

It is surely good to occasionally test whether unforeseen events will cause disruption. Do you know how often such a test is performed?
For example, nuclear weapons systems are tested occasionally, and seeing a failure is probably important information.
Giving moderators practice with moderating will probably improve their performance when they are actually needed.
How do you know this is a con?