I'm not sure what point you're trying to make but in this context there is no difference. If you know someone is cheating, you prevent further cheating by banning them.
Now I'll ask: how do you detect someone wall hacking automatically? No human review and no false flags. Go!
> If you know someone is cheating, you prevent further cheating by banning them.
If you think it's statistically likely that someone might be cheating, but you're not sure, you can matchmake them with other people who might be cheating.
That seems flawed as you would punish people who are playing well. Statistics are great, but you'd inevitably match legitimate players with misfits, ruining their experience.
A prevention model would be like the xbox where technical measures are used to prevent user code. A detection model is server side and detects anomalies for bans.
This doesn't work well in real time games. The client needs to know another player is on the other side of that wall so it can
* Play sounds from their actions
* Actually be able to render them when either player comes around the corner without them obviously materializing out of thin air.
Now I'll ask: how do you detect someone wall hacking automatically? No human review and no false flags. Go!