Python dict order
Python implements "hash randomization"; the hash of a string is different from run to run. This is done to avoid DOS attacks on the hash algorithm (i.e. an attacker may create malicious input which intentionally causes hash collisions). As a result, set and dict iteration order may change from run to run.
On an unrelated note: do the participants get feed-back on how well they did? I.e. if they passed all the tests, or which ones they failed?