[kqr]

If you think of incidents as component failures, and the solution as increasing automation related to getting the faulty component back online again, you're under the old view of system failure. This view works for simpler systems.

More complex systems experience failures due to interactions between fully functioning components. The teams that made them didn't, for one reason or another, foresee that mode of interaction.

These are errors designed deeply into the system, and you can't automate recovery. You need to fix the problem at the cause.

Proper analysis is required, and if a game is what it takes to do that then why not? Additionally, it helps people learn to do that analysis on the fly. That is a crucial skill because those incidents are normal in complex systems. They will happen.