You overstate your case, unless you wish to argue that Satoshi was not competent.
Considering some of the bugs that made it into his initial release, it's at least safe to say he's a better cryptographer than software engineer.
In any case, the reissue fraud is more or less independent of mutation. The way you protect yourself against double payment when reissuing is that you must spend at least some of the same coins so that only one transaction or the other can get mined. If you do this, no amount of mutation will result in funds loss (though it might confuse people!), if you don't do this then you can still have payments doubled up even with no mutation at all. (E.g. first payment delay, second authored, second gets confirmed, first gets rebroadcast and makes it in too).
It appears that Mt Gox was trying to function using "
dead reckoning" accounting by treating their internal database as canonical when in fact only the blockchain is canonical. This is incompetent accounting even in the fiat world - it's like going off your checkbook balance exclusively and never checking your actual bank account balance. (For everybody not born in ancient times, this is an analogy from the dark ages of the 1980s and 1990s).
That's what I meant by being incompetent. If one of your UTXOs gets spent by being included in a transaction that makes it into the blockchain, your wallet should notice this regardless of anything else you think should be happening. Theirs didn't - if it did they wouldn't have this problem.
This is in addition to other errors like not accounting for coin maturation and also their long, multi-year history of failing to respond to problems proactively when they are first pointed out as opposed to waiting for catastrophe to take action.
