One flaw with Whirlpool was you could only mix 1 output at once. Which could be used for correlation.
An RNG that newer outputs five 9’s in a row, or always passes statistical tests, is also broken even though it may “seem” random
Discussion
I don't understand how that can be correlated.
Also, if you were able to have more than 1 input in a mix, you might be ably to sybil it an unmix it.
The one input per mix restriction was enforced client side.
That can be correlated to say that 5 UTXOs lack common ownership with each other.
If the coordinator could detect you mixing 2 UTXOs at once, it means the coordinator has deanonymized you
So who the hell would know if 80% of Whirlpool liquidity was controlled by one entity. Sure it might cost but once it’s in they can remix those forever and imagine the returns they could get…
You're right, that's definitely a possible attack vector (besides a malicious coordinator, who can just do that by only paying miner fees).
I never thought of that. But still don't understand how allowing more than 1 utxo being remixed per client would sove it... Sounds like it would just make it easier to attack
> That can be correlated to say that 5 UTXOs lack common ownership with each other.
That's still not very useful information to have. It's not exactly a secret that it's a zerolink coinjoin.
I dont think chain analysis knowing that they might have common ownership would make it that much better.
But how would you accomplish that in a decentralized way?
alright, so when coinjoin coordinator over nostr?
Communication layer is not exactly the issue thus far. Real problems are architectural
Something of. Using irc chat bots to coordinate is ridiculous. Using a centralized coordinator like whirlpool clearly has risks. Coordinating over a protocol layer like nostr seems like a good fit?
True, centralized coordinator doesn't fly anymore obviously so that has to be solved as well. Nostr would be awesome for it