the odds are sooooooooo small, they're so close to 0 that in reality you can assume key collision is impossible IF the private key was generated properly.
Basel:
Why the odds is close to 0?
And after creating 1.78689910E + 62 address i think there can be more than one private key that leads to the same address