If a fortress could shift its walls with a single thought, the first thing I'd notice is that the static advantage of a fixed perimeter evaporates. Every attacker would have to guess not just where the walls are, but when they'll move. That turns defense into a game of timing and prediction. You'd need a dedicated unit to monitor the fortress's own motion patterns, maybe using a rotating camera array or a network of motion sensors that feed back in real time. Attackers would still be able to study the fortress's logic, but they'd have to do it quickly, or they'll hit a wall that’s already moved on to the next phase. It also forces you to think about redundancy; if one wall moves, you must have backup positions ready to fill the gap. In short, you shift from brute strength to a dance of anticipatory moves, and that dance is as much about outthinking the opponent as it is about outlasting them.

You’re right, the guards would have to be trained to anticipate the moves, or they’ll be running into a wall that just slid into their line of fire. It’s a risk that the fortress could end up shielding the wrong flank if its logic misfires. A mock siege is exactly the kind of test that reveals those blind spots before you let real troops live through it.

That’s the plan: drills until the guards and the walls move in lockstep. If either falls behind, the whole system collapses. It’s a zero‑margin trade‑off, but in war, the margin you lose is the cost of failure.