I don't disagree with your overall point, but if you care about accuracy, you shouldn't also conflate an engineering goal with an estimation. I couldn't find th

I don't disagree with your overall point, but if you care about accuracy, you shouldn't also conflate an engineering goal with an estimation. I couldn't find the original estimation, but considering the unknowns we're talking about and the sunk costs, surely a 4-8x safety margin (1-2 years estimation) is not unreasonable. So we're at most talking about a 15-fold wrong estimation, not 50.