In some environments, thread creation can fail. The weakest link here is attempting to set the stack size for the crawler thread which will bail if the stack size doesn't meet various requirements. Debugging this without an error message requires far too much effort because the seeder will silently continue running even though no crawler thread is live.
Since the risk of running the seeder without any of the threads (except maybe stats) is too high, it is best to check all of them for errors.