a stop-loss feature been discussed before? Surely I'm not the only investor who has doubts about being exposed to this kind of variance, especially just after big losses. Why not have a user-settable investment threshold, and if a user's investment falls below that level they'd be auto-divested?

I take the viewpoint that too many people are invested anyway. People who want a safe investment should go elsewhere (Don't want to sound mean, but making JD safer for investors is a little odd, its a casino after all). Remember one day Nakowa will lose if he continues to play, and on that day you want to be invested.

[juhakall]

Has a stop-loss feature been discussed before? Surely I'm not the only investor who has doubts about being exposed to this kind of variance, especially just after big losses. Why not have a user-settable investment threshold, and if a user's investment falls below that level they'd be auto-divested?

I take the viewpoint that too many people are invested anyway. People who want a safe investment should go elsewhere (Don't want to sound mean, but making JD safer for investors is a little odd, its a casino after all). Remember one day Nakowa will lose if he continues to play, and on that day you want to be invested.

I figured this would be one reason not to have a stop-loss feature. Maybe it's even a good one, I'm hardly qualified to say. But wouldn't those investors who are willing to take higher variance reap more profits from the whales anyway, if all the scared investors were auto-divested almost immediately when high variance shows up?

[infested999]

Has a stop-loss feature been discussed before? Surely I'm not the only investor who has doubts about being exposed to this kind of variance, especially just after big losses. Why not have a user-settable investment threshold, and if a user's investment falls below that level they'd be auto-divested?

mechs placed a 7 BTC bounty on anyone who makes a script that does this.

[drawingthesun]

I figured this would be one reason not to have a stop-loss feature. Maybe it's even a good one, I'm hardly qualified to say. But wouldn't those investors who are willing to take higher variance reap more profits from the whales anyway, if all the scared investors were auto-divested almost immediately when high variance shows up?

You make a good point.

[bulltrap]

I figured this would be one reason not to have a stop-loss feature. Maybe it's even a good one, I'm hardly qualified to say. But wouldn't those investors who are willing to take higher variance reap more profits from the whales anyway, if all the scared investors were auto-divested almost immediately when high variance shows up?

You make a good point.

Where does the site need investors who divest when large max bet is required? These kind of "investors" just leech the low variance profit and make profitability worse for everyone.

[romerun]

I figured this would be one reason not to have a stop-loss feature. Maybe it's even a good one, I'm hardly qualified to say. But wouldn't those investors who are willing to take higher variance reap more profits from the whales anyway, if all the scared investors were auto-divested almost immediately when high variance shows up?

You make a good point.

Where does the site need investors who divest when large max bet is required? These kind of "investors" just leech the low variance profit and make profitability worse for everyone.

but they will miss the high variance profits, although if most people hate high variance why not lower the max bet to 100 or 50 making site more immune to nakowa. The problem right now is most people are betting less than 5, and that proprietor is doing 100+, screwing the math.

[molecular]

Has a stop-loss feature been discussed before? Surely I'm not the only investor who has doubts about being exposed to this kind of variance, especially just after big losses. Why not have a user-settable investment threshold, and if a user's investment falls below that level they'd be auto-divested?

mechs placed a 7 BTC bounty on anyone who makes a script that does this.

Link? Sounds like easy money to me.

[RationalSpeculator]

I also think the house taking a beating was a good thing. The risk/reward was becoming unfavorable due to too many investors.

But why do many investors try to lower the risk by changing the rules instead of just lowering their invested amount?

Because they know that lowering their invested amount means less potential profit for them. But changing the rules reduces potential profit for all, so they don't miss out while reducing their risk also.


Makes me think of people that go live near an airport with cheap land and good job opportunities then complain about airplanes flying over their house and lobby politicians into more regulation to have less flights over their house.

Sure less flights means you have less risk/annoyance and your real estate will go up but at the same time job opportunities in the region go down.  


You want less risk? Reduce your exposure instead of trying to reduce the potential reward for all of us.

[Oleander]

But why do many investors try to lower the risk by changing the rules instead of just lowering their invested amount?

Lowering the invested amount reduces risk and reward at the same time.  It doesn't change the risk/reward ratio.

Changing the max bet policy might reduce the risk a lot, while lowering the reward a little.  In other words, it might make the risk/reward ratio more favorable.

[RationalSpeculator]

But why do many investors try to lower the risk by changing the rules instead of just lowering their invested amount?

Lowering the invested amount reduces risk and reward at the same time.  It doesn't change the risk/reward ratio.

Changing the max bet policy might reduce the risk a lot, while lowering the reward a little.  In other words, it might make the risk/reward ratio more favorable.

I think it is scientifically proven that your statement is incorrect.

Applying the Kelly Criterion gives you the best risk/reward ratio.

Lowering max bet amount, gives lower volatility indeed, but lowers the potential returns more than the loss in volatitlity.

Ie: you will lower the risk/reward ratio of JD by lowering the max bet amount.


So you want less risk/volatility? Just reduce your exposure. Indeed you will have less potential reward, but the coins that you leave invested will have the best possible risk/reward ratio, and this indeed means high volatility.
  

[infested999]

Has a stop-loss feature been discussed before? Surely I'm not the only investor who has doubts about being exposed to this kind of variance, especially just after big losses. Why not have a user-settable investment threshold, and if a user's investment falls below that level they'd be auto-divested?

mechs placed a 7 BTC bounty on anyone who makes a script that does this.

Link? Sounds like easy money to me.

https://bitcointalk.org/index.php?topic=263522.0

"easy money"

[Oleander]

Applying the Kelly Criterion gives you the best risk/reward ratio.

No, it doesn't.  The Kelly Criterion maximizes the reward, regardless of risk.  Read the Wikipedia article.

The "best" risk/reward ratio is a complicated subject.  There are a number of approaches, such as Sharpe ratio, Treynor ratio, etc.  Your mileage may vary.  But the Kelly Criterion doesn't address this at all.

[trout]

forget about the risk. Kelly's for the win. Win in the long run. Just sit and wait.

The only real problem is that there's only one  whale.

[dooglus]

There we go. How convenient. As you can see, we have just rolled over 2.4 million bets at 500 max bet. Therefore we arrive at the following connundrum:

1. 2,250,000 bets at 500 BTC is enough to guarantee variance within 0.9% < profit < 1.1%.
2. Actual sample size is a minimum of 2,400,000 because not all bets were made at max kelly bet.
3. Actual site profit is less than 0.2%.

The site has had 5 bets of 500 BTC or more.  Not 2,250,000.  I expect that maybe invalidates your conclusions.

Here are the 5 bets:

#7093577 bet 561.29526541 BTC at 98% and won 5.72750179 BTC
#13076611 bet 500.00000000 BTC at 90% and won 50.00000000 BTC
#13856884 bet 500.00000000 BTC at 90% and won 50.00000000 BTC
#13865610 bet 500.00000000 BTC at 90% and won 50.00000000  BTC
#14740956 bet 640.00000000 BTC at 90% and lost 640 BTC

In total, there have been 2073 bets of 100 BTC or more,
901 of which were for 200 BTC or more,
129 of which were for 300 BTC or more,
61 of which were for 400 BTC or more,
5 of which were for 500 BTC or more,
1 of which was for more than 600 BTC.

(Each total includes all following totals).

This is nowhere near your estimate of 2 million bets over 500 BTC.

[dooglus]

If dooglus can make a database-dump with all bets-by-size-and-chance again, we can compute what the expected standard deviation of the profits is and how (un)likely our current situation is. The previous time we had such a dump, the site was at -2 sigma, so the conclusion was that the house was very unlucky, but not unreasonably so. At that time, we had around half the amount wagered, so it's been a while.

I remember before I made several reports.  What would be most useful?

There are two variables: bet size and chance of winning.  Do you want a simple count of bets by stake, ignoring chance?  Or counts grouped by stake and chance?

Here's an example.  It is all bets of 1 BTC or more, grouped by 'bet' (stake) and 'chance'.

http://privatepaste.com/931a6b5a2a

'bet' is the stake in satoshis, and 'chance' is the percentage chance of winning times 10,000.

+----------+-------------+--------+
| count(*) | bet         | chance |
+----------+-------------+--------+
|        1 | 64000000000 | 900000 |
|        1 | 56129526541 | 980000 |
|        3 | 50000000000 | 900000 |

So there was 1 bet of 640 BTC at 90% chance,
1 bet of 561.29xxx BTC at 98% chance,
and 3 bets of 500 BTC at 90% chance.

Is that enough?  Do you need counts for the sub 1 BTC bets too?

[Rannasha]

If dooglus can make a database-dump with all bets-by-size-and-chance again, we can compute what the expected standard deviation of the profits is and how (un)likely our current situation is. The previous time we had such a dump, the site was at -2 sigma, so the conclusion was that the house was very unlucky, but not unreasonably so. At that time, we had around half the amount wagered, so it's been a while.

I remember before I made several reports.  What would be most useful?

There are two variables: bet size and chance of winning.  Do you want a simple count of bets by stake, ignoring chance?  Or counts grouped by stake and chance?

Here's an example.  It is all bets of 1 BTC or more, grouped by 'bet' (stake) and 'chance'.

http://privatepaste.com/931a6b5a2a

'bet' is the stake in satoshis, and 'chance' is the percentage chance of winning times 10,000.

+----------+-------------+--------+
| count(*) | bet         | chance |
+----------+-------------+--------+
|        1 | 64000000000 | 900000 |
|        1 | 56129526541 | 980000 |
|        3 | 50000000000 | 900000 |

So there was 1 bet of 640 BTC at 90% chance,
1 bet of 561.29xxx BTC at 98% chance,
and 3 bets of 500 BTC at 90% chance.

Is that enough?  Do you need counts for the sub 1 BTC bets too?

The report I parsed previously is the first one listed in this post:
https://bitcointalk.org/index.php?topic=242962.msg2777073#msg2777073

It contains all the necessary data to compute what the expected profits are (obviously, one could take the shortcut and compute 1% of the amount wagered, but where's the fun in that?) as well as the standard deviation of the expected profits. With that, we can see how far off expectation we currently are.

I think I have the code that parses such a file on my laptop somewhere, so if I can find it back, I could run it without much effort on a new dataset.

[dooglus]

Is that enough?  Do you need counts for the sub 1 BTC bets too?

http://just-dice.com/grouped.txt.bz2 has the same information for all bets, down to 0 BTC staked.

I removed all the formatting, so it's just 3 numbers per line.

[mechs]

But why do many investors try to lower the risk by changing the rules instead of just lowering their invested amount?

Lowering the invested amount reduces risk and reward at the same time.  It doesn't change the risk/reward ratio.

Changing the max bet policy might reduce the risk a lot, while lowering the reward a little.  In other words, it might make the risk/reward ratio more favorable.

That is how I see it