Satoshi Dice -- Statistical Analysis - page 77.

dooglus

legendary

Activity: 2940

Merit: 1333

Quote

Results: 2012-Jul-02 01:23pm (up to block 187242)

Address Target Should Win | #Bets | Win | Lose | Refunds | BTC In | BTC Out | Profit | RTP
-----------------------------------------------------------------------------------------------------------------------------------------
1dice1e6p 1 0.00002 | 6250 | 0 (0.00000) | 5992 | 257 | 21.71 | 0.01 | 21.69 | 0.088
1dice1Qf4 2 0.00003 | 888 | 0 (0.00000) | 821 | 67 | 7.15 | 0.00 | 7.14 | 0.025
1dice2pxm 4 0.00006 | 1397 | 0 (0.00000) | 1367 | 30 | 12.15 | 0.00 | 12.15 | 0.046
1dice2vQo 8 0.00012 | 1102 | 0 (0.00000) | 1064 | 38 | 15.19 | 0.00 | 15.18 | 0.031
1dice2WmR 16 0.00024 | 1358 | 0 (0.00000) | 1330 | 28 | 21.01 | 0.01 | 20.99 | 0.076
1dice2xkj 32 0.00049 | 3092 | 1 (0.00032) | 3080 | 11 | 89.89 | 100.38 | -10.48 | 111.666
1dice2zdo 64 0.00098 | 4460 | 5 (0.00113) | 4438 | 17 | 176.73 | 110.74 | 65.99 | 62.661
1dice37Ee 128 0.00195 | 5901 | 12 (0.00205) | 5840 | 47 | 1197.34 | 1142.16 | 55.17 | 95.392
1dice3jkp 256 0.00391 | 4399 | 20 (0.00456) | 4366 | 12 | 458.86 | 330.67 | 128.18 | 72.064
1dice4J1m 512 0.00781 | 5462 | 32 (0.00587) | 5424 | 5 | 748.28 | 330.60 | 417.67 | 44.182
1dice5wwE 1000 0.01526 | 6656 | 93 (0.01398) | 6558 | 2 | 1377.12 | 711.48 | 665.64 | 51.664
1dice61SN 1500 0.02289 | 5172 | 108 (0.02091) | 5057 | 6 | 1913.72 | 2135.76 | -222.03 | 111.602
1dice6DPt 2000 0.03052 | 5628 | 174 (0.03096) | 5447 | 3 | 1786.52 | 1635.04 | 151.47 | 91.521
1dice6gJg 3000 0.04578 | 5518 | 277 (0.05028) | 5232 | 7 | 3791.62 | 4916.50 | -1124.87 | 129.667
1dice6GV5 4000 0.06104 | 5586 | 344 (0.06163) | 5238 | 3 | 1619.49 | 1429.16 | 190.33 | 88.247
1dice6wBx 6000 0.09155 | 9550 | 893 (0.09358) | 8650 | 4 | 6146.27 | 6745.93 | -599.66 | 109.756
1dice6YgE 8000 0.12207 | 10612 | 1297 (0.12232) | 9306 | 0 | 2952.77 | 2544.50 | 408.26 | 86.173
1dice7EYz 12000 0.18311 | 14629 | 2789 (0.19087) | 11823 | 5 | 5962.82 | 6312.56 | -349.74 | 105.865
1dice7fUk 16000 0.24414 | 35975 | 8723 (0.24258) | 27236 | 6 | 9011.21 | 8694.07 | 317.13 | 96.481
1dice7W2A 24000 0.36621 | 24066 | 8887 (0.36977) | 15147 | 30 | 10854.86 | 11103.44 | -248.57 | 102.290
1dice8EMZ 32000 0.48828 | 272702 | 132734 (0.48701) | 139817 | 121 | 80984.06 | 82596.47 | -1612.40 | 101.991
1dice97EC 32768 0.50000 | 106458 | 53176 (0.49984) | 53210 | 64 | 36476.09 | 35805.28 | 670.81 | 98.161
1dice9wcM 48000 0.73242 | 66349 | 48731 (0.73487) | 17581 | 32 | 38074.31 | 36695.62 | 1378.68 | 96.379
1dice9wVt 64000 0.97656 | 5694 | 4598 (0.97457) | 120 | 975 | 4901.49 | 4709.70 | 191.79 | 96.087
-----------------------------------------------------------------------------------------------------------------------------------------
| 608904 |
-----------------------------------------------------------------------------------------------------------------------------------------

Cumulative Wagers: 212865.75014757 BTC
Cumulative Rewards: 212275.45778249 BTC
Cumulative Fees Paid: 306.38627500 BTC
Cumulative Unreturned: 39.71568201 BTC
----
SD Profit on Completed Bets : 244.19040807 BTC
----
Since Satoshi Dice started, there have been:
Blockchain Tx: 1761683 : SatoshiDice Tx: 1136393 (64.5%)
Blockchain MB: 746.3 : SatoshiDice Tx: 462.7 (62.0%)

dooglus

legendary

Activity: 2940

Merit: 1333

Quote from: bitclown on July 02, 2012, 04:48:05 AM

Quote from: dooglus on July 01, 2012, 02:14:43 PM

If it weren't for the big bets, SatoshiDice would be almost 4000 BTC up by now, but instead they're 500 BTC down.

Sounds like SatoshiDice is lazy and use big amounts for their statistics-skewing transactions masqueraded as real bets.

It occurred to me that it was their way of taking a profit. But if that's the case, why have they lowered the max bet and kept it low? It doesn't make any sense to be decreasing the bet sizes as the profits grow.

bitclown

full member

Activity: 185

Merit: 100

Quote from: dooglus on July 01, 2012, 02:14:43 PM

If it weren't for the big bets, SatoshiDice would be almost 4000 BTC up by now, but instead they're 500 BTC down.

Sounds like SatoshiDice is lazy and use big amounts for their statistics-skewing transactions masqueraded as real bets.

dooglus

legendary

Activity: 2940

Merit: 1333

Quote from: ZiggiStar on July 01, 2012, 04:38:29 PM

You mean Satoshi lose more money with big bets ? is the graph now right ?

He lose alote of money with big bets ans makes money with small bets ?
is this right so ?

Yes, that's what I mean. The graph is right, but a little out of date. Here's an up-to-date one;

Quote

Results: 2012-Jul-01 02:44pm (up to block 187092)

Address Target Should Win | #Bets | Win | Lose | Refunds | BTC In | BTC Out | Profit | RTP
-----------------------------------------------------------------------------------------------------------------------------------------
1dice1e6p 1 0.00002 | 6201 | 0 (0.00000) | 5943 | 256 | 21.36 | 0.01 | 21.34 | 0.090
1dice1Qf4 2 0.00003 | 882 | 0 (0.00000) | 814 | 67 | 7.12 | 0.00 | 7.12 | 0.025
1dice2pxm 4 0.00006 | 1382 | 0 (0.00000) | 1351 | 30 | 11.99 | 0.00 | 11.99 | 0.047
1dice2vQo 8 0.00012 | 1092 | 0 (0.00000) | 1055 | 36 | 15.05 | 0.00 | 15.05 | 0.031
1dice2WmR 16 0.00024 | 1349 | 0 (0.00000) | 1321 | 27 | 20.65 | 0.01 | 20.64 | 0.075
1dice2xkj 32 0.00049 | 3075 | 1 (0.00033) | 3062 | 11 | 89.10 | 100.38 | -11.27 | 112.650
1dice2zdo 64 0.00098 | 4438 | 5 (0.00113) | 4415 | 17 | 175.99 | 110.74 | 65.24 | 62.927
1dice37Ee 128 0.00195 | 5843 | 12 (0.00206) | 5803 | 26 | 1176.62 | 1142.07 | 34.54 | 97.064
1dice3jkp 256 0.00391 | 4369 | 20 (0.00459) | 4335 | 12 | 458.47 | 330.67 | 127.80 | 72.125
1dice4J1m 512 0.00781 | 5402 | 32 (0.00593) | 5364 | 5 | 723.43 | 330.50 | 392.93 | 45.685
1dice5wwE 1000 0.01526 | 6570 | 92 (0.01402) | 6472 | 2 | 1306.15 | 695.34 | 610.80 | 53.236
1dice61SN 1500 0.02289 | 5088 | 108 (0.02125) | 4974 | 6 | 1598.74 | 2134.22 | -535.47 | 133.494
1dice6DPt 2000 0.03052 | 5583 | 173 (0.03103) | 5403 | 3 | 1750.55 | 1619.07 | 131.48 | 92.489
1dice6gJg 3000 0.04578 | 5445 | 275 (0.05057) | 5163 | 6 | 3608.20 | 4704.79 | -1096.58 | 130.392
1dice6GV5 4000 0.06104 | 5515 | 340 (0.06171) | 5170 | 3 | 1602.86 | 1402.21 | 200.65 | 87.481
1dice6wBx 6000 0.09155 | 9358 | 879 (0.09400) | 8472 | 4 | 5961.81 | 6560.49 | -598.68 | 110.042
1dice6YgE 8000 0.12207 | 10325 | 1259 (0.12204) | 9057 | 0 | 2891.30 | 2498.25 | 393.04 | 86.406
1dice7EYz 12000 0.18311 | 14500 | 2768 (0.19109) | 11717 | 5 | 5905.06 | 6230.00 | -324.94 | 105.503
1dice7fUk 16000 0.24414 | 35833 | 8688 (0.24253) | 27134 | 6 | 8967.36 | 8620.49 | 346.86 | 96.132
1dice7W2A 24000 0.36621 | 23916 | 8841 (0.37015) | 15044 | 30 | 10783.37 | 11044.96 | -261.59 | 102.426
1dice8EMZ 32000 0.48828 | 270535 | 131680 (0.48700) | 138711 | 120 | 80495.15 | 82188.83 | -1693.67 | 102.104
1dice97EC 32768 0.50000 | 105990 | 52945 (0.49986) | 52975 | 61 | 36154.48 | 35516.80 | 637.68 | 98.236
1dice9wcM 48000 0.73242 | 66124 | 48556 (0.73470) | 17534 | 31 | 38033.84 | 36657.31 | 1376.52 | 96.381
1dice9wVt 64000 0.97656 | 5692 | 4598 (0.97457) | 120 | 973 | 4901.49 | 4709.70 | 191.79 | 96.087
-----------------------------------------------------------------------------------------------------------------------------------------
| 604507 |
-----------------------------------------------------------------------------------------------------------------------------------------

Cumulative Wagers: 210867.93922406 BTC
Cumulative Rewards: 210773.98925246 BTC
Cumulative Fees Paid: 304.16610000 BTC
Cumulative Unreturned: 30.63706275 BTC
----
SD Profit on Completed Bets : -240.85319115 BTC
----
Since Satoshi Dice started, there have been:
Blockchain Tx: 1737546 : SatoshiDice Tx: 1128216 (64.9%)
Blockchain MB: 736.9 : SatoshiDice Tx: 459.5 (62.4%)

ZiggiStar

full member

Activity: 210

Merit: 100

Quote from: dooglus on July 01, 2012, 02:14:43 PM

Here's an interesting graph:

It splits the profit/loss into that caused by bets under 4 BTC and that caused by bets of 4 BTC or more.

There's a marked difference. Small bets are very profitable. Big bets are very expensive.

If it weren't for the big bets, SatoshiDice would be almost 4000 BTC up by now, but instead they're 500 BTC down.

Everything was looking fine until the first week of June. Since then there's been a big difference between the profitability of "big" and "small" bets.

Edit: fixed the graph. the >= and < labels were the wrong way around

You mean Satoshi lose more money with big bets ? is the graph now right ?

He lose alote of money with big bets ans makes money with small bets ?
is this right so ?

dooglus

legendary

Activity: 2940

Merit: 1333

Here's an interesting graph:

It splits the profit/loss into that caused by bets under 4 BTC and that caused by bets of 4 BTC or more.

There's a marked difference. Small bets are very profitable. Big bets are very expensive.

If it weren't for the big bets, SatoshiDice would be almost 4000 BTC up by now, but instead they're 500 BTC down.

Everything was looking fine until the first week of June. Since then there's been a big difference between the profitability of "big" and "small" bets.

Edit: fixed the graph. the >= and < labels were the wrong way around

byronbb

legendary

Activity: 1414

Merit: 1000

HODL OR DIE

Reminds me of my poker playing days where you can only ever lose. But I was never good enough to run this badly:

dooglus

legendary

Activity: 2940

Merit: 1333

Here are updated stats, along with a graph showing profit over time, along with expected profits given 1.0% and 1.5% house edges.

Quote

Results: 2012-Jun-29 12:47pm

Address Target Should Win | #Bets | Win | Lose | Refunds | BTC In | BTC Out | Profit | RTP%
-----------------------------------------------------------------------------------------------------------------------------------------
1dice1e6p 1 0.00002 | 5968 | 0 (0.00000) | 5716 | 251 | 20.07 | 0.01 | 20.05 | 0.096
1dice1Qf4 2 0.00003 | 865 | 0 (0.00000) | 798 | 67 | 7.06 | 0.00 | 7.06 | 0.025
1dice2pxm 4 0.00006 | 1372 | 0 (0.00000) | 1342 | 30 | 11.93 | 0.00 | 11.92 | 0.047
1dice2vQo 8 0.00012 | 1087 | 0 (0.00000) | 1051 | 36 | 14.93 | 0.00 | 14.93 | 0.032
1dice2WmR 16 0.00024 | 1335 | 0 (0.00000) | 1308 | 27 | 20.23 | 0.01 | 20.21 | 0.075
1dice2xkj 32 0.00049 | 3067 | 1 (0.00033) | 3055 | 11 | 88.65 | 100.38 | -11.72 | 113.224
1dice2zdo 64 0.00098 | 4425 | 5 (0.00113) | 4403 | 17 | 175.83 | 110.74 | 65.08 | 62.983
1dice37Ee 128 0.00195 | 5739 | 12 (0.00210) | 5702 | 25 | 1087.25 | 1141.67 | -54.42 | 105.006
1dice3jkp 256 0.00391 | 4352 | 19 (0.00438) | 4321 | 12 | 458.06 | 256.56 | 201.50 | 56.010
1dice4J1m 512 0.00781 | 5285 | 32 (0.00606) | 5248 | 5 | 701.77 | 330.41 | 371.35 | 47.083
1dice5wwE 1000 0.01526 | 6367 | 87 (0.01367) | 6277 | 2 | 1257.79 | 655.93 | 601.85 | 52.150
1dice61SN 1500 0.02289 | 5019 | 107 (0.02134) | 4907 | 5 | 1380.24 | 2091.00 | -710.75 | 151.495
1dice6DPt 2000 0.03052 | 5498 | 171 (0.03112) | 5324 | 3 | 1595.74 | 1459.85 | 135.88 | 91.484
1dice6gJg 3000 0.04578 | 5394 | 272 (0.05048) | 5116 | 6 | 3476.60 | 4491.24 | -1014.64 | 129.185
1dice6GV5 4000 0.06104 | 5378 | 333 (0.06195) | 5042 | 3 | 1578.36 | 1378.16 | 200.19 | 87.316
1dice6wBx 6000 0.09155 | 9136 | 856 (0.09373) | 8277 | 3 | 5842.87 | 6421.04 | -578.17 | 109.895
1dice6YgE 8000 0.12207 | 9609 | 1166 (0.12134) | 8443 | 0 | 2813.96 | 2430.33 | 383.62 | 86.367
1dice7EYz 12000 0.18311 | 14200 | 2709 (0.19088) | 11483 | 5 | 5857.29 | 6191.28 | -333.98 | 105.702
1dice7fUk 16000 0.24414 | 35530 | 8609 (0.24235) | 26914 | 5 | 8893.77 | 8541.12 | 352.65 | 96.035
1dice7W2A 24000 0.36621 | 23601 | 8726 (0.37022) | 14844 | 30 | 10701.69 | 10988.45 | -286.76 | 102.680
1dice8EMZ 32000 0.48828 | 266842 | 129874 (0.48693) | 136844 | 117 | 80094.30 | 81849.88 | -1755.57 | 102.192
1dice97EC 32768 0.50000 | 105161 | 52539 (0.49990) | 52561 | 60 | 35780.42 | 35176.72 | 603.69 | 98.313
1dice9wcM 48000 0.73242 | 65717 | 48254 (0.73463) | 17431 | 31 | 37953.57 | 36577.94 | 1375.63 | 96.375
1dice9wVt 64000 0.97656 | 5687 | 4598 (0.97457) | 120 | 968 | 4901.49 | 4709.70 | 191.79 | 96.087
-----------------------------------------------------------------------------------------------------------------------------------------
| 596634 |
-----------------------------------------------------------------------------------------------------------------------------------------

Cumulative Wagers: 208853.68253522 BTC
Cumulative Rewards: 209038.94171709 BTC
Cumulative Fees Paid: 300.23065000 BTC
Cumulative Unreturned: 3.31029257 BTC
----
SD Profit on Completed Bets : -488.80012444 BTC
----
Since Satoshi Dice started, there have been:
Blockchain Tx: 1692649 : SatoshiDice Tx: 1113632 (65.8%)
Blockchain MB: 718.9 : SatoshiDice Tx: 453.4 (63.1%)

dooglus

legendary

Activity: 2940

Merit: 1333

Quote from: unk1911 on June 27, 2012, 07:38:41 PM

i have noticed something off regarding SatoshiDice. on june 25th, it appears they made some modifications to the system, and it appears that if i add up all of the bets that i have sent to the following address:
1dice7fUkz5h4z2wPc1wLMPWgB5mDwKDx

that the payout odds coming back, were on average 1:7.91 instead of the approximately 1:4 that is advertised on the front of the website.

How many bets have you made to that address since June 25th, and how many won?

SatoshiDice is provably honest - they use a publicly known mathematical function to decide whether you won or not, using secret numbers decided before the game started. A hash of the secret numbers was also published before the game started, so they can't change it.

Looking at the blockchain since June 25th (I started counting at block 186101), I can see that they have processed 3808 bets to that address. 1 was refunded for being over the max bet, 3 haven't yet been paid out either way, and of the remaining 3804, 885 won. That's 23.26% of them that won (1:4.298). We would expect 24.41% to win in the long run (1:4.096), so that's pretty close. It's certainly nowhere near the 1:7.91 that you saw.

unk1911

newbie

Activity: 15

Merit: 0

i have noticed something off regarding SatoshiDice. on june 25th, it appears they made some modifications to the system, and it appears that if i add up all of the bets that i have sent to the following address:
1dice7fUkz5h4z2wPc1wLMPWgB5mDwKDx

that the payout odds coming back, were on average 1:7.91 instead of the approximately 1:4 that is advertised on the front of the website. of course, the payout had stayed at 4x. something seems to be broken and maybe they need to look into this. of course i'm a little bit pissed off because i have already lost a lot of money before i realized this was going on. i could sense that something was off but by the time i actually sat down and did the counting of # of wins vs # of losses i didn't realize that the odds were stacked up against me at approx 1:8...

dooglus

legendary

Activity: 2940

Merit: 1333

Quote from: Stephen Gornick on June 25, 2012, 04:27:15 PM

How about the potential that there are multiple wagers with the same transaction hash? That would appear in the results.

That's an interesting attack. I'll check and see if there are any duplicate transaction IDs with SatoshiDice bets in them.

I looked at just the bets during the period of heaviest losses for SatoshiDice, the 800 blocks 182800 through 183599, and found the following. My theory of it being double-spent longshot bets is clearly wrong. The vast majority of the losses were from "under 32000" bets - pretty much coin tosses. "under 4000" had the highest RTP, but much smaller losses, due to smaller volume of bets:

Quote

Address Target Should Win | #Bets | Win | Lose | In | Out | Profit | RTP
-----------------------------------------------------------------------------------------------------------
1dice1e6p 1 0.00002 | 51 | 0 | 37 | 0.20 | 0.00 | 0.20 | 0.0%
1dice1Qf4 2 0.00003 | 32 | 0 | 29 | 0.27 | 0.00 | 0.27 | 0.0%
1dice2pxm 4 0.00006 | 20 | 0 | 17 | 0.26 | 0.00 | 0.26 | 0.0%
1dice2vQo 8 0.00012 | 47 | 0 | 44 | 0.96 | 0.00 | 0.96 | 0.0%
1dice2WmR 16 0.00024 | 26 | 0 | 25 | 1.18 | 0.00 | 1.18 | 0.0%
1dice2xkj 32 0.00049 | 20 | 0 | 20 | 1.48 | 0.00 | 1.48 | 0.1%
1dice2zdo 64 0.00098 | 49 | 0 | 47 | 10.43 | 0.03 | 10.40 | 0.3%
1dice37Ee 128 0.00195 | 506 | 0 | 502 | 42.46 | 0.11 | 42.35 | 0.3%
1dice3jkp 256 0.00391 | 115 | 1 | 111 | 19.39 | 10.06 | 9.33 | 51.9%
1dice4J1m 512 0.00781 | 107 | 0 | 106 | 34.24 | 0.14 | 34.09 | 0.4%
1dice5wwE 1000 0.01526 | 222 | 2 | 215 | 120.98 | 1.27 | 119.71 | 1.1%
1dice61SN 1500 0.02289 | 37 | 1 | 35 | 56.35 | 0.31 | 56.03 | 0.6%
1dice6DPt 2000 0.03052 | 94 | 2 | 92 | 56.92 | 1.25 | 55.66 | 2.2%
1dice6gJg 3000 0.04578 | 105 | 3 | 101 | 83.61 | 1.26 | 82.34 | 1.5%
1dice6GV5 4000 0.06104 | 135 | 12 | 123 | 94.54 | 224.38 | -129.84 | 237.3%
1dice6wBx 6000 0.09155 | 232 | 20 | 211 | 851.87 | 1168.98 | -317.11 | 137.2%
1dice6YgE 8000 0.12207 | 384 | 37 | 340 | 186.28 | 112.71 | 73.57 | 60.5%
1dice7EYz 12000 0.18311 | 4309 | 819 | 3488 | 626.54 | 861.54 | -235.00 | 137.5%
1dice7fUk 16000 0.24414 | 3124 | 793 | 2322 | 427.20 | 467.04 | -39.83 | 109.3%
1dice7W2A 24000 0.36621 | 1342 | 501 | 824 | 1483.96 | 1635.54 | -151.58 | 110.2%
1dice8EMZ 32000 0.48828 | 50458 | 24498 | 25726 | 24846.77 | 26874.90 | -2028.12 | 108.2%
1dice97EC 32768 0.50000 | 3569 | 1783 | 1754 | 2187.61 | 2476.62 | -289.00 | 113.2%
1dice9wcM 48000 0.73242 | 4849 | 3571 | 1251 | 4339.44 | 4312.58 | 26.86 | 99.4%
1dice9wVt 64000 0.97656 | 283 | 263 | 10 | 294.68 | 269.88 | 24.79 | 91.6%
-----------------------------------------------------------------------------------------------------------

(I'm ignoring bets which weren't paid out within the 800 blocks in question, though they count in the 'bets' column, and am also ignoring fees paid)

nibor

sr. member

Activity: 438

Merit: 291

Only defence to premined blocks seems to be to ensure payout is never more that 50BTC then there is no point withholding the blocks when you do win.

Stephen Gornick

legendary

Activity: 2506

Merit: 1010

Quote from: dooglus on June 25, 2012, 03:04:00 PM

More likely is that they could double-spend their losing bets using pre-mined blocks, and only let their winning bets confirm.

How about the potential that there are multiple wagers with the same transaction hash? That would appear in the results.

- https://bitcointalksearch.org/topic/m.883119

dooglus

legendary

Activity: 2940

Merit: 1333

Quote from: etotheipi on June 25, 2012, 03:10:31 PM

Even with millions of bets there should be no way to reverse-engineer the secret from the HMAC operation, especially knowing only the last two bytes. (is the full output of HMAC(secret,TxID) shown?)

sha256(secret) has been published for all the secrets that will be used in the next 10 years, and the full output of HMAC is shown. See this 'recent bet':

Quote

Secret and Transaction Hash

Verify that the hmac sha512 with secret and transaction_id hash to the bet hash
hmac_sha512(hidden,69468dfc1a10bc9ae98d6913a715436eb1bd795bf023ea2aa0ff1d5f07f22b97) -> 1695547cca9f140b91a127664c43369d8ec9a70cc3c509d2c25f179a4d4c8dc3

etotheipi

legendary

Activity: 1428

Merit: 1093

Core Armory Developer

Quote from: apetersson on June 25, 2012, 02:55:27 PM

Quote from: etotheipi on June 25, 2012, 02:19:52 PM

My guess is that someone found a vulnerability

how likely is it someone could brute-force the daily secret hashed and published in advance?

In a perfect world, the methodology is secure, and this would be extremely difficult. Even with millions of bets there should be no way to reverse-engineer the secret from the HMAC operation, especially knowing only the last two bytes. (is the full output of HMAC(secret,TxID) shown?)

It's my understanding that there has not been a lot of confirmed double-spend [attempts]. I could be wrong though.

But the world isn't perfect. The machine holding the secrets and doing the calculations must be connected to the internet. Maybe someone found a way to get the server to leak information. I'm sure interested peoples have isolated a "direct" connection to this node, and even if they don't "hack" it, may be able to get useful information out of communicating with it.

dooglus

legendary

Activity: 2940

Merit: 1333

Quote from: apetersson on June 25, 2012, 02:55:27 PM

Quote from: etotheipi on June 25, 2012, 02:19:52 PM

My guess is that someone found a vulnerability

how likely is it someone could brute-force the daily secret hashed and published in advance?

Not very. More likely is that they could double-spend their losing bets using pre-mined blocks, and only let their winning bets confirm.

If I control a lot of hashpower, I make sure that every block I mine contains a spend to myself. Each time I find a block, I double-spend that transaction in the form of a "less than 512" bet at SatoshiDice. I wait a few seconds to see if I win. 99% of the time I'll lose, but if I win, I throw away the block I mined and let my winning bet confirm, otherwise I submit it to the network as normal, causing the lost bet to never be confirmed. By waiting to see if I win the bet, I risk my block being orphaned, but my expected win is enough to cover that risk so long I can bet high enough at SatoshiDice.

Is there any flaw in that system? Have any steps been taken to stop it happening today?

apetersson

hero member

Activity: 668

Merit: 501

Quote from: etotheipi on June 25, 2012, 02:19:52 PM

My guess is that someone found a vulnerability

how likely is it someone could brute-force the daily secret hashed and published in advance?

DeepBit

donator

Activity: 532

Merit: 501

We have cookies

Quote from: byronbb on June 25, 2012, 10:28:36 AM

This satoshidice guy is pretty unlucky.

Depends on what his goal is.

etotheipi

legendary

Activity: 1428

Merit: 1093

Core Armory Developer

My guess is that someone found a vulnerability and exploited it for 1000 blocks, and is waiting for the site to "refill" before doing it again. The sequence of losses is just way too extreme over that 182k-184k region to be luck, unless it's based mostly on long-shot bets. But from the previous postings, there haven't been a ton of long-shot bets..

Or the vulnerability was found and fixed. I'm not really sure. But as an attacker, it might be more beneficial to let the site recover and skim money instead. So you get the benefits of running a gambling website without actually having to run one!

dooglus

legendary

Activity: 2940

Merit: 1333

Here's the same thing but with dates instead of block numbers on the X axis:

Topic: Satoshi Dice -- Statistical Analysis - page 77. (Read 192918 times)