Hacker News Reader

Comment by paulsutter | Hacker News Reader

Believe in your lucky sunglasses all you want, but how much would you really stake on a tumbler?

"Such tumblers do a very poor job of hiding transactions, especially when dealing with a thousand or a hundred thousand bitcoins. The simplest approach is to simply look for common flow patterns."

- http://www.forbes.com/sites/valleyvoices/2015/01/15/how-pros...

View on HN

Interesting article, but isn't that survivor bias?

All the tumblers seem inept when the only detectable tumblers are inept. You don't see the competent or succesful tumblers.

There is also an inherent risk in trusting that the tumbler operator is not willing to cooperate with (or happens to be operated by) western governments. It doesn't seem like a reliably sustainable strategy, especially for any substantial amount of funds.

Even if they don't control it, wouldn't a heavy user be able to build a probabilistic model of where the coins went? If you are 99% of the volume on a tumbler, you know with near certainty that your adversary's coins are being swapped with your own. Paying tumbler fees or even making massive amounts of small-loss bets on a gambling site is well within the resources of a state actor. Bitcoin is small. Eventually real money has to be taken in or out of the system, which gives them a smaller set of individuals to target for investigation.

Right, so you mix multiple times. You use unlinkable wallets. Electrum and Multibit wallets in Whonix VMs work well. And you successively use different mixing services, so deanonymization would require collusion. Doing three successive mixes is prudent.

If a tumbler is used by 1 person, they get their own coins back. To work, it needs as many users as possible. Therefore a researcher doesn't scan transactions looking for "possible tumbling", they look for the tumblers, use them and follow those transactions.

I meant tumbler as in a person, not a repeat service. It would be pretty easy to detect mass tumbling services through pattern recognition.

Why, though? Once coins from two separate sources enter a common wallet, they become impossible to differentiate from each other.

Not true. Each piece of each coin is traceable. With good mixers, you don't get your own coin back. And you check for taint.

True. If you want to anonymize that much coin, you must break it down into numerous small flows.

Reply on news.ycombinator.com