Hello everyone, I just wanted to summarise my thoughts about AI in the Return of Rome DLC. I really like this game and must say - I see a potential for that to be more fun with some improvements of AI. Sadly, there are areas which are not adequate for the year 2025. I listed those I’ve noticed below.
1.Factors related to territory - RoR AI does not understand that… :
1.a) it is better to steal gold/other sources from enemy’s or neutral territory first and conserve sources in its own territory. Especially gold requiring civilizations should behave differently to the important resource which decides if they win or fail.
1.b) how to build walls and gates to defend its own territory.
1.c) it is a waste to send its own villagers to the area, where I am attacking with my soldiers, so those villagers get killed immediately and buildings they tried to construct/repair are razed (which leads to the loss of food and wood).
1.d) it is a waste of population capacity when the AI builds ships on a map like the Continental, where it is possible (and much more probable!) to win without ships. Greater number of land units will just avoid the coast and outnumber & destroy enemy’s land units and buildings.
2.Factors related to warfare - RoR AI does NOT understand… :
2.a) how formations work
2.b) how counter units work
2.c) it’s preferred to not create units for which AI’s civilization does not have an upgrade (example: Macedonians create broad swordsmen.)
2.d) which buildings are important to defend and when to do so.
2.e) it should scout areas near its base to predict from which side I will attack and what units I will use.
2.f) it is a waste to send only one or few soldiers to defend its city while the numbers of my soldiers are much bigger. AI should try to lure me somewhere else and build up its soldier’s quantity in order to have some chance to win the fight.
2.e) it does not understand priorities: to build its own economy and defend it, to build its own army, to destroy the enemy’s army, to gain ability to rebuild army quickly, to destroy enemy’s economy, to take enemy’s resources. AI should not be lured so easily to any other area and keep its focus on the enemy’s army/production buildings/economy.
3.AI performs unrealistically well in these cases:
3.a) giving orders to its ranged units: every foot archer tries to lure my units in a different direction.
3.b) when every monk tries to convert a different unit.
In these cases, I just pause the game and give my units separate orders to mimic the AI performance. I don’t like this, as I see it as cheating, and it became a bad habit for me, which I should stop.
Basically I see the main improvement of AI in its economy, which now works much better than in the original Rise of Rome datadisc. Now, AI does not stop its economy because of any hidden bug or early attack. It is also difficult to make a rush on it, as it is growing faster. It can also recover from many attacks, unlike in the Rise of Rome. It’s also good that AI can surrender at correctly evaluated moments, as this makes finishing the game easier. I just expected a little bit more, as I guess this is not about not available tools for AI learning (it may learn from the human players too), but more about the money investment part… 