Definitely have been cases where they have not been able to reverse transfers of this type, even for huge sums, where they know exactly where the money went ie. the bank and the specific accounts:

https://en.wikipedia.org/wiki/Bangladesh_Bank_robbery

Right, not having worked with it my knowledge of SWIFT is at a superficial level. So I actually followed Bangladesh heist because it's a fascinating at multiple levels.

The net loss was in fact minuscule (in % terms) compared to the attempted amount. Summarising from the article you linked.

> ....illegally transfer close to US$1 billion

> ..The Federal Reserve Bank of New York blocked..US$850 million

> ..with US$20 million traced to Sri Lanka...recovered

> ..traced..US$81 million to the Philippines...out of which US$18 million has been recovered.

So the actual loss is ~US$60 million. I would say the system has worked as intended, for sure there's a scope for improvement which I'm sure are being worked upon. Also I imagine the 4 individuals, the final recipients of the money are being pursued.

If I do a thought exercise with bitcoin; what are the potential options to prevent such a heist? I would imagine within an hour or so $1B would get distributed to to thousands of wallets. Sure, you could trace all of that, but then what? Would you instruct the Blockchain network to not authorise any spends originating from those addresses? Kind of maintain a black list?

There have been several proposals that would delay large crypto transactions by 24 hours or require them to be approved by multiple people. In no case are transactions reversed, but the idea is to introduce enough friction that hacking/fraud won't be effective.