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

[chriswen]

I don't see why there should be an investment delay.  There's nothing wrong with using a daytrading strategy.  And if you actually tried to day trade there is actually a delay when investing and divesting.

First when you click the edit button there is a delay.  Then you have to input a value and click okay and there's a delay for that to.

If you were discussing a delay, how long would it be?

[mechs]

The site so far has had 145,551,114 bets and cumulative luck for all players is at 100.49%.  This seems like a large enough sample size that it should be converging on 100%.  With a 1% edge assuming the most common bet is a 2x 49.5% bet, it would seem a player luck of 100.49% essentially means a historical house edge of 0.51%.  Is that logic correct?  If so, is this a probable result with a 1% house edge?

If I understand things correctly, the luck percentage has the house edge built in already.

The luck percentage displayed shows how many rolls you have won compared to how many you 'should' have won.

The luck percentage should, over time, be very close to 100%.

Really?  I think luck is supposed to approach around 101%.  Dooglus made a comment about it being above 101% before.

I thought luck is deviation from expected results.  So with enough wagers, it invariably trend towards 100%  It does not take into account bet sizes, only outcomes so it should have minimal effect from variance.  It is a good test of how well the RNG is working.
I feel like after 146 million wagers, a player luck of 100.49% seems high. Not sure how one would go about testing that.

[nrd525]

If you were flipping a coin a hundred million times, I'm guessing your luck (with two standard deviations) would be around 99.98% to 100.02%  (or would it be 99.96 to 100.04?)

The approximation formula for this (if I remember stats 101) is: std deviation is
1/square root of (n)  
1/square root (100,000,000)  = 1/10000 or 0.01%

In this case, for it to deviate to 100.4% would be extremely improbable (40 or 20? standard deviations).

Now it's more complex because the probability of the bet ranges from 0 to 1.  But still it seems unreasonable.  Maybe someone with a better explanation of the luck number can explain?

[elm]

"Fair" means they don't cheat the player.  "Provable" means they can prove it. so the players are on the secure side.

how about the investors?

It would be trivial for a traditional casino to say "we had a high roller come and win a million dollars from us.  Sorry, no dividends this month".  How would the investors/shareholders argue with that?

I can imagine that the investors wouldnt like it and would argue heavily

[Jls.]

The site so far has had 145,551,114 bets and cumulative luck for all players is at 100.49%.  This seems like a large enough sample size that it should be converging on 100%.  With a 1% edge assuming the most common bet is a 2x 49.5% bet, it would seem a player luck of 100.49% essentially means a historical house edge of 0.51%.  Is that logic correct?  If so, is this a probable result with a 1% house edge?

If I understand things correctly, the luck percentage has the house edge built in already.

The luck percentage displayed shows how many rolls you have won compared to how many you 'should' have won.

The luck percentage should, over time, be very close to 100%.

Really?  I think luck is supposed to approach around 101%.  Dooglus made a comment about it being above 101% before.

I thought luck is deviation from expected results.  So with enough wagers, it invariably trend towards 100%  It is not take into account bet sizes, only outcomes so it should have minimal effect from variance.  It is a good test of how well the RNG is working.
I feel like after 146 million wagers, a player luck of 100.49% seems high. Not sure how one would go about testing that.

If I remember well, doog explained once that the luck % had been boosted by numerous successful very low odds bets.

Variance. Variance everywhere !

[nrd525]

FAQ should include
-why is luck <> 100%
-why is the profit/loss deviating so much from 1%?

[mechs]

The site so far has had 145,551,114 bets and cumulative luck for all players is at 100.49%.  This seems like a large enough sample size that it should be converging on 100%.  With a 1% edge assuming the most common bet is a 2x 49.5% bet, it would seem a player luck of 100.49% essentially means a historical house edge of 0.51%.  Is that logic correct?  If so, is this a probable result with a 1% house edge?

If I understand things correctly, the luck percentage has the house edge built in already.

The luck percentage displayed shows how many rolls you have won compared to how many you 'should' have won.

The luck percentage should, over time, be very close to 100%.

Really?  I think luck is supposed to approach around 101%.  Dooglus made a comment about it being above 101% before.

I thought luck is deviation from expected results.  So with enough wagers, it invariably trend towards 100%  It is not take into account bet sizes, only outcomes so it should have minimal effect from variance.  It is a good test of how well the RNG is working.
I feel like after 146 million wagers, a player luck of 100.49% seems high. Not sure how one would go about testing that.

If I remember well, doog explained once that the luck % had been boosted by numerous successful very low odds bets.

Variance. Variance everywhere !

That makes sense with 1 million bets, not with nearly 150 million

[nicolaennio]

More importantly however, whatever you think the reason for this is, please look at nakowa's yesterday winning charts and tell me that there is no "exploitable" pattern to it:

That's all easy with hindsight. The hard part is being sure of the exploitable pattern of Nakowa's next run.

Similarly, I've noticed BTC trading in the 130-145 range over the last weeks. What's stopping you from selling at say 142 and buying back at 135 repeatedly? Would have been a winning strategy over the last weeks/months. What could possibly go wrong...

I'm addressing that point in my post: I said, at around bet 6000, some people decided it's a "pattern" and traded accordingly.

You also noted that I said, several times in fact, that I'm aware that in the long run, investment "trading" should be -EV, right?

But that wasn't my argument anyway: in the short run, it can be +EV. So a loss like the one we suffer now isn't distributed evenly over investors. So the losing investors will be unhappy. So the site will lose investors, which is bad for everyone.

That, in a nutshell, is my argument why invest-trading hurts j-d. And the complementary argument is: there's no good reason to *want* invest-trading: j-d already has a gambling mechanism. Investments should be just that: investments.

The discussion about investing/divesting is getting long and I must tell that I am convinced that people in the chat were actually cheating people by asking them to divest when Nakowa was -4K for at least two reasons. First, if you have the "secret" strategy you are not going to tell anyone, secondly as everyone agrees divesting is another fallacy.

Now also the forum is getting plenty of this and it looks to me an army of trolls trying to cheat other people with "secret" tricks.

Maybe I should stay silent and stay invested, but this is getting too much spam in my opinion.

[chriswen]

More importantly however, whatever you think the reason for this is, please look at nakowa's yesterday winning charts and tell me that there is no "exploitable" pattern to it:

That's all easy with hindsight. The hard part is being sure of the exploitable pattern of Nakowa's next run.

Similarly, I've noticed BTC trading in the 130-145 range over the last weeks. What's stopping you from selling at say 142 and buying back at 135 repeatedly? Would have been a winning strategy over the last weeks/months. What could possibly go wrong...

I'm addressing that point in my post: I said, at around bet 6000, some people decided it's a "pattern" and traded accordingly.

You also noted that I said, several times in fact, that I'm aware that in the long run, investment "trading" should be -EV, right?

But that wasn't my argument anyway: in the short run, it can be +EV. So a loss like the one we suffer now isn't distributed evenly over investors. So the losing investors will be unhappy. So the site will lose investors, which is bad for everyone.

That, in a nutshell, is my argument why invest-trading hurts j-d. And the complementary argument is: there's no good reason to *want* invest-trading: j-d already has a gambling mechanism. Investments should be just that: investments.

The discussion about investing/divesting is getting long and I must tell that I am convinced that people in the chat were actually cheating people by asking them to divest when Nakowa was -4K for at least two reasons. First, if you have the "secret" strategy you are not going to tell anyone, secondly as everyone agrees divesting is another fallacy.

Now also the forum is getting plenty of this and it looks to me an army of trolls trying to cheat other people with "secret" tricks.

Maybe I should stay silent and stay invested, but this is getting too much spam in my opinion.

Divesting is not a fallacy.  It is a risk management strategy.  You minimize losses but you may miss out on gains.

Have you ever heard the term take profit in investing?  This is what you're doing with the divesting strategy.

You take profits that you are content with.  You might miss out on further profits.  But if the profit goes down, you'll have limited your losses.

[integrity42]

On first view it seems that 'day trading' nakowa has a zero expectation.  But if he keeps going up and down like today then "divest at +1000, invest at -1000" is a profitable strategy.

Invest at - 5k, divest at +5k like yesterday.
You will make 20% profit daily

Or you'll get stuck divested at 5k and missing out on profits from 5k to 50k...  (we can dream...)

I've made a change to the roll algorithm.

From bet 145,000,000 it will be doing:
  hmac_sha512(nonce+':'+server_seed+':'+nonce, nonce+':'+client_seed+':'+nonce)
instead of the previous:
  hmac_sha512(          server_seed          ,           client_seed+':'+nonce)

ie. using the nonce in 4 places instead of 1.  sha512 is meant to completely change its output for any single bit change in its inputs, but it can't do any harm to change more than one bit each time.

Edit: it was pointed out in chat that this change is pointless, since even a small change in the message will (with overwhelming probability) result in a mostly different hash, due to the avalanche effect.

Aside from this change being absolutely pointless,  you risk nakowa having something else to complain about.

*Any* change you make to the site in any way, Nakowa will complain if he loses that you're cheating somehow.  So if he plays a new session with the new seed algo, and loses, he'll complain in some incoherent anti-mathematically fashion.  IF you change the kelly number, its also cheating....

[Rampion]

More importantly however, whatever you think the reason for this is, please look at nakowa's yesterday winning charts and tell me that there is no "exploitable" pattern to it:

That's all easy with hindsight. The hard part is being sure of the exploitable pattern of Nakowa's next run.

Similarly, I've noticed BTC trading in the 130-145 range over the last weeks. What's stopping you from selling at say 142 and buying back at 135 repeatedly? Would have been a winning strategy over the last weeks/months. What could possibly go wrong...

I'm addressing that point in my post: I said, at around bet 6000, some people decided it's a "pattern" and traded accordingly.

You also noted that I said, several times in fact, that I'm aware that in the long run, investment "trading" should be -EV, right?

But that wasn't my argument anyway: in the short run, it can be +EV. So a loss like the one we suffer now isn't distributed evenly over investors. So the losing investors will be unhappy. So the site will lose investors, which is bad for everyone.

That, in a nutshell, is my argument why invest-trading hurts j-d. And the complementary argument is: there's no good reason to *want* invest-trading: j-d already has a gambling mechanism. Investments should be just that: investments.

The discussion about investing/divesting is getting long and I must tell that I am convinced that people in the chat were actually cheating people by asking them to divest when Nakowa was -4K for at least two reasons. First, if you have the "secret" strategy you are not going to tell anyone, secondly as everyone agrees divesting is another fallacy.

Now also the forum is getting plenty of this and it looks to me an army of trolls trying to cheat other people with "secret" tricks.

Maybe I should stay silent and stay invested, but this is getting too much spam in my opinion.

Divesting is not a fallacy.  It is a risk management strategy.  You minimize losses but you may miss out on gains.

Have you ever heard the term take profit in investing?  This is what you're doing with the divesting strategy.

You take profits that you are content with.  You might miss out on further profits.  But if the profit goes down, you'll have limited your losses.

I agree on this (divesting to secure profits is not gambling), in fact I should have protected the profits I had when I went to sleep yesterday (I was up 5%), but greed and faith in both math and Nak's compulsive gambling behavior made me keep my investment there... And I woke up with a 12% loss.

This does not change the fact that trying to "ride the waves" or "predict the patterns" by investing/divesting multiple times while Nak is playing is just another form of gambler's fallacy. There are no predictable waves, no patterns, just random events. When the site is +3k coins, the chance of it going to -3k or 6k is exactly the same.

[eltopo]

Not exactly, the chance is higher we're getting to +6k because of the house edge.

[nicolaennio]

More importantly however, whatever you think the reason for this is, please look at nakowa's yesterday winning charts and tell me that there is no "exploitable" pattern to it:

That's all easy with hindsight. The hard part is being sure of the exploitable pattern of Nakowa's next run.

Similarly, I've noticed BTC trading in the 130-145 range over the last weeks. What's stopping you from selling at say 142 and buying back at 135 repeatedly? Would have been a winning strategy over the last weeks/months. What could possibly go wrong...

I'm addressing that point in my post: I said, at around bet 6000, some people decided it's a "pattern" and traded accordingly.

You also noted that I said, several times in fact, that I'm aware that in the long run, investment "trading" should be -EV, right?

But that wasn't my argument anyway: in the short run, it can be +EV. So a loss like the one we suffer now isn't distributed evenly over investors. So the losing investors will be unhappy. So the site will lose investors, which is bad for everyone.

That, in a nutshell, is my argument why invest-trading hurts j-d. And the complementary argument is: there's no good reason to *want* invest-trading: j-d already has a gambling mechanism. Investments should be just that: investments.

The discussion about investing/divesting is getting long and I must tell that I am convinced that people in the chat were actually cheating people by asking them to divest when Nakowa was -4K for at least two reasons. First, if you have the "secret" strategy you are not going to tell anyone, secondly as everyone agrees divesting is another fallacy.

Now also the forum is getting plenty of this and it looks to me an army of trolls trying to cheat other people with "secret" tricks.

Maybe I should stay silent and stay invested, but this is getting too much spam in my opinion.

Divesting is not a fallacy.  It is a risk management strategy.  You minimize losses but you may miss out on gains.

Have you ever heard the term take profit in investing?  This is what you're doing with the divesting strategy.

You take profits that you are content with.  You might miss out on further profits.  But if the profit goes down, you'll have limited your losses.

I do not get this. I am referring to "daytrading" divesting/investing. I agree with you that at some point, if you have earned bitcoins, you can quit just-dice and take the profit and never invest them again on the website. In all other cases you will reinvest the money, putting yourself back into the situation on the possibility of experiencing losses. Or maybe you wait the whale will finish to play for reinvest? But then why not adopting a smaller Kelly (like 0.125%) which would rule out the whale and really secure your profit?

[Draino]

can i just point out: when the max profit was less than 100BTC site profit consistently went up over a period of days, and now that it's over 100BTC site profit has swung wildly is now ultimately thousands of coins down over a period of hours (again)

i'm sure you probably-math people will have something that sounds alot like furrrmp durrrrmp vrrrrance to say but uh, results are real

[Oleander]

This does not change the fact that trying to "ride the waves" or "predict the patterns" by investing/divesting multiple times while Nak is playing is just another form of gambler's fallacy. There are no predictable waves, no patterns, just random events.

If this is true, then investing / divesting does no harm (other than the opportunity cost of not being invested part of the time).

On the other hand, investing / divesting seems to work very well.

If I'm not 100% convinced of the Gambler's Fallacy explanation, then it is reasonable to invest / divest.  At best, I come out ahead.  At worst, I come out about even.

[chriswen]

More importantly however, whatever you think the reason for this is, please look at nakowa's yesterday winning charts and tell me that there is no "exploitable" pattern to it:

That's all easy with hindsight. The hard part is being sure of the exploitable pattern of Nakowa's next run.

Similarly, I've noticed BTC trading in the 130-145 range over the last weeks. What's stopping you from selling at say 142 and buying back at 135 repeatedly? Would have been a winning strategy over the last weeks/months. What could possibly go wrong...

I'm addressing that point in my post: I said, at around bet 6000, some people decided it's a "pattern" and traded accordingly.

You also noted that I said, several times in fact, that I'm aware that in the long run, investment "trading" should be -EV, right?

But that wasn't my argument anyway: in the short run, it can be +EV. So a loss like the one we suffer now isn't distributed evenly over investors. So the losing investors will be unhappy. So the site will lose investors, which is bad for everyone.

That, in a nutshell, is my argument why invest-trading hurts j-d. And the complementary argument is: there's no good reason to *want* invest-trading: j-d already has a gambling mechanism. Investments should be just that: investments.

The discussion about investing/divesting is getting long and I must tell that I am convinced that people in the chat were actually cheating people by asking them to divest when Nakowa was -4K for at least two reasons. First, if you have the "secret" strategy you are not going to tell anyone, secondly as everyone agrees divesting is another fallacy.

Now also the forum is getting plenty of this and it looks to me an army of trolls trying to cheat other people with "secret" tricks.

Maybe I should stay silent and stay invested, but this is getting too much spam in my opinion.

Divesting is not a fallacy.  It is a risk management strategy.  You minimize losses but you may miss out on gains.

Have you ever heard the term take profit in investing?  This is what you're doing with the divesting strategy.

You take profits that you are content with.  You might miss out on further profits.  But if the profit goes down, you'll have limited your losses.

I do not get this. I am referring to "daytrading" divesting/investing. I agree with you that at some point, if you have earned bitcoins, you can quit just-dice and take the profit and never invest them again on the website. In all other cases you will reinvest the money, putting yourself back into the situation on the possibility of experiencing losses. Or maybe you wait the whale will finish to play for reinvest? But then why not adopting a smaller Kelly (like 0.125%) which would rule out the whale and really secure your profit?

When you reinvest you have the possibility of losing more this is true.  But if you didn't divest in the first place your loss would be up to the point you invested and then if it went down you would have lost even more.  So, of course you could lose more but you already minimized losses.

[GOB]

@pascal257

re: withdrawal delay. Maybe not a withdrawal delay per se, because it could make the *gamblers* nervous, but an 'investment from new deposit' delay should have the same effect, and only affect investors

@rampion

re: house edge 1% vs 1.5%. Thanks for your support. I said I'd rather wait a bit with raising house edge. Let's be honest: if nakowa moves away completely now, we're in a deep minus that'll take a long time to reverse (though I hope nakowa's as compulsive as we think he is). So the edge might not affect him anyway.

On the other hand, if we wait another week, let's see what happens. If site profit keeps going down (to new lows), I agree, raise the house edge. But if it stabilizes or goes up, we won't have chased customers away to the competition.

But, yes, I ultimately agree with you: a site that is not profitbale for investors in a very very long time (say, many months), will ultimately become uninteresting.

"Chased customers away to the competition"

Don't forget that Nakowa is the principal investor in Letsdice, that he has stated that he does this crazy betting here at least in part to promote letsdice, and that he seems to be on a mission to convince doog to change the Edge to 2% (ever since the beginning of just-dice). So if you increase the edge to 1.5 or 2%, Nakowa WOULD win, as that is his stated goal. And once he accomplishes his goal of making his main competitor remove their defining characteristic, their comparative advantage, who's to say he won't just be done?

[GOB]

So a loss like the one we suffer now isn't distributed evenly over investors. So the losing investors will be unhappy. So the site will lose investors, which is bad for everyone.

That, in a nutshell, is my argument why invest-trading hurts j-d. And the complementary argument is: there's no good reason to *want* invest-trading: j-d already has a gambling mechanism. Investments should be just that: investments.

Had Nak's run gone the other way, they would have been the winning investors, and they would have been quite happy to take a larger share of the gains I am sure. Nobody would now be calling for any tinkering of the system either.

There's no need to do anything. Investing in JD is a gamble anyway and in the long term daytrading will likely make no difference to those who stay invested. Besides we've only heard from the winners because they feel they've been smart, just like Nak feels he's a smart gambler whatever he calls himself.

My feeling is that some people who took a hit are pissed off that other people managed to make some money. While this is understandable, tinkering with the system just to appease them, with measures that probably don't work anyway and make no difference overall are just a waste of time.

There's even a chance that such tinkering is mildly detrimental for the site. What if a gambler decides to invest while he doesn't play? Say he wants to make a few rolls each day and always stops after 15 minutes, hoping to have made a profit, then reinvests. Why would you want to prevent that and either dissuade him from gambling or from investing or from keeping his money on the site. My impression is that are quite a lot of people doing both gambling and investing.

re: "only people who lost want to tinker with the system". Nope. I'm still ahead in invest profit. And I'm in favor of a delay.

re: gamble&re-invest. You're right, I've done so as well. And you still can do that, only with one hour delay.

Also, you're not really addressing my point: invest-trading/gambling serves absolutely no purpose (the example you gave, of gambling, then re-investing is something else), and there is arguably negative consequences. If something has no positive side, and quite possibly a negative side, then it is only rational to get rid of it.

As for it having no purpose, a few pages ago somebody did come up with a good reason: day traders tend to divest when the whale is losing and investment is high, and reinvest when the whale is losing and investing is low. So they are counter-cyclical. That means they tend to keep the max bet stable. Also, if there was a crazy run of bad luck and a bunch of divestments, you might have a situation where people would like to jump in and bring investment up, but they can't and we reach a tipping point where a sort of bank-run happens and the site collapses as a whale gets lucky, people panic-withdraw, and potential investors are locked out.

Also, isn't all this talk of preventing day-trading academic? There's no limit to how many user names you can create, so determined day-traders could just create dozens or hundreds of user names and get around most plausible restrictions.

[GOB]

Given that the probabilities don't change, it wouldn't make sense to invest/divest on winning/losing streaks. That's investors fallacy.

Unless for example the hash distribution is not random, or the server seed is not safe.

As pointed out before, when nakowa is betting, there is a lot of moves, and profit get easily + 10 k or - 6k within hours.
At the end of the day, it always end up with a large negative profit.


Some people says that nakowa got 3 or 4 times the website bankroll and can take a lot of losses ...
So it makes sense to divest when he is loosing a lot, and invest at the end of the period, or when he got lucky.


As I said, you can make up to 30 % profit daily with that, with only investing and divesting once (at the correct time).

No you cannot, you are trying to time the market: a losing proposition. Staying invested throughout is the most profitable strategy. If you want to avoid a lot of variance you can divest. This is why Dooglus divested!

This.

There is no regression to the mean. If we're at +5000, we can go to +10000 as likely as to 0. If someone wants to time the market, do it. You can win or you can lose, but math is not on your side.

For those whining around everytime the profit is down, that's simply because you can't stand the variance. Don't call out for further reglementations, just divest partly. I have done this yesterday myself, I divested about 60% before going to sleep, because I don't want to wake up with a 20% loss.

Regarding day-trading, I have been of the opinion that it is largely irrelevant from the position of the day-trader: you have no knowledge of the next rolls, so if you believe you know when wins or losses are coming, you are falling for the same good'ole gambler's fallacy. Additionally, from the perspective of passive investors, ex-ante it's irrelevant because daytraders could just as easily divest before a huge drop as they could before a huge gain.

Also, don't forget that you're likely to only be hearing from those day traders that have done well, but not from those that do poorly. I'll break the silence of the failed daytraders right here: I tried it once and divested before a HUGE nakowa 7000btc runup, and then invested and immediately took an even bigger drop. F*ck that. So it's not all sunshine and happiness in daytrade land.

HOWEVER!!!

Watching yesterday's ridiculous action and reading the comment above from eltopo (There is no regression to the mean. If we're at +5000, we can go to +10000 as likely as to 0.) have made me realize something: while that statement is true, there is a way you can successfully daytrade because while you do not know future rolls, there is some publicly available info about the future. Bear with me:

IF:

A) You know nokawa follows a random walk, up and down, betting close to max bet, until he loses his entire last deposit OR he makes X BTC (for example, yesterday it was 2k + making up for his previous losses, so 5k)
B) You can see what his balance is (people seem to be able to do this, not sure how)
C) You can see when he is depositing again
D) All other betting going on is relatively so small that we can safely ignore it.

THEN:

A strategy where you divest when he is near losing his last deposit and invest when he is near reaching his goal, then you can make the decision to daytrade EV+ with respect to passive investing. This is because when he is near losing his entire deposit, you only stand to gain very little from him losing it all, and a lot to gain if he doesn't; similarly, when he is near his goal, you stand to lose little if he reaches his goal as compared to how much you stand to gain if he dips back down toward losing his latest deposit. Finally, because (somehow) people know how much his balance is and they see when he has incoming deposits, even if you divest and he loses another, say, 500BTC and loses his deposit, there will be a break while he reloads, so you can invest again and start a new "session".

The point is, like dooglus said, the downside to daytrading is that you might miss a huge 30k slide by nakowa at any point. But the problem with this is that you WON'T miss it (if you're paying attention and day trading), because you'll see nakowa lose his whole balance and then see the new deposit coming.


Does this make sense? So, this would push me toward being anti-daytrading being permitted, but so far I haven't heard any effective methods for preventing it given that you can open unlimited accounts.

[romerun]

Interesting indeed. How do people see Nak balance, and his incoming deposit ?