More importantly, the comparison is useless. The odds of running into issues with SHA-1 collisions in Git is a very different question from just the odds of two random SHA-1 hashes colliding.