No that's clear.. if it leaves liquid, of course it's not confidential anymore.. it needs to be settled on chain after all... the question is... is it possible in any way to trace back from the receiving address on chain to the buy on kraken if time has been randomized and amount has been altered... my guess is no.
Discussion
The thing I can't stress enough is that only amounts are hidden. It is trivial to see sender and receiver in Liquid so can easily be traced in and followed the whole way out. Waiting and breaking it up does nothing to help this. It has the exact same traceability as doing a normal Bitcoin transaction.
You need to do one of the following:
1) Coinjoin/Whirlpool (Alice sent to either Bob, Charlie, or Dave)
2) Ring signatures + stealth addresses (Maybe Alice sent to [?])
3) Full ZKPs ([?] sent to [?])
These are the only things that would help resist or eliminate tracing. Liquid does none of this.
All someone has to do is follow your first transaction in, follow the multiple branches going thru and out of Liquid, add up the amounts that left and know you are extremely likely the same person.
Thanks for clarifying.. Seems like LN -> LN -> loop out will give better privacy. One could let a few sats accumulate on LN, then send to liquid for "semi" cold storage before sending to final CS onchain... LN -> LN -> LQ should sufficiently break the link imo.
PS: I whirlpooled what I bought KYC so far but I'm looking to find a cheaper way to do it with sufficient privacy as well.
WabiSabi coinjoins cost far less in mining fees and provide superior privacy compared to Whirlpool since there's no toxic change created and no common input ownership revealed. You can participate in WabiSabi coinjoins with BTCPay Server's coinjoin plugin, Trezor's coinjoin account, or Wasabi Wallet for desktop.
Yeah... no... Hard pass on Wasabi and Trezor's coinjoin feature. ;)
...why would you pay more in fees to get worse privacy using Whirlpool?
Do you mean worse obfuscation per mixing round? Doesn't that depend on how many rounds you Whirlpool? (remixing is free, Wasabi isn't)
This is why I find the bolted-on wallet-based "privacy" of bitcoin coinjoins tedious and constant drama tiring.
It's easier just to press send on a Monero transaction and gain superior privacy to either one with way cheaper fees. Larger anon set and no centralized coordinators involved either.
While the naive per round privacy expectations from Whirlpool are objectively smaller than WabiSabi (5 input minimum in Whirlpool compared to 150 input minimum in WabiSabi), that's just an implementation detail. There's no coordinator fee charged for remixing using zkSNACKs' WabiSabi coordinator either.
The advancement that makes the WabiSabi coinjoin protocol so much better is that you gain COMPLETE privacy on your coins, whereas Whirlpool links your transactions together due to common input ownership being revealed and toxic change being created during the premix transaction.
WabiSabi coinjoins gives the same privacy UX to Bitcoin as Monero does to Monero 🎠The only exception is if you are the largest whale in your coinjoin round - since Bitcoin does not have confidential transactions, a whale's coins might require multiple remixes before gaining full privacy.
You can read more from the mailing list post if you enjoy the technical details: https://lists.linuxfoundation.org/pipermail/bitcoin-dev/2022-April/020202.html
There is no coordinator fee charged but you still have mining fees for every round on WabiSabi, no?
Yea I know about the toxic change part of Whirlpool. Seems like a trade off for completely uniform indistinguishable input/output amounts.
While the decreased privacy for the being the largest amount in WabiSabi rounds is the trade off for increased flexibility of input/output amounts.
Yes, it's a very important detail that WabiSabi does not incentivize free Sybil attacks like Whirlpool does: In Whirlpool, the victims of Sybil attacks pay for the block space used by an attacker, who remixes for completely free. WabiSabi is more resilient to Sybil attacks because the attacker has to pay for the block space they use.
Fair point
Censored TXs in Wasabi should be reason enough not to use them.
That's like saying no one should use Nostr because a single relay censors notes. Anyone can run a WabiSabi coordinator just like anyone can run a Nostr relay, stop blaming the protocol for the actions of a single user of that protocol.