It is about statistics. I don't think Ford would let that happen if it failed instantly just once. If the car was driven for hours from San Francisco to Tejon, then the connector is probably already warmed up, and then a high throttle climb for 3000 ft in 10 mins would weaken the connector a bit but won't fail right away. Then repeat this once a week and after 1 year the connector "may" fail in a single percent probability.
These kind of things cannot be tested just a couple times before a product is released. I think they just didn't add enough safety margin in either algorithm (limit power output if temperature is X) or simulation didn't have enough condition like gradual aging of non-limiting components (i.e. having plastic in power connector in addition to the battery, motor, inverter, coolant, gasket, etc). Some companies would add more safety margin and not release their products until they have enough of it, while others would just rush it out and start selling asap. I think this is why a $13k Corolla in 93 would only have 3 speed auto and distributor cap/rotor whereas a $10k Escort would have 4 speed auto and waste spark, but the corolla would end up lasting longer.
