The old/legacy AI in relation to our walls was way too easy. The new AI in combination with the new campaign large +upgraded armies is unnecessarily hard as the AI attacks an unwanted entrance even if it has zero chance of succeeding. In order to avoid the old wall cheesing or the new found wall cheesing with an armory behind the wall and make the AI’s behavior more logical, I propose this:
Simply put - if the AI army would be attacking our entrance point and detect that in its army’s range there is too strong defenses, it would disengage and look for another, less defended entrance. Each defensive building and unit would be worth of some defensive potential. The same each AI’s attacking army would calculate its offensive potential. It could be possible to make rules for the AI’s code logic to work like this.
Only if the AI’s offensive potential would be larger than the player’s defensive potential, it would attack that wall / position. Funneling would be possible, but only with building enough of defensive buildings in order to make the AI to use a different approach.
I know pure text is boring, so here are images of specific situations:
How would it go if we didn’t have any upgraded towers or walls?
And how would it go if we would have towers and walls more upgraded?
Summary of reasons for suggesting this:
- legacy AI vs walls was too easy
- new AI vs walls behaves way too aggressive unnecesarily, hard to defend multiple entrances as the walls are ineffective without the cheesing
- new AI can be cheesed with 1 armory/eco behind the wall - easy with Egyptians with free eco buildings = unbalanced
- just to be more realistic/logical, to more behave like human players (who would never choose the most defended entrance while knowing the whole attacking army will die, while there is much easier alternative at the other entrance. At least not without a large enough army at disposal.)