5 ms·
I'd like to know what Gabe meant by fraudulent too. The whole idea of crypto is that if you send it and it's confirmed you can't get it back, which makes fraud
by m348e912 5y ago
I'd like to know what Gabe meant by fraudulent too. The whole idea of crypto is that if you send it and it's confirmed you can't get it back, which makes fraud (on the payers side) nearly impossible.
You brought up the point about zero confirmations. Zero-conf itself is fairly reliable, it certainly doesn't account for 50% transactions being fraudulent. Bitcoin introduced replace by fee in 2016. In 2017, bitcoin transition fees were high, and there was large increase in the bitcoin mempool, which led to a backlog of stuck low-fee transactions.
Replace by fee adds a major loophole to zeroconf, low-fee, stuck transactions could be rebroadcast with a higher fee and sent somewhere else even after they were "received" but not yet "confirmed" by Steam. Users could sign up, pay with Bitcoin, steam sees the payment but doesn't want the user to wait 3 days for confirmation, so steam completes the signup. After that the user rebroadcasts the transaction back to themselves with a higher fee and quicker confirmation. The original transaction doesn't go through.
Here's and discussion on how reliable zero-conf is when RBF is not part of the protocol and instead bigger blocks are mined.
https://www.youtube.com/watch?v=TIt96gFh4vw https://www.youtube.com/watch?v=TIt96gFh4vw
- rosndo 5y agoI know this is about zero confirmation transactions because I know people who were deliberately double spending on Steam. > Zero-conf itself is fairly reliable, it certainly doesn't account for 50% transactions being fraudulent Of course it does. Someone using this to spend the same bitcoin multiple times will generate many times more transactions than a regular user.
- m348e912 5y agoYou’re right and I take it back. If Bitcoin didn’t have replace-by-fee, then maybe it would have been a more accurate statement.
- rosndo 5y agoYou can fairly reliably double spend regardless of replace-by-fee, you just need to coordinate which nodes you broadcast your transactions to.
- m348e912 5y agoI'd love it if you can share some evidence if that's as feasible as you think it is. I posted a link above that demonstrates otherwise.
- rosndo 5y agoAs this video makes clear it is possible to relatively safely accept unconfirmed non-rbf transactions, but that requires you to take extra measures like waiting for a few seconds.
- m348e912 5y agoYes, if you were going to send me $100 of non RBF-enabled crypto, I'd have no issue with waiting 3 seconds to feel reasonably sure the transaction would confirm.
- UncleMeat 5y ago> The whole idea of crypto is that if you send it and it's confirmed you can't get it back, which makes fraud (on the payers side) nearly impossible. I consider "obtained somebody's credentials or private key through illegal means and then used that to purchase a good with stolen BTC" to be fraud. That the network considers this to be a normal transaction is not relevant.
- leereeves 5y agoI wonder, is there any legal precedent for that? Analogy with other transactions could support either interpretation. (Spending stolen cash is not fraud, but using a stolen credit card is.)
- michaelmrose 5y agoCouldn't steam use longevity to decide whether to accept crypto transactions and cancel entire accounts for fraudulent activity? Don't let bob create a new account with zero games and double spend his way to a free game because he has nothing to lose. If Sam who has 100 games in his account with 10 year longevity cheats you insist he make good on his transaction or ban his account and don't ever let him create a new one with the same name and billing address. Sam is highly unlikely to value playing a singular game more than his account. This seems absolutely trivial to avoid problems with.