That being said, I do think patio11 is technically correct, given his unconventionally strict definition of "finality".
Probability is at the core of Bitcoin's proof of work security mechanism. There is a non zero probability that I could rewrite the whole Bitcoin transaction history in the next 15 minutes, if my miner gets really, really, really lucky. The chances of that happening are of course abysmally small and not worth considering. For all practical purposes, Bitcoin transactions are final past a few confirmations.
I estimate the probability of a similar Bitcoin action to be small, but much larger than the "miner gets really, really, really lucky" scenario you propose. Do you disagree?
> I estimate the probability of a similar Bitcoin action to be small, but much larger than the "miner gets really, really, really lucky" scenario you propose. Do you disagree?
Of course, I agree. That being said, I'd still put the probability in the vanishingly small category, especially for any individual transaction that isn't an exploit, and not worth considering in the context of "finality", as the term is commonly used (and not as patio11 defines it).