Satoshi Dice -- Statistical Analysis - page 42.

dooglus

legendary

Activity: 2940

Merit: 1333

Quote from: m0mchil on January 30, 2013, 04:08:03 AM

Quote from: dooglus on January 29, 2013, 04:42:58 PM

I don't really understand that double-spends page - I would expect it to show pairs of conflicting transactions, but it doesn't as far as I can see.

Perhaps piuk will provide more details about the double-spends page. I'd guess that it lists transactions that spend inputs which are used in previously known (to blockchain.info) transactions. The majority of them should not enter a block. Every SDICE profit which is double-spent and confirmed, as with the ones you noted, is a dent in the edge.

I found that if I click on any of the transactions on the double-spends page it shows me a link to the corresponding double-spend transaction. I posted about 3 such transactions here, and retep replied, pointing out that these double-spends could have been made by anyone, not necessarily the big spender. I hadn't thought of that before, but it's quite interesting. Anyone can double-spend any transaction they see on the network by rewriting the transaction without needing access to the sender's private keys. It doesn't change any of the inputs or outputs, but it does change the transaction hash, and so changes the SD lucky number.

dooglus

legendary

Activity: 2940

Merit: 1333

Quote

Total of 92 bets unaccounted for.

Results: 2013-Jan-30 12:51pm (up to block 218786)

Address Target Should Win | #Bets | Win | Lose | Refunds | BTC In | BTC Out | Refund | Profit | RTP
----------------------------------------------------------------------------------------------------------------------------------------------
1dice1e6p 1 0.00002 | 77190 | 0 (0.00000) | 75898 | 1292 | 794.57 | 0.01 | 104.64 | 794.55 | 0.003
1dice1Qf4 2 0.00003 | 3819 | 0 (0.00000) | 3322 | 497 | 62.90 | 0.01 | 17.03 | 62.89 | 0.020
1dice2pxm 4 0.00006 | 6030 | 1 (0.00018) | 5608 | 421 | 87.96 | 159.97 | 8.46 | -72.01 | 181.860
1dice2vQo 8 0.00012 | 8656 | 3 (0.00036) | 8227 | 426 | 131.00 | 176.01 | 10.03 | -45.00 | 134.353
1dice2WmR 16 0.00024 | 10494 | 1 (0.00010) | 10081 | 412 | 311.63 | 4.56 | 18.66 | 307.07 | 1.465
1dice2xkj 32 0.00049 | 11583 | 3 (0.00027) | 11187 | 393 | 619.83 | 303.62 | 1.36 | 316.20 | 48.985
1dice2zdo 64 0.00098 | 13889 | 14 (0.00104) | 13455 | 420 | 887.34 | 230.74 | 55.72 | 656.59 | 26.004
1dice37Ee 128 0.00195 | 16175 | 31 (0.00197) | 15712 | 432 | 1954.53 | 1379.19 | 48.32 | 575.34 | 70.564
1dice3jkp 256 0.00391 | 17970 | 78 (0.00443) | 17514 | 378 | 2249.15 | 3542.76 | 13.17 | -1293.60 | 157.515
1dice4J1m 512 0.00781 | 26955 | 201 (0.00765) | 26071 | 683 | 4447.35 | 4122.08 | 9.98 | 325.26 | 92.686
1dice5wwE 1000 0.01526 | 86267 | 1318 (0.01536) | 84500 | 449 | 22933.29 | 19402.30 | 1.95 | 3530.99 | 84.603
1dice61SN 1500 0.02289 | 20369 | 476 (0.02378) | 19543 | 350 | 6530.84 | 6608.00 | 15.07 | -77.15 | 101.181
1dice6DPt 2000 0.03052 | 51463 | 1582 (0.03097) | 49495 | 386 | 25602.82 | 22872.34 | 9.32 | 2730.47 | 89.335
1dice6gJg 3000 0.04578 | 23694 | 1080 (0.04640) | 22196 | 418 | 8622.25 | 9770.89 | 25.11 | -1148.64 | 113.322
1dice6GV5 4000 0.06104 | 29297 | 1774 (0.06133) | 27153 | 370 | 6673.05 | 6287.39 | 31.28 | 385.66 | 94.221
1dice6wBx 6000 0.09155 | 36314 | 3376 (0.09400) | 32540 | 398 | 14193.64 | 15913.14 | 7.14 | -1719.50 | 112.115
1dice6YgE 8000 0.12207 | 145727 | 17882 (0.12309) | 127393 | 452 | 71680.20 | 71740.95 | 100.23 | -60.75 | 100.085
1dice7EYz 12000 0.18311 | 59724 | 10936 (0.18447) | 48347 | 441 | 143370.18 | 147417.44 | 3314.66 | -4047.26 | 102.823
1dice7fUk 16000 0.24414 | 178918 | 43498 (0.24377) | 134941 | 479 | 328033.09 | 314390.96 | 2056.09 | 13642.12 | 95.841
1dice7W2A 24000 0.36621 | 174521 | 64038 (0.36803) | 109963 | 520 | 503330.95 | 509788.75 | 1012.97 | -6457.79 | 101.283
1dice8EMZ 32000 0.48828 | 926538 | 451613 (0.48817) | 473506 | 1419 | 734450.90 | 716571.23 | 2924.16 | 17879.67 | 97.566
1dice97EC 32768 0.50000 | 422891 | 210756 (0.49941) | 211258 | 877 | 555106.22 | 543980.04 | 5720.40 | 11126.18 | 97.996
1dice9wcM 48000 0.73242 | 271032 | 198946 (0.73548) | 71551 | 535 | 262910.68 | 257133.39 | 5454.88 | 5777.28 | 97.803
1dicec9k7 52000 0.79346 | 53579 | 42195 (0.79387) | 10956 | 428 | 54896.71 | 53754.86 | 1187.20 | 1141.84 | 97.920
1dicegEAr 56000 0.85449 | 39908 | 33766 (0.85627) | 5668 | 474 | 58548.72 | 58077.37 | 400.24 | 471.35 | 99.195
1diceDCd2 60000 0.91553 | 54119 | 49144 (0.91567) | 4526 | 449 | 56642.49 | 56234.29 | 0.22 | 408.20 | 99.279
1dice9wVt 64000 0.97656 | 13137 | 12126 (0.98067) | 239 | 772 | 22469.29 | 22074.52 | 239.70 | 394.77 | 98.243
----------------------------------------------------------------------------------------------------------------------------------------------
small (bets < 4 BTC) | 2684278 | 1099463 | 1570533 | 14282 | 614337.60 | 601241.32 | 225.42 | 13096.28 | 97.868
big (bets >= 4 BTC) | 95981 | 45375 | 50317 | 289 | 2273204.12 | 2240695.62 | 22562.70 | 32508.50 | 98.570
----------------------------------------------------------------------------------------------------------------------------------------------
| 2780259 | 1144838 | 1620850 | 14571 | 2887541.72 | 2841936.94 | 22788.12 | 45604.78 | 98.421
----------------------------------------------------------------------------------------------------------------------------------------------

SD Profit before fees: 45604.78423613 BTC (1.579%)
Cumulative Fees Paid: 1978.33107500 BTC
SD Profit after fees: 43626.45316113 BTC (1.511%)
Pending Liabilities: 12.49110279 BTC
Final SD Profit: 43613.96205834 BTC (1.510%)
----
Since Satoshi Dice started, there have been:
Blockchain Tx: 9057473 : SatoshiDice Tx: 5119162 (56.5%)
Blockchain MB: 3865.4 : SatoshiDice MB: 2102.1 (54.4%)

dooglus

legendary

Activity: 2940

Merit: 1333

Quote from: Kryptox on January 30, 2013, 10:25:26 AM

Yes I read that. I'm a noob when it comes to cryptography. How do you actually verify it? Is there software for this?

EDIT: Ok, I guess maybe I'm being unclear. Here is an example of a previous transaction from SatoshiDice.

http://www.satoshidice.com/full.php?tx=38f381893d3cdcc8f7ea3e11d098f8e7b4daa60e904a62ef0ab4798fa996def8

hmac_sha512(nJawTR7RIrciXb3JIYp9Qoh7YJ1QALxlnQJw3veDN4UkecWcMNa2woxV6y891GXz,38f381893d3cdcc8f7ea3e11d098f8e7b4daa60e904a62ef0ab4798fa996def8) -> 003edf82a3b5a6680a6ede36a573cbf59f7f8d9f1200884755573c66d6d68190

003e -> 62

So that's lucky number was 62. How do you get from:

This
nJawTR7RIrciXb3JIYp9Qoh7YJ1QALxlnQJw3veDN4UkecWcMNa2woxV6y891GX

And this
38f381893d3cdcc8f7ea3e11d098f8e7b4daa60e904a62ef0ab4798fa996def8

To this
003edf82a3b5a6680a6ede36a573cbf59f7f8d9f1200884755573c66d6d68190

Do we just take it for granted that this is somehow correct?

Download and install Python. It's free, and comes with all the crypto libraries you need.

Then you can calculate lucky numbers for yourself, as follows.

Your example is from before the rule-change of Jan 3rd. Before then, all the bets in a transaction got the same lucky number. Here's how you would calculate it for your example (the bold parts are the parts you type). See how it spits out '62' at the end?

Quote

$ python
Python 2.7.3 (default, Sep 26 2012, 21:51:14)
[GCC 4.7.2] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> import hashlib, hmac
>>> txid = '38f381893d3cdcc8f7ea3e11d098f8e7b4daa60e904a62ef0ab4798fa996def8'
>>> secret = 'nJawTR7RIrciXb3JIYp9Qoh7YJ1QALxlnQJw3veDN4UkecWcMNa2woxV6y891GXz'
>>> int(hmac.new(secret, txid, hashlib.sha512).hexdigest()[:4], 16)
62
>>>

For bets made after Jan 2nd 2013, the lucky number is different for each bet in the transaction, so you need to specify the output number (nout), as follows. Let's use a bet I made as an example, and calculate all 10 lucky numbers:

Quote

$ python
Python 2.7.3 (default, Sep 26 2012, 21:51:14)
[GCC 4.7.2] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> import hashlib, hmac
>>> txid = 'df7a5cb1ed439f1e0f314cfed88f77ab0f0fdf4a4d2e4714a8287834791fbd95'
>>> secret = 'DIGtgGbebAIjJeuh1tlzMW5Xzs8ds0V9azgs40JvWOd24NpFEttZEFM32OHgEqFd'
>>> for nout in range(10): print nout, int(hmac.new(secret, "%s:%d" % (txid, nout), hashlib.sha512).hexdigest()[:4], 16)
0 10351
1 37909
2 63099
3 19569
4 13550
5 36796
6 28032
7 53908
8 64653
9 36560
>>>

See how they match up with the numbers shown on the SD page (although the SD page shows them in the wrong order, it does now show the 'Idx' column so you can match them up):

ThickAsThieves

hero member

Activity: 518

Merit: 500

I asked the owner about this topic a few days ago actually, and he convinced me that all that would happen if there were inside betting like this, would be that shareholders would eventually get all the money.

But, to run with this theory, this is only "harmless" fluffing of numbers if the whale loses. But, if the whale has a winning record, does that simply snake dividends from the 10% stakeholders?

I'm probably missing something here that makes this impossible, but if it's possible, there's essentially very little reason for the 90% stakeholder(s) not to constantly bet big on themselves, no?

StarenseN

legendary

Activity: 2478

Merit: 1362

Quote from: ThickAsThieves on January 30, 2013, 09:28:27 AM

Quote from: EskimoBob on January 30, 2013, 09:12:40 AM

Quote from: dooglus on January 29, 2013, 04:42:58 PM

...
The high roller uses the following 8 addresses, by the way:

Code:

15HJgkvj6roZQvNRC3pUJNKP2YgfEuD8wo
17sjYPaYJAvtwXM2q5ojzrCiYewnQMBi9y
1A3z7M6GhpUejQ5eqgSyjVWRuezeepdKKV
1B5h9sf6xNLYGa3NBk48hwYbxyNZ22PRJk
1DBH7Xum2re7LXd8YNSe1uRznttsqeMXGK
1Di6ux3CQBNXCZLD76BsnYWDo6XnFGzQah
1EP4rM8hLZCRSWPNZDWJp9zdvmoWYkkgbQ
1MxgyKiKiHEemMUTJuq9NAbjmhKnUL5Ric

You obviously know what you are doing

Can you figure out, how and where did the coin get to those accounts?
I still like my "idea" (more like a speculation) that this mysterious whale is sdice guy(s) himself and he is only trying to make sdice looks stronger and more profitable to possible future investors.
Only problem is, he "fucked up" and sdice "lost" a boat load of coin

Even if S.DICE was being fluffed, it would come at 10% expense to the "whale" while losing, because 10% of the dividends are public.

Anyone can bet what they want on the site, even majority shareholders with less to lose (because they'll see some of their losses come back thru dividends). This is only a problem if the odds are in the whale's favor, I think...

Actualy long term it's 10% of the house edge (=0.19%) given to stakeholders. So less tah 200$ lost for the whale every 100,000$ bet. That's cheap for a good advertising (even if i don't trust the conspiracy)

Kryptox

member

Activity: 98

Merit: 10

Quote from: Kryptox on January 30, 2013, 10:25:26 AM

I'm a noob when it comes to cryptography.

Quote from: kgo on January 30, 2013, 11:40:17 AM

You run the function yourself with the same inputs with a crypto library that you trust.

Thanks kgo. Do these crypto libraries exist somewhere on the net? I did a google search and yielded nothing.

kgo

hero member

Activity: 548

Merit: 500

Quote from: Kryptox on January 30, 2013, 10:25:26 AM

Do we just take it for granted that this is somehow correct?

You run the function yourself with the same inputs with a crypto library that you trust.

Kryptox

member

Activity: 98

Merit: 10

Yes I read that. I'm a noob when it comes to cryptography. How do you actually verify it? Is there software for this?

EDIT: Ok, I guess maybe I'm being unclear. Here is an example of a previous transaction from SatoshiDice.

http://www.satoshidice.com/full.php?tx=38f381893d3cdcc8f7ea3e11d098f8e7b4daa60e904a62ef0ab4798fa996def8

hmac_sha512(nJawTR7RIrciXb3JIYp9Qoh7YJ1QALxlnQJw3veDN4UkecWcMNa2woxV6y891GXz,38f381893d3cdcc8f7ea3e11d098f8e7b4daa60e904a62ef0ab4798fa996def8) -> 003edf82a3b5a6680a6ede36a573cbf59f7f8d9f1200884755573c66d6d68190

003e -> 62

So that's lucky number was 62. How do you get from:

This
nJawTR7RIrciXb3JIYp9Qoh7YJ1QALxlnQJw3veDN4UkecWcMNa2woxV6y891GX

And this
38f381893d3cdcc8f7ea3e11d098f8e7b4daa60e904a62ef0ab4798fa996def8

To this
003edf82a3b5a6680a6ede36a573cbf59f7f8d9f1200884755573c66d6d68190

Do we just take it for granted that this is somehow correct?

ingrownpocket

legendary

Activity: 952

Merit: 1000

Quote from: Kryptox on January 30, 2013, 10:18:27 AM

How does the secret compute against the Transaction ID to produce the lucky number?

Quote

The lucky number used to determine the winner of games is simple. It is simply the first bytes of hmac_sha512(secret,txid:out_idx). That would be the secret string as the key and the transaction ID of your bet transaction as the data.

BTW, there's a error in SD site: "hmac_sha512(secert,txid:out_idx)".

Kryptox

member

Activity: 98

Merit: 10

I've been searching around and can't find an answer to this so I apologize if this has been covered (I'm sure it has) but earlier in this thread there was a discussion about how SD cannot be rigged and it's provable.

How exactly do you go about verifying that the lucky number is calculated correctly from the Transaction ID? How does the secret compute against the Transaction ID to produce the lucky number?

ingrownpocket

legendary

Activity: 952

Merit: 1000

Quote from: dooglus on January 29, 2013, 04:42:58 PM

...
The high roller uses the following 8 addresses, by the way:

Code:

15HJgkvj6roZQvNRC3pUJNKP2YgfEuD8wo
17sjYPaYJAvtwXM2q5ojzrCiYewnQMBi9y
1A3z7M6GhpUejQ5eqgSyjVWRuezeepdKKV
1B5h9sf6xNLYGa3NBk48hwYbxyNZ22PRJk
1DBH7Xum2re7LXd8YNSe1uRznttsqeMXGK
1Di6ux3CQBNXCZLD76BsnYWDo6XnFGzQah
1EP4rM8hLZCRSWPNZDWJp9zdvmoWYkkgbQ
1MxgyKiKiHEemMUTJuq9NAbjmhKnUL5Ric

15HJgkvj6roZQvNRC3pUJNKP2YgfEuD8wo	590 BTC
17sjYPaYJAvtwXM2q5ojzrCiYewnQMBi9y	4036 BTC
1A3z7M6GhpUejQ5eqgSyjVWRuezeepdKKV	1475 BTC
1B5h9sf6xNLYGa3NBk48hwYbxyNZ22PRJk	0 BTC
1DBH7Xum2re7LXd8YNSe1uRznttsqeMXGK	1371 BTC
1Di6ux3CQBNXCZLD76BsnYWDo6XnFGzQah	9768 BTC
1EP4rM8hLZCRSWPNZDWJp9zdvmoWYkkgbQ	5080 BTC
1MxgyKiKiHEemMUTJuq9NAbjmhKnUL5Ric	800 BTC

Total: 23120 BTC ($447,140)

ThickAsThieves

hero member

Activity: 518

Merit: 500

Quote from: EskimoBob on January 30, 2013, 09:12:40 AM

Quote from: dooglus on January 29, 2013, 04:42:58 PM

...
The high roller uses the following 8 addresses, by the way:

Code:

15HJgkvj6roZQvNRC3pUJNKP2YgfEuD8wo
17sjYPaYJAvtwXM2q5ojzrCiYewnQMBi9y
1A3z7M6GhpUejQ5eqgSyjVWRuezeepdKKV
1B5h9sf6xNLYGa3NBk48hwYbxyNZ22PRJk
1DBH7Xum2re7LXd8YNSe1uRznttsqeMXGK
1Di6ux3CQBNXCZLD76BsnYWDo6XnFGzQah
1EP4rM8hLZCRSWPNZDWJp9zdvmoWYkkgbQ
1MxgyKiKiHEemMUTJuq9NAbjmhKnUL5Ric

You obviously know what you are doing

Can you figure out, how and where did the coin get to those accounts?
I still like my "idea" (more like a speculation) that this mysterious whale is sdice guy(s) himself and he is only trying to make sdice looks stronger and more profitable to possible future investors.
Only problem is, he "fucked up" and sdice "lost" a boat load of coin

Even if S.DICE was being fluffed, it would come at 10% expense to the "whale" while losing, because 10% of the dividends are public.

Anyone can bet what they want on the site, even majority shareholders with less to lose (because they'll see some of their losses come back thru dividends). This is only a problem if the odds are in the whale's favor, I think...

EskimoBob

legendary

Activity: 910

Merit: 1000

Quality Printing Services by Federal Reserve Bank

Quote from: dooglus on January 29, 2013, 04:42:58 PM

...
The high roller uses the following 8 addresses, by the way:

Code:

15HJgkvj6roZQvNRC3pUJNKP2YgfEuD8wo
17sjYPaYJAvtwXM2q5ojzrCiYewnQMBi9y
1A3z7M6GhpUejQ5eqgSyjVWRuezeepdKKV
1B5h9sf6xNLYGa3NBk48hwYbxyNZ22PRJk
1DBH7Xum2re7LXd8YNSe1uRznttsqeMXGK
1Di6ux3CQBNXCZLD76BsnYWDo6XnFGzQah
1EP4rM8hLZCRSWPNZDWJp9zdvmoWYkkgbQ
1MxgyKiKiHEemMUTJuq9NAbjmhKnUL5Ric

You obviously know what you are doing

Can you figure out, how and where did the coin get to those accounts?
I still like my "idea" (more like a speculation) that this mysterious whale is sdice guy(s) himself and he is only trying to make sdice looks stronger and more profitable to possible future investors.
Only problem is, he "fucked up" and sdice "lost" a boat load of coin

m0mchil

full member

Activity: 171

Merit: 127

Quote from: dooglus on January 29, 2013, 04:42:58 PM

I don't really understand that double-spends page - I would expect it to show pairs of conflicting transactions, but it doesn't as far as I can see.

Perhaps piuk will provide more details about the double-spends page. I'd guess that it lists transactions that spend inputs which are used in previously known (to blockchain.info) transactions. The majority of them should not enter a block. Every SDICE profit which is double-spent and confirmed, as with the ones you noted, is a dent in the edge.

nimda

hero member

Activity: 784

Merit: 1000

0xFB0D8D1534241423

Quote from: Bowjob on January 29, 2013, 03:51:03 PM

Dice shares will now go down the drain

Profit is still up for the month

dooglus

legendary

Activity: 2940

Merit: 1333

Quote from: ThickAsThieves on January 29, 2013, 04:46:34 PM

Is that to say, that once the double spends iron out, the total could be different, and these numbers aren't necessarily totally confirmed yet?

No, the numbers I've published are only for bets that have been confirmed, and had their payout transactions confirmed too.

The double-spends page just makes me think that maybe the big player has been deliberately double-spending some of his losing bets.

Quote from: StarenseN on January 29, 2013, 04:49:41 PM

Thanks for the deep analysis dooglus.

+0.5 tip

Thank you.

StarenseN

legendary

Activity: 2478

Merit: 1362

Thanks for the deep analysis dooglus.

+0.5 tip

ThickAsThieves

hero member

Activity: 518

Merit: 500

Quote from: dooglus on January 29, 2013, 04:42:58 PM

Except... that http://blockchain.info/double-spends lists a few big bets from the high roller, for example this one:

Is that to say, that once the double spends iron out, the total could be different, and these numbers aren't necessarily totally confirmed yet?

dooglus

legendary

Activity: 2940

Merit: 1333

Quote from: dooglus on January 29, 2013, 04:19:40 PM

It looks legit. It's the same big player who has been playing all week.

Except... that http://blockchain.info/double-spends lists a few big bets from the high roller, for example this one:

I don't really understand that double-spends page - I would expect it to show pairs of conflicting transactions, but it doesn't as far as I can see.

The high roller uses the following 8 addresses, by the way:

Code:

15HJgkvj6roZQvNRC3pUJNKP2YgfEuD8wo
17sjYPaYJAvtwXM2q5ojzrCiYewnQMBi9y
1A3z7M6GhpUejQ5eqgSyjVWRuezeepdKKV
1B5h9sf6xNLYGa3NBk48hwYbxyNZ22PRJk
1DBH7Xum2re7LXd8YNSe1uRznttsqeMXGK
1Di6ux3CQBNXCZLD76BsnYWDo6XnFGzQah
1EP4rM8hLZCRSWPNZDWJp9zdvmoWYkkgbQ
1MxgyKiKiHEemMUTJuq9NAbjmhKnUL5Ric

dooglus

legendary

Activity: 2940

Merit: 1333

It looks legit. It's the same big player who has been playing all week.

The 10k loss is pretty much entirely accounted for by 222 bets of 250 BTC on lessthan 24000.

123 lost, losing 248.7505 BTC per bet for a total loss of 30596.3115 BTC,
99 won, winning 417.53216666 BTC per bet for a total win of 41335.68449934 BTC,
making a net win of 10739.37299934 BTC over the 222 bets.

Here's a graph of the probability of winning X out of 222 bets on lessthan 24000. The most likely number of wins is 81:

Code:

>>> 222 * float(24000)/65536
81.298828125

and 99 is quite unlikely:

Edit: here's some Python code to generate the probabilities:

Code:

#!/usr/bin/env python

import string, sys
lessthan, trials = string.atoi(sys.argv[1]), string.atoi(sys.argv[2])

def factorial(x):
  if x < 2: return 1
  return x * factorial(x - 1)

def c(n, k): return factorial(n) / (factorial(k) * factorial(n - k))

p = float(lessthan) / 65536
for x in range(trials + 1): print "%3d %f" % (x, c(trials, x) * p**x * (1-p)**(trials-x))

and an example of it running (for 6 plays on lessthan 32000):

Code:

$ ./binomial.py 32000 6
  0 0.017955
  1 0.102796
  2 0.245220
  3 0.311985
  4 0.223272
  5 0.085218
  6 0.013553

Topic: Satoshi Dice -- Statistical Analysis - page 42. (Read 192896 times)