Nick Szabo  🔑

Bitcoin finality isn’t binary. Anyone telling you six is sufficient in all cases is lying. It was a semi arbitrary number Satoshi offhandedly mentioned

For casual use 1 or 2 is fine. For large transactions the jury is still out - there is no widespread model. If I were receiving $1b I would probably wait for a few days worth of confs

@NickSzabo4 suggests to wait a few month for $1b.

A few days should be safe because after that point a reorg causes so much collateral damage that the attacker can't expect the network to converge on his version of the chain. The social layer would definitely reject it.

the attacker can control the size of the collateral damage by simply replaying transactions 1:1 that are not part of the attack.

I was assuming that already, but he can't replay coinbase transactions. These coins and their ancestors will definitely be destroyed, causing havoc in the network.

those can only be spend 100 blocks after mined. So for a 200 block reorg we are talking about 100*12.5 = 1250 BTC ~$12.500.000 I am aware that IN THEORY those missing UTXOs can block an unlimited amount of total tx. In practice this seems unlikely.

So you think Bitcoin nodes would accept a reorg more than a couple days deep? The recent Binance incident implies otherwise. There was immediate support, incl. from core devs, to reject that reorg which wasn't even a day deep and not coming from a hostile entity.