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

[BlueBanderBear]

Again I have been accused of suspicious activity and IP address blocked- Dooglus, what gives?

What account number, or IP address, please?

81.159.39.121. Thanks, just been blocked again!

[TaxReturn]

Looks like celeste is at it again. Do I see some nervous fingers hovering over that divest button?

[Logik]

#whalewatch

[themagicdrake]

Ouch!

https://i.imgur.com/QQ71z52.png

[stripykitteh]

I've written a whale simulation program and have run some trials with it (it is only a quick and dirty perl script and I am not a great programmer but it seems to work).

Look at the comments in the code to see what it does. Basically a whale takes on the bank, betting max until one or the other goes broke.

It takes the following parameters:

• number of trials
• whale's bankroll
• house's bankroll
• maximum bet (as a % of the house's roll
  house edge (this is expressed as a number between 0 and 199, 100 represents a 1% house edge, 101 would be a 2% edge, etc)

It reports results like this:

Code:

bash-3.2$ ./whale_sim.pl 10 25000 25000 0.01 100
roll_num => 5294, bet => 387.43, jd_bal => 50000.00, whale_bal => 0.00, max => 50000.00, min => 18335.95, jw => 2695, ww => 2599
roll_num => 10742, bet => 3.18, jd_bal => 50000.00, whale_bal => 0.00, max => 50000.00, min => 14505.12, jw => 5433, ww => 5309
roll_num => 571, bet => 81.16, jd_bal => 50000.00, whale_bal => 0.00, max => 50000.00, min => 24007.70, jw => 322, ww => 249
roll_num => 4960, bet => 102.77, jd_bal => 50000.00, whale_bal => 0.00, max => 50000.00, min => 23369.29, jw => 2527, ww => 2433
roll_num => 2166, bet => 409.23, jd_bal => 50000.00, whale_bal => 0.00, max => 50000.00, min => 20695.56, jw => 1122, ww => 1044
roll_num => 29386, bet => 234.59, jd_bal => 50000.00, whale_bal => 0.00, max => 50000.00, min => 4441.13, jw => 14801, ww => 14585
roll_num => 8725, bet => 119.31, jd_bal => 50000.00, whale_bal => 0.00, max => 50000.00, min => 18439.80, jw => 4419, ww => 4306
roll_num => 14902, bet => 407.67, jd_bal => 50000.00, whale_bal => 0.00, max => 50000.00, min => 7332.98, jw => 7523, ww => 7379
roll_num => 21831, bet => 231.58, jd_bal => 50000.00, whale_bal => 0.00, max => 50000.00, min => 4008.85, jw => 11005, ww => 10826
roll_num => 619, bet => 200.83, jd_bal => 50000.00, whale_bal => 0.00, max => 50000.00, min => 24242.92, jw => 346, ww => 273

The legend is:

roll_num - the roll number when the simulation ended
bet - the bet size
jd_bal - Just-Dice's ending balance
whale_bal - Whale's ending balance
max - Just-Dice's highest balance during the simulation
min - Just-Dice's lowest balance during the simulation
jw - winning rolls for J-D
ww - winning rolls for the whale

I've done a lot of runs, and can summarize as follows:

- the house always wins in the end
- volatility is a bitch (when I gave the whale 4x the balance of the house the house was down 80% a few times, but always won eventually)

Code:

bash-3.2$ ./whale_sim.pl 10 100000 25000 0.01 100
roll_num => 15429, bet => 759.48, jd_bal => 125000.00, whale_bal => 0.00, max => 125000.00, min => 20707.98, jw => 7832, ww => 7597
roll_num => 9460, bet => 572.40, jd_bal => 125000.00, whale_bal => 0.00, max => 125000.00, min => 21912.81, jw => 4834, ww => 4626
roll_num => 18783, bet => 364.74, jd_bal => 125000.00, whale_bal => 0.00, max => 125000.00, min => 13347.83, jw => 9519, ww => 9264
roll_num => 29059, bet => 288.24, jd_bal => 125000.00, whale_bal => 0.00, max => 125000.00, min => 8580.01, jw => 14683, ww => 14376
roll_num => 8500, bet => 541.51, jd_bal => 125000.00, whale_bal => 0.00, max => 125000.00, min => 24750.00, jw => 4352, ww => 4148
roll_num => 13605, bet => 984.95, jd_bal => 125000.00, whale_bal => 0.00, max => 125000.00, min => 18119.94, jw => 6917, ww => 6688
roll_num => 4898, bet => 528.69, jd_bal => 125000.00, whale_bal => 0.00, max => 125000.00, min => 19163.66, jw => 2542, ww => 2356
roll_num => 17691, bet => 960.43, jd_bal => 125000.00, whale_bal => 0.00, max => 125000.00, min => 10519.66, jw => 8970, ww => 8721
roll_num => 17107, bet => 585.96, jd_bal => 125000.00, whale_bal => 0.00, max => 125000.00, min => 21547.66, jw => 8677, ww => 8430
roll_num => 32050, bet => 232.41, jd_bal => 125000.00, whale_bal => 0.00, max => 125000.00, min => 10047.04, jw => 16186, ww => 15864

See, with a 100,000BTC starting bankroll for the whale the house gets down to 8580 in trial 4. I've seen lower.

One last sim before my bedtime, this one has a 2% max bet and a 2% house edge:

Code:

bash-3.2$ ./whale_sim.pl 10 25000 25000 0.02 101
roll_num => 6831, bet => 349.12, jd_bal => 50000.00, whale_bal => 0.00, max => 50000.00, min => 5577.00, jw => 3467, ww => 3364
roll_num => 2518, bet => 817.76, jd_bal => 50000.00, whale_bal => 0.00, max => 50000.00, min => 14035.39, jw => 1289, ww => 1229
roll_num => 217, bet => 604.93, jd_bal => 50000.00, whale_bal => 0.00, max => 50000.00, min => 24890.22, jw => 127, ww => 90
roll_num => 6483, bet => 917.34, jd_bal => 50000.00, whale_bal => 0.00, max => 50000.00, min => 4204.63, jw => 3291, ww => 3192
roll_num => 179, bet => 826.50, jd_bal => 50000.00, whale_bal => 0.00, max => 50000.00, min => 24830.54, jw => 107, ww => 72
roll_num => 2920, bet => 837.70, jd_bal => 50000.00, whale_bal => 0.00, max => 50000.00, min => 14270.50, jw => 1492, ww => 1428
roll_num => 713, bet => 236.40, jd_bal => 50000.00, whale_bal => 0.00, max => 50000.00, min => 20488.12, jw => 377, ww => 336
roll_num => 805, bet => 688.58, jd_bal => 50000.00, whale_bal => 0.00, max => 50000.00, min => 19237.02, jw => 424, ww => 381
roll_num => 315, bet => 58.77, jd_bal => 50000.00, whale_bal => 0.00, max => 50000.00, min => 22507.79, jw => 176, ww => 139
roll_num => 4534, bet => 501.58, jd_bal => 50000.00, whale_bal => 0.00, max => 50000.00, min => 12465.50, jw => 2307, ww => 2227

In trial 4 the house goes down to 4204BTC!


Here's the source code for reference:

Code:

! /usr/bin/perl                                                                                                                                                     

#                                                                                                                                                                    
# Simulation of a battle between a whale player and Just-Dice                                                                                                        
#                                                                                                                                                                    
# Purpose                                                                                                                                                            
# -------                                                                                                                                                            
# Theory is that the Kelly Criterion should keep the bank safe from ruin                                                                                             
# (though not from volatility) and eventually always cause the house to                                                                                              
# win the whale's roll.                                                                                                                                              
#                                                                                                                                                                    
# I am interested to see                                                                                                                                             
# - if this is true in practice                                                                                                                                      
# - how long might it take                                                                                                                                           
# - how far the house might fall behind before it happens                                                                                                            
#                                                                                                                                                                    
# Assumptions                                                                                                                                                        
# -----------                                                                                                                                                        
# - Whale always bets the maximum.                                                                                                                                   
# - If Just-Dice's roll drops below 250btc (1% of starting roll), it is all-in                                                                                       
#   (just to give the Whale a chance of winning :-) )                                                                                                                
# - Whale keeps betting until one or the other goes broke.                                                                                                           
#                                                                                                                                                                    
# Purpose                                                                                                                                                            
# -------                                                                                                                                                            
# Theory is that the Kelly Criterion should keep the bank safe from ruin                                                                                             
# (though not from volatility) and eventually always cause the house to                                                                                              
# win the whale's roll.                                                                                                                                              
#                                                                                                                                                                    
# I am interested to see                                                                                                                                             
# - if this is true in practice                                                                                                                                      
# - how long might it take                                                                                                                                           
# - how far the house might fall behind before it happens                                                                                                            
#                                                                                                                                                                    

$num_trials = $ARGV[0];

for ($i=0; $i<$num_trials; $i++) {
    $whale_bal = $ARGV[1];
    $jd_bal = $ARGV[2];
    $max_bet = $ARGV[3];

# bets in the range 0 -> $house_edge -> 199 win                                                                                                                      
    $house_edge = $ARGV[4];
    $report_freq = 100000;

# Keep track of the highest and lowest house balances seen                                                                                                           
    $jd_max = $jd_bal;
    $jd_min = $jd_bal;
    $jd_orig = $jd_bal;

    while ( ($whale_bal > 0) && ($jd_bal > 0) ) {

        $roll_num++;

        if ( $jd_bal < ($jd_orig * $max_bet) ) {
            $bet = $jd_bal;
        } elsif ( $whale_bal < ($jd_bal * $max_bet) ) {
            $bet = $whale_bal;
        } else {
            $bet = $jd_bal * $max_bet;
        }

        $roll = int(rand(200));

        if ( $roll > $house_edge ) {
            $whale_bal = $whale_bal + $bet;
            $jd_bal = $jd_bal - $bet;
            $whale_wins++;
        } else {
            $whale_bal = $whale_bal - $bet;
            $jd_bal = $jd_bal + $bet;
            $jd_wins++;
        }

        if ( $jd_bal > $jd_max ) {
            $jd_max = $jd_bal;
        } elsif ( $jd_bal < $jd_min ) {
            $jd_min = $jd_bal;
        }

        if ( 0 == ($roll_num % $report_freq) ) {
            printf ("roll_num => $roll_num, bet => %.2f, jd_bal => %.2f, whale_bal => %.2f, max => %.2f, min => %.2f, jw => $jd_wins, ww => $whale_wins\n", $bet, $j\
d_bal, $whale_bal, $jd_max, $jd_min);
        }
    }

    printf ("roll_num => $roll_num, bet => %.2f, jd_bal => %.2f, whale_bal => %.2f, max => %.2f, min => %.2f, jw => $jd_wins, ww => $whale_wins\n", $bet, $jd_bal, $\
whale_bal, $jd_max, $jd_min);

}

exit 0;

[romerun]

if the whale ever gets 50% of the house, I guess most investors are likely to devest and the house is defeated shortly after.

[infested999]

if the whale ever gets 50% of the house, I guess most investors are likely to devest and the house is defeated shortly after.

If the profit goes to -15,000 BTC?

[petrescuerz]

Thanks to Emily Spaven at CoinDesk for a great article on Just-Dice!

http://www.coindesk.com/bitcoin-gambler-cheats-satoshidice-competitor-just-dice-out-of-1300-btc/

And in other news, the whale is back!

[RationalSpeculator]

I've done a lot of runs, and can summarize as follows:

- the house always wins in the end
- volatility is a bitch (when I gave the whale 4x the balance of the house the house was down 80% a few times, but always won eventually)

Wauw, thank you so much for doing such extensive testing stripykitteh. That's very valuable information for any investor.

This means that as an investor you better be prepared for some serious losses, because if you are not you will bail out, and never make it back.

Turns out dooglus his decision to divest is not an emotional one but indeed a rational one. If you are risk averse this is not the place to put your money.

[dooglus]

Turns out dooglus his decision to divest is not an emotional one but indeed a rational one. If you are risk averse this is not the place to put your money.

I left my coins invested overnight accidentally.  'celeste' had promised me he had retired from playing, and I believed him, so I wasn't overly concerned.

When I woke up, he had already lost one deposit of 200 BTC and was half way through losing his 2nd 200 BTC.  I divested full as soon as I saw what was happening, and stayed divested until he lost the 2nd 200 BTC as well.

I reinvested, but then he started playing again, so I pulled out again until he finally lost his 3rd deposit of 600 BTC.

So I missed most of the upside by being too scared to roll with his enormous variance.

Remember, however, that I (hopefully, eventually) get to take commission on any new profits the site makes - so investing isn't my only source of income from the site.  That probably makes it easier for me to divest during dangerous levels of play than for other investors.

Edit: even after divesting, I still couldn't bare to watch.  I went out for a ride around the area, then came back and brushed my dog's coat.  She's gotten all matted where I've neglected her over the last couple of months.  All my time has been spent working on Just-Dice, to the detriment of everything else...  Watching the whale destroy the site profits has been very upsetting to me, and I find it hard to remember "variance is a bitch".  I need to keep telling myself that.

[Lohoris]

When I woke up, he had already lost one deposit of 200 BTC and was half way through losing his 2nd 200 BTC.  I divested full as soon as I saw what was happening, and stayed divested until he lost the 2nd 200 BTC as well.

I reinvested, but then he started playing again, so I pulled out again until he finally lost his 3rd deposit of 600 BTC.

I'm sorry, but you doing such a thing kind of demonstrates that the design strongly needs to be adjusted.

Sure, this will still be a risky investment, but it's no reason to let it be "as it is" instead of improving it.

Increasing the house rake for the bets with high payouts is really a must and can't be avoided anymore, IMAO.

[dooglus]

I'm sorry, but you doing such a thing kind of demonstrates that the design strongly needs to be adjusted.

I don't think it demonstrates any such thing.

What it demonstrates is that I'm a pussy...

I'm happy to let braver souls take on the whale, and plenty seem prepared to do so.

[RationalSpeculator]

All the people saying "you don't understand variance" seem to, themselves, not understand it; the maths depends on 'perfect conditions' - in this case that there are many whales and that they bet constantly and consistently. That isn't the case. We need a few dozen (or a few hundred) more whales and then you could trust the maths.

Some saying that may not understand it.  But some saying it (such as myself) DO understand variance.

You counter variance with volume.  One way is to have lots of whales.  Another is to have lots of different investments.  If someone has a bunch of different investments all with similar risk profiles to J-D then they don't need any individual investment to have low variance itself.  Similarly if someone manages risk across their portfolio so that J-D is part of their high-variance portion then they also aren't too concerned.

The high variance of J-D only really matters to investors who are trying to use it as something it isn't - or who have too large a portion of their investment capital in it.  For the rest of us the main concern is maximising EV - then we can manage our investments ourself to deal with variance.

There's only two scenarios in which it makes sense to reduce the max bet % of J-D :

1.  If dooglus believes that doing so would attract so much more investment that the actual max bet would increase, allowing even bigger whales and so more total expected house profit (and thus more expected commission).
2.  If dooglus needs J-D to be profitable for himself in the short-term - in which case he should probably drop max bet to 5 BTC, remove the investment option (as he could bank-roll that himself) and settle for being just another small-stakes dice site.

For investors who are interested in maximising EV any reduction of max bet % is bad news.  Not only is it sub-optimal from a kelly perspective (and so even more inefficient use of capital than the current inefficient system - which is a seperate topic) but reducing max bet has a double whammy impact on likely  bet volume:

1.  Reducing the attraction of the site to whales.
2.  Reducing the attraction of the site to smaller bettors who only come because they can watch the whales play - and because in theory they can keep doubling up to a large win.

It also has a double-whammy impact on EV for investors:

1.  Loss of volume,
2.  Likely increase in investment diluting the already reduced volume.

Anything which reduces EV is bad for all sensible investors.  Variance is something you manage across your investment portfolio - not something every investment needs to handle itself.  Precisely how you do that depends on what invesments are available to you and what profile you want your investments as a whole to match.  Now I appreciate that can be hard to do - as there's a lack of investments available for the low'risk of your portfolio.  But asking for high returns AND low variance is just unrealistic.

Anyone not invested who wants max bet reduced so they can feel safe investing is totally missing the point.  Your investment is NOT needed if the condition for it is that the site makes itself less attractive to players than it already is.  What is so important about YOU wanting to invest that players should get reduced options and existing investors have their EV destroyed just to make YOU feel happy?

If dooglus wanted to run a low max-bet site than he has the funds to back it himself without investors at all : be careful what you ask for in case you actually get it.  It's debatable whether the current investment model is sensible at all - for any significantly lower max bet it almost certainly isn't.

What a great post Deprived. Thank you for sharing.

I agree that if you decide to invest in just-dice you want it to be only a small part of your portfolio, and have different other investments also. I also think it is wise to keep it stable. So say you invest 10% of your capital in it, then you keep it at 10%. This means that if a whale hits the house hard to say -80%, you lose only 8% of your capital.

And since you want to keep it at 10% you will invest more into just-dice as it goes down. This way you will do the inverse of what most do as they divest out of fear seeing all the losses, giving you a much larger percentage of the house, right at the time when the turnover is going up and up as more gamblers want to join the party seeing all the winnings.

Then when the inevitable turn comes and the whale loses it all you make a killing, and since you have more then 10% of your capital in just-dice you now divest a lot, while others seeing the gains mostly invest. This way you lose most of the percentage in the house, but there is not much money to be made anyway as the whale just got wiped out and the turnover of bets will be much lower.  

[Lohoris]

I'm sorry, but you doing such a thing kind of demonstrates that the design strongly needs to be adjusted.

I don't think it demonstrates any such thing.

What it demonstrates is that I'm a pussy...

I'm happy to let braver souls take on the whale, and plenty seem prepared to do so.

Again, this is no reason to believe that "as it is now" is "just holy perfect" and that it can't be improved.

[RationalSpeculator]

When I woke up, he had already lost one deposit of 200 BTC and was half way through losing his 2nd 200 BTC.  I divested full as soon as I saw what was happening, and stayed divested until he lost the 2nd 200 BTC as well.

I reinvested, but then he started playing again, so I pulled out again until he finally lost his 3rd deposit of 600 BTC.

I'm sorry, but you doing such a thing kind of demonstrates that the design strongly needs to be adjusted.

Sure, this will still be a risky investment, but it's no reason to let it be "as it is" instead of improving it.

Increasing the house rake for the bets with high payouts is really a must and can't be avoided anymore, IMAO.

As deprived explained very well that is not the solution Lohoris. Please debunk his arguments instead of ignoring them and repeating your poor solution. 

The solution for dooglus and all the others that find the whale too risky is to invest less of your capital in just-dice so that you are not feeling bad when the worst happens and the whale takes -80% of your investment.

I am very risk averse. That's why I keep my exposure low to any single investment, bitcoins included!

[Lohoris]

As deprived explained very well that is not the solution Lohoris. Please debunk his arguments instead of ignoring them and repeating your poor solution. 

The solution for dooglus and all the others that find the whale too risky is to invest less of your capital in just-dice so that you are not feeling bad when the worst happens and the whale takes -80% of your investment.

I am very risk averse. That's why I keep my exposure low to any single investment, bitcoins included!

Deprived was answering to people that were asking for increasing the house rake for all bets. I agree with him.

Later someone posted a list of competitors, noticing that most of them have quite low max bets, hence I proposed to increase the rake on bets bigger than that.
Doing so would not hurt small bets (it would still be 1%) and likely it wouldn't hurt large bets either, since gamblers basically have nowhere else to go.

Try to debunk this.

[Deprived]

I'm sorry, but you doing such a thing kind of demonstrates that the design strongly needs to be adjusted.

I don't think it demonstrates any such thing.

What it demonstrates is that I'm a pussy...

I'm happy to let braver souls take on the whale, and plenty seem prepared to do so.

Again, this is no reason to believe that "as it is now" is "just holy perfect" and that it can't be improved.

Agreed.

If raising the edge on higher bets would increase EV then I'm all for it.  Question is whether it would.  If EV remains the same but variance reduces that's also good.

An approximation to the math is pretty simple.

If at 1% edge X BTC are wagered/week in the higher bracket then an increase of edge to (for example) 1.5% only makes sense if the BTC wagered in the higher bracket would drop to no lower that .667X (giving the same EV with lower variance). It's not QUITE that simple as there's other factors but as an approximation it isn't bad.  Here's some of the factors against it:

1.  It makes it more complicated for bettors - and potentially puts off some using betting strategies which start low but have the potential to end up in the higher band (people running martingles aren't goign to like the idea that if they lose a lot then sudenly the house makes the edge worse whilst they're trying to get back what they just lost).
2.  Whales draw attention which leads to other trade.  ANY reduction in whale action is likley to be accompanied by a drop (or reduction in growth) of smaller bets.
3.  Raising the edge makes life easier for existing and new competitors - which is an unnecessary risk when there's no problem raising the capital necessary at current edge.

But make a convincing case why with a 1.5% edge we'd still get at least 2/3 the volume of large bets (or whatever ratio applies for whichever edge you propose) and it's worth considering.  Personally I think even then the other factors make it unattractive - and the moment some other site offered 1% edge with a max bet above the higher-rate limit the change would likely have to be reversed anyway.

Whilst changes shouldn't be ruled out just because they're changes (or because the current system is assumed to be perfect) it remains the case that changes that are obviously detrimental to bettors are NOT going to attract more bettors.  And there's no need to worry about chnages being beneficial to investors - as, unlike bettors, there's no massive incentive to attract more investors.  Unless, that is, you believe max bet should be a lot higher and the way to raise it is to heavily increase investment - which I disagree with (not because I'm against a higher max-bet but because the price of it is diluted EV and higher variance for existing investors).

In short, changes shouldn't be rejected out of hand - but it's the responsibility of whoever suggests them to demonstrate the change is likely to result in increased EV.   That's because any change which is negative to bettors is almost certain to reduce betting volume (unless very cunningly disguised).

[RationalSpeculator]

even after divesting, I still couldn't bare to watch.  I went out for a ride around the area, then came back and brushed my dog's coat.  She's gotten all matted where I've neglected her over the last couple of months.  All my time has been spent working on Just-Dice, to the detriment of everything else...  Watching the whale destroy the site profits has been very upsetting to me, and I find it hard to remember "variance is a bitch".  I need to keep telling myself that.

I'm sorry to hear you have been neglecting your dog and everything else

I also work a lot, enjoy it very much, but at the expense of the other parts of my life. Time is ticking though and I worry often about my life choices.

I'm in therapy and it's very helpful. Though I still have a lot of self-work to do I feel that I am becoming happier and healthier slowly.

I don't think it demonstrates any such thing.

What it demonstrates is that I'm a pussy...

I'm happy to let braver souls take on the whale, and plenty seem prepared to do so.

I'm sorry to hear you think you are a pussy.

I don't think that is true. It's completely rational what you do, though somewhat impractical to continuously invest/divest. I can imagine this puts extra stress on you.

What I am learning in therapy is to treat myself better. I also had parts that called me names like 'you are such a loser'. That's hard

[Lohoris]

[...]
In short, changes shouldn't be rejected out of hand - but it's the responsibility of whoever suggests them to demonstrate the change is likely to result in increased EV.   That's because any change which is negative to bettors is almost certain to reduce betting volume (unless very cunningly disguised).

Good points.

My theory is

assuming that:
1. the only competitor accepting large bets is satoshi dice
2.1. satoshi dice is experiencing turmoils, and getting out of flavour for various reasons
2.2. anyway, their rake is 1.9%, so if we raised to 1.6% we would still be well below them
3. whales will likely want to play somewhere

the consequence would be:
A. whales have nowhere else to go, and will keep playing here even if we raise the rake a bit

I think the only "dubious" point of this theory is (3), but I don't think it's weak enough not to be worth trying anyway.

So, if someone fears that increasing the rake would make the whales choose not to play anymore, fine.
Other than that, I see no other reason to refute this theory.

(that said, I think I won't press the issue any further)

[RationalSpeculator]

As deprived explained very well that is not the solution Lohoris. Please debunk his arguments instead of ignoring them and repeating your poor solution.  

The solution for dooglus and all the others that find the whale too risky is to invest less of your capital in just-dice so that you are not feeling bad when the worst happens and the whale takes -80% of your investment.

I am very risk averse. That's why I keep my exposure low to any single investment, bitcoins included!

Deprived was answering to people that were asking for increasing the house rake for all bets. I agree with him.

Later someone posted a list of competitors, noticing that most of them have quite low max bets, hence I proposed to increase the rake on bets bigger than that.
Doing so would not hurt small bets (it would still be 1%) and likely it wouldn't hurt large bets either, since gamblers basically have nowhere else to go.

Try to debunk this.

Makes sense Lohoris. I can't debunk your argument.

I'm sorry, I made false conclusions about you.

I should have read the posts better before responding.