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

[nicolaennio]

I had been talking in the chat about increasing the house edge for bets with profit over 50 BTC (since none of the competition with a 1% edge allows bets that big anyway).  This was obviously his way of 'showing me' that it won't help...

Yes, but if he bets lower amounts, like 20B, and he consistently wins in an unlikely fashion (for example he wagers 20000 and wins other 20000), then we could really think he compromised the server.

[1715503671]

1715503671

[1715503671]

1715503671

[1715503671]

1715503671

[1715503671]

1715503671

[1715503671]

1715503671

[1715503671]

1715503671

[molecular]

Following on from my last post.  If he waits for 2 small bets where the first wins and the 2nd loses, then bets big, then shifting by 1 will result in the big bet losing, while shifting by 2 will result in it winning.

Pretty nice result. You guys found his betting pattern by shifting. shift 1: mostly loss, shift 2: mostly win... hence pattern is: wait for "WL", then bet high.

[Rampion]

Following on from my last post.  If he waits for 2 small bets where the first wins and the 2nd loses, then bets big, then shifting by 1 will result in the big bet losing, while shifting by 2 will result in it winning.

Pretty nice result. You guys found his betting pattern by shifting. shift 1: mostly loss, shift 2: mostly win... hence pattern is: wait for "WL", then bet high.

Which is still unprofitable in the very long run, despite of his rational "betting management" because of the 1% house edge, and would be much more unprofitable if the house edge would be higher.

[dooglus]

If you can break down his behaviour into rules then in theory at least you could check to see whether it was just luck or he'd somehow spotted a real pattern (meaning the seeding system is broken).  That would mean running loads of tests where you ran the betting rules against a long series of bets from multiple sets of seeds.

The problem is, I can't.  Here's the most recent betting session, as posted a few posts up from here.  I've separated the small (left column) and large (right column) bets out, and tidied up the formatting.  I don't see any obvious rules being followed.  Do you?

Read as follows: he bet small, saw a win then a loss, then bet 1 BTC and lost.  Then he bet 5 more small bets, seeing 4 wins and a loss, then bet 2 BTC twice, losing both.  Etc.  Those first two sequences follow the "wait for WL then bet big" rule.  But the rest of the session doesn't see it happen again at all:

W L                              1 lose

W W W W L                        2 lose
                                 2 lose

L L W                            5 win
                                 5 win
                                 5 lose

L                                2 win

L                                2 lose
                                 2 lose
                                 4 win
                                22 lose
                                22 lose

W                               11 win
                                11 lose
                                20 win

W                               20 win

L                               10 lose
                                10 win

L                                2 lose
                                 4 lose
                                16 lose

L W W                           10 win
                                10 lose
                                20 win

W                               10 lose
                                20 lose

L L L W W                       10 win
                                10 lose
                                20 win
                                10 win
                                10 win
                                10 win
                                10 win
                                10 win
                                10 win
                                10 win
                                10 lose
                                20 lose

L W                             10 lose
                                20 lose

W                               20 lose
                                20 lose

W L                             20 lose
                                20 lose

L L L L L W W                   20 win
                                20 win
                                20 lose
                                10 win

L                               20 win

L                               10 lose
                                20 win

L                               10 win

W                               10 lose
                                20 lose

L W W L L L W W L L L W L       10 win

W                               10 win
                                10 lose
                                 5 win

L                               20 win

L                               10 lose
                                20 lose

W L                             40 win

L                               10 win

W                               10 lose
                                20 lose
                                40 win
                                10 lose
                                20 lose

W L                             20 lose
                                20 win
                                40 win
                                20 win
                                20 lose
                                40 win

W                               10 lose
                                20 lose

L                               30 win

W                               10 lose

W L                             10 lose
                                20 win

L                               10 win

L                               10 win

W                               10 win
                                10 win
                                10 win
                                10 win
                                10 lose
                                20 win
                                20 lose
                                40 win

L                               10 lose

L L L W W                       20 win
                                20 win
                                20 win
                                20 lose
                                40 win
                                20 win

[nicolaennio]

2013-09-25 07:13:07 (153338) ACTION (debug, (none)) bitcoin account has 156 BTC
2013-09-25 07:27:04 (153338) chat: <percent> I think I've proved only raising HE for large bets (profit larger than 50) is not effective to JD.
2013-09-25 07:27:24 (153338) chat: <percent> I only deposited 156, now 406.
2013-09-25 07:27:57 (153338) chat: <percent> I'm not in hurry.
2013-09-25 07:28:25 (153338) chat: <percent> done for today. already 75% up.
2013-09-25 07:38:38 (153338) ACTION (withdraw, "125.34.52.64") success: 255.99990000 to 17ziML8b3hZEKpYeJu3UeSx8TTBDKTDHoU txid 77df6a3b86765386793b0033e56fc29f5a1cae37b2ee7c001ffaa0cf99b144bb

If he was losing he would not have said on the chat, he boasts after getting the lucky strike... a real statistical proof would be to repeat the lucky strike ten or more times.

Secondly, he is reading the forum and he is worried that dooglus actually lowers the max bet, and if he is worried it means that it is not cheating at all, but he is just trying to play out the bankroll (which is easier with max bets).

Thirdly, I fear that dooglus is getting stressed about this issue, for many reasons (people are getting nervous and the situation is new). I hope no one here is really thinking that the whale has the "right martingale" or that they can "analyse the betting pattern". COME ON! What is happening is that the whale just wants to make dooglus nervous by exploiting the unpredictability of luck.

dooglus please do not get stressed, do not believe in martingales or as such and do not lose confidence in yourself. Until a STRONG AND CONVINCING statistical proof arrives, the whale is just a gambler with tons of bitcoins.

[drawingthesun]

Increasing house edge of large bets is also a really good idea. This should be implemented immediately.

[Peter R]

3. Now, if all you do is calculate what he WOULD have won had the dice result been shifted by 2, then I calculate +420,784 BTC! 

He has said before that "the most common sequence of wins and losses is WLWLWL" or words to that effect.

If he bets small, waiting for a win followed by a loss, and then bets big, expecting a win next, he'll be right about half the time.

If you shift his big bets back two steps then he'll be right every time (since the two bets before the big one were W and L) - the big bet will land on the first observed W rather than the third hoped-for W.

Surely that explains your observed result doesn't it?  Or were you shifting in the opposite direction?

Nice work dooglus!  That explains everything.  I checked, and yes, the shifting is in the right direction.

So, this work exposes part of Nakowa's strategy, but in no way suggests that he has access to the server seed.  The outcome in my graphs was extremely improbable due to "randomness" because it wasn't randomness, it was the result of him weighting his bets based on the outcome of his previous rolls. 

I should point out that Oleander suggested this a few pages back, so cudos to him too. 

[drawingthesun]

3. Now, if all you do is calculate what he WOULD have won had the dice result been shifted by 2, then I calculate +420,784 BTC! 

He has said before that "the most common sequence of wins and losses is WLWLWL" or words to that effect.

If he bets small, waiting for a win followed by a loss, and then bets big, expecting a win next, he'll be right about half the time.

If you shift his big bets back two steps then he'll be right every time (since the two bets before the big one were W and L) - the big bet will land on the first observed W rather than the third hoped-for W.

Surely that explains your observed result doesn't it?  Or were you shifting in the opposite direction?

Nice work dooglus!  That explains everything.  I checked, and yes, the shifting is in the right direction.

So, this work exposes part of Nakowa's strategy, but in no way suggests that he has access to the server seed.  The outcome in my graphs was extremely improbable due to "randomness" because it wasn't randomness, it was the result of him weighting his bets based on the outcome of his previous rolls. 

I should point out that Oleander suggested this a few pages back, so cudos to him too. 

Which means his entire play is a gamblers fallacy. Correct?

[Rampion]

3. Now, if all you do is calculate what he WOULD have won had the dice result been shifted by 2, then I calculate +420,784 BTC!  

He has said before that "the most common sequence of wins and losses is WLWLWL" or words to that effect.

If he bets small, waiting for a win followed by a loss, and then bets big, expecting a win next, he'll be right about half the time.

If you shift his big bets back two steps then he'll be right every time (since the two bets before the big one were W and L) - the big bet will land on the first observed W rather than the third hoped-for W.

Surely that explains your observed result doesn't it?  Or were you shifting in the opposite direction?

Nice work dooglus!  That explains everything.  I checked, and yes, the shifting is in the right direction.

So, this work exposes part of Nakowa's strategy, but in no way suggests that he has access to the server seed.  The outcome in my graphs was extremely improbable due to "randomness" because it wasn't randomness, it was the result of him weighting his bets based on the outcome of his previous rolls.  

I should point out that Oleander suggested this a few pages back, so cudos to him too.  

Which means his entire play is a gamblers fallacy. Correct?

Exactly. We spotted that long time ago. He bets "low" until he "spots" the pattern (eg: many losses in a row), and then he starts to bet high because he seems to think that after X losses or a certain WL pattern the probability to win is higher. It definitely looks to me like a semi-martingale strategy (because his "high" bet is at least x2 than his low bet) totally based on gambler's phallacy (eg: after many L in a row the probability of a W is higher).

IMO his biggest advantage compared to the house is that he is a) willing to lose insane amounts of coins without stopping; b) he seems quite disciplined, and he quits as soon as he is winning. Obviously the house cannot "stop when its winning", and with Nakowa's bankroll and daring strategy, the house edge is too low and the max bet too high to avoid massive variance.

[dooglus]

Dooglus, I admire how you react on the situation. You payout every time and I think that is worth a compliment. I know for sure that many other sites would just shutdown / block people like him to prevent those huge losses.

Thanks.

But what option do I have other than paying out?  Who would play on a site which doesn't pay out if you get too lucky?

Blocking him is an interesting one.  Personally I would like if I could stop him playing.  But I can't, since he constantly switches accounts and IP addresses, and also the majority of investors seem to welcome his play, expecting him to lose everything one day.

Shutting down is always on my mind.  But again, investors want the chance to win their coins back.  I would encourage those who have had enough carnage to divest - like I did!

Nakowa is playing as I write this - he's currently "percent (153338)":

02:46:18 *** percent (153338) [#138901997] bet 10 BTC at 49.5% and won 10 BTC ***
02:46:19 *** percent (153338) [#138902026] bet 10 BTC at 49.5% and lost ***
02:46:22 *** percent (153338) [#138902063] bet 20 BTC at 49.5% and lost ***

He seems to be on a mission to demonstrate that he doesn't need to bet more than 40 BTC at a time to win.

[Rampion]

Dooglus, I admire how you react on the situation. You payout every time and I think that is worth a compliment. I know for sure that many other sites would just shutdown / block people like him to prevent those huge losses.

Thanks.

But what option do I have other than paying out?  Who would play on a site which doesn't pay out if you get too lucky?

Blocking him is an interesting one.  Personally I would like if I could stop him playing.  But I can't, since he constantly switches accounts and IP addresses, and also the majority of investors seem to welcome his play, expecting him to lose everything one day.

Shutting down is always on my mind.  But again, investors want the chance to win their coins back.  I would encourage those who have had enough carnage to divest - like I did!

Nakowa is playing as I write this - he's currently "percent (153338)":

02:46:18 *** percent (153338) [#138901997] bet 10 BTC at 49.5% and won 10 BTC ***
02:46:19 *** percent (153338) [#138902026] bet 10 BTC at 49.5% and lost ***
02:46:22 *** percent (153338) [#138902063] bet 20 BTC at 49.5% and lost ***

He seems to be on a mission to demonstrate that he doesn't need to bet more than 40 BTC at a time to win.

Wow, you divested? I didn't and I don't plan to do it.. I'm willing to lose everything because I trust you, and because I trust you I truly believe nakowa's strategy is pure luck and there is no flaw on the site.

Let me ask you a direct question: are you entertaining the idea he might be actually cheating somehow? Honestly, I've played a lot of poker and I've seen much crazier and unlikely variance many times. Nakowa's winning do not seem so strange to me.

[Progressive]

dooglus, if you think you would like single one player to stop playing in the casino, you should change the rules that allows him (or others playing just like him) to bother you.

[dooglus]

IMO his biggest advantage compared to the house is that he is a) willing to lose insane amounts of coins without stopping; b) he seems quite disciplined, and he quits as soon as he is winning. Obviously the house cannot "stop when its winning"

I'm not sure about that.  I've seen it mentioned a few times recently, also in connection with how the Kelly criterion doesn't apply to Just-Dice because the house can't decide when to stop playing.

Nakowa doesn't stop playing.  He pauses to sleep when he's ahead, but then he plays again.  As far as the numbers are concerned his bets are a single continuous stream.  He's playing right now.  He didn't "quit when he was ahead", he just paused for some hours.

Checking the logs, actually he's not playing at the moment, because he busted:

09:47:49 bet #138902848: 0.01 BTC @ 49.5% hi: lucky:489187 profit:-0.01000000 bal:33.20237624 n:400 percent tot:133.20237624
09:47:52 bet #138902894: 2 BTC @ 49.5% hi: lucky:499765 profit:-2.00000000 bal:31.20237624 n:401 percent tot:131.20237624
09:47:54 bet #138902932: 4 BTC @ 49.5% hi: lucky:247064 profit:-4.00000000 bal:27.20237624 n:402 percent tot:127.20237624
09:47:58 bet #138902981: 16 BTC @ 49.5% hi: lucky:75006 profit:-16.00000000 bal:11.20237624 n:403 percent tot:111.20237624
09:48:01 bet #138903016: 0.01 BTC @ 49.5% hi: lucky:855437 profit:0.01000000 bal:11.21237624 n:404 percent tot:111.21237624
09:48:04 bet #138903076: 0.01 BTC @ 49.5% hi: lucky:559395 profit:0.01000000 bal:11.22237624 n:405 percent tot:111.22237624
09:48:07 bet #138903121: 3 BTC @ 49.5% hi: lucky:202450 profit:-3.00000000 bal:8.22237624 n:406 percent tot:108.22237624
09:48:17 bet #138903160: 6 BTC @ 49.5% hi: lucky:266205 profit:-6.00000000 bal:2.22237624 n:407 percent tot:102.22237624
09:48:21 bet #138903215: 0.01 BTC @ 49.5% hi: lucky:804467 profit:0.01000000 bal:2.23237624 n:408 percent tot:102.23237624
09:48:22 bet #138903241: 0.01 BTC @ 49.5% hi: lucky:180762 profit:-0.01000000 bal:2.22237624 n:409 percent tot:102.22237624
09:48:28 bet #138903329: 2.22237624 BTC @ 49.5% hi: lucky:49616 profit:-2.22237624 bal:0.00000000 n:410 percent tot:100.00000000
09:52:31 chat: <percent> I'm still 100 up. don't panic.
09:53:43 chat: <percent> I won't bet big though. since the house edge is going to raise when profit larger than 50.

So he's 'paused', waiting for his next deposit to confirm.

But my point is, being able to chose when to stop isn't an advantage for the players compared to the site, unless they really stop, and nobody with a similar bankroll appears to take their place.

[Rampion]

IMO his biggest advantage compared to the house is that he is a) willing to lose insane amounts of coins without stopping; b) he seems quite disciplined, and he quits as soon as he is winning. Obviously the house cannot "stop when its winning"

I'm not sure about that.  I've seen it mentioned a few times recently, also in connection with how the Kelly criterion doesn't apply to Just-Dice because the house can't decide when to stop playing.

Nakowa doesn't stop playing.  He pauses to sleep when he's ahead, but then he plays again.  As far as the numbers are concerned his bets are a single continuous stream.  He's playing right now.  He didn't "quit when he was ahead", he just paused for some hours.

Checking the logs, actually he's not playing at the moment, because he busted:

09:47:49 bet #138902848: 0.01 BTC @ 49.5% hi: lucky:489187 profit:-0.01000000 bal:33.20237624 n:400 percent tot:133.20237624
09:47:52 bet #138902894: 2 BTC @ 49.5% hi: lucky:499765 profit:-2.00000000 bal:31.20237624 n:401 percent tot:131.20237624
09:47:54 bet #138902932: 4 BTC @ 49.5% hi: lucky:247064 profit:-4.00000000 bal:27.20237624 n:402 percent tot:127.20237624
09:47:58 bet #138902981: 16 BTC @ 49.5% hi: lucky:75006 profit:-16.00000000 bal:11.20237624 n:403 percent tot:111.20237624
09:48:01 bet #138903016: 0.01 BTC @ 49.5% hi: lucky:855437 profit:0.01000000 bal:11.21237624 n:404 percent tot:111.21237624
09:48:04 bet #138903076: 0.01 BTC @ 49.5% hi: lucky:559395 profit:0.01000000 bal:11.22237624 n:405 percent tot:111.22237624
09:48:07 bet #138903121: 3 BTC @ 49.5% hi: lucky:202450 profit:-3.00000000 bal:8.22237624 n:406 percent tot:108.22237624
09:48:17 bet #138903160: 6 BTC @ 49.5% hi: lucky:266205 profit:-6.00000000 bal:2.22237624 n:407 percent tot:102.22237624
09:48:21 bet #138903215: 0.01 BTC @ 49.5% hi: lucky:804467 profit:0.01000000 bal:2.23237624 n:408 percent tot:102.23237624
09:48:22 bet #138903241: 0.01 BTC @ 49.5% hi: lucky:180762 profit:-0.01000000 bal:2.22237624 n:409 percent tot:102.22237624
09:48:28 bet #138903329: 2.22237624 BTC @ 49.5% hi: lucky:49616 profit:-2.22237624 bal:0.00000000 n:410 percent tot:100.00000000
09:52:31 chat: <percent> I'm still 100 up. don't panic.
09:53:43 chat: <percent> I won't bet big though. since the house edge is going to raise when profit larger than 50.

So he's 'paused', waiting for his next deposit to confirm.

But my point is, being able to chose when to stop isn't an advantage for the players compared to the site, unless they really stop, and nobody with a similar bankroll appears to take their place.

Yes, I guess you're absolutely right - he never "quits", he just keeps playing over and over. Did you already increase the house edge for bigger bets?

One question: you have done thorough analysis of dice sites like SatoshiDice, you know math, does this looks like cheating to you? I didn't do such a thorough analysis, but what Nakowa is doing doesn't look like cheating to me, just variance.

[Peter R]

3. Now, if all you do is calculate what he WOULD have won had the dice result been shifted by 2, then I calculate +420,784 BTC!  

He has said before that "the most common sequence of wins and losses is WLWLWL" or words to that effect.

If he bets small, waiting for a win followed by a loss, and then bets big, expecting a win next, he'll be right about half the time.

If you shift his big bets back two steps then he'll be right every time (since the two bets before the big one were W and L) - the big bet will land on the first observed W rather than the third hoped-for W.

Surely that explains your observed result doesn't it?  Or were you shifting in the opposite direction?

Nice work dooglus!  That explains everything.  I checked, and yes, the shifting is in the right direction.

So, this work exposes part of Nakowa's strategy, but in no way suggests that he has access to the server seed.  The outcome in my graphs was extremely improbable due to "randomness" because it wasn't randomness, it was the result of him weighting his bets based on the outcome of his previous rolls.  

I should point out that Oleander suggested this a few pages back, so cudos to him too.  

Which means his entire play is a gamblers fallacy. Correct?

Exactly. ...

Or at least we haven't proven that it's not

[herzmeister]

soooooo is it a gambler's fallacy to invest when the profit of the site "is down" by over 5,000? 

[Rampion]

soooooo is it a gambler's fallacy to invest when the profit of the site "is down" by over 5,000?  

If you believe that the probability of the profit going up is now higher because it went down a lot, then yes: that would be a gambler's fallacy.

[Rampion]

Nokowa is playing again, with much smaller bets. It looks like he is in a losing strike...

[Rannasha]

soooooo is it a gambler's fallacy to invest when the profit of the site "is down" by over 5,000?  

It's Gambler's Fallacy if you let the wager-history factor into your decission whether to invest or not. You should look at expected amount wagered, total amount invested and expected volatility and nothing more.

[dooglus]

dooglus, if you think you would like single one player to stop playing in the casino, you should change the rules that allows him (or others playing just like him) to bother you.

You're right.  I need to work out how to change the rules.

I'd like to increase the house edge for large bets.  But how exactly?  The people betting 500 BTC at 98% are getting a pretty bad deal already, and don't need the house edge increased.  So it shouldn't be based solely on stake.

Perhaps on potential profit?  But then I'm stuck.  Profit depends on payout, which depends on house edge, which now depends on profit.  That's a confusing circular dependency.  How do I communicate that to the player via the interface in a clear way?

If they're martingale betting, 49.5% for 2x, bet 32 BTC and lose, then hit 'x2', the potential profit of the next bet is 64 BTC, so the edge needs to go up.  Do I just quietly adjust the 2x to 1.9x or whatever?  Then if they win, their win doesn't cover the previous losses.  Or do I adjust the chance of winning?  Or prompt for which to change?  Or have two separate tables, one with a 50 BTC max profit and 1% house edge, and another with the full 1% max profit but higher house edge?  Or should the house edge increase linearly with potential profit?

I'd like to get it right, so as not to overwhelm the players with too many changes.  I would appreciate any feedback.