An effective attack against SatoshiDice

dooglus

legendary

Activity: 2940

Merit: 1333

Quote from: paybitcoin on January 03, 2013, 02:16:47 AM

Meh, SD would just start keeping the small transactions or roll out some better fraud protection if it becomes an issue.

Or they could pay out each incoming transaction with a single outgoing transaction.

If you want to make 100 bets in a single transaction, you won't mind getting your payout in a single transaction either. If you want to look into exactly which bets won and which lost, you can check their website for details.

paybitcoin

member

Activity: 85

Merit: 10

1h79nc

Meh, SD would just start keeping the small transactions or roll out some better fraud protection if it becomes an issue.

Third Way

full member

Activity: 238

Merit: 100

Christ how horrifying. Dmanit bitcoin I just got my first five now this scary ass fuck news.

jl2012

legendary

Activity: 1792

Merit: 1111

I find an interesting transaction: http://blockchain.info/tx/9106ac6859097079d39127aaac86208ac2a2b9bade92c3ae109192b7bc340872

Paying 0.01BTC fee, it sent 228 x 0.000001BTC outputs to SatoshiDice. SD returned the fund (because they were lower than the minimum bet) with 228 transactions with 0.001BTC fee for each, with a total of 0.228BTC or about 3USD.

Comparing the loss of the attacker and SD, the ratio is 22.8x and seems quite effective. Actually there were 243 outputs in the attacking transaction (15 of them were not sent to SD) so the real ratio should be 24.3x. Since outputs do not contain public keys thus are small in size, I think 0.01BTC fee could actually pay for more than 243 outputs, making the attack even more effective.

So the question is, what could be the maximum harm done if the attacker paid 1BTC fee?

Topic: An effective attack against SatoshiDice (Read 1264 times)