- Joined
- Jul 18, 2017
I think the most frustrating thing about Paradox's AI is its inability to set proper priorities. This works in both directions. Its very easy to "Game" the AI, especially in crusader Kings 3. You might only have an army of 800 dudes against their combined alliance of 5,000. But you also know that once they start an 18 month siege of your capital you can swoop down with your 800 and start a 9 month siege of theirs. Now, you would think given the danger of this the first thing they will do is realize 9<18 and immediately pull off the siege of your capital and go after you. But no. They don't. So you siege down their capital, capture the leader, and boom, your 800 dudes just won a war against a massive coalition of great powers.
And in the opposite direction if you are trying an overly complicated start (like say Cornwall), you would think that after an 8 year long war with Wessex, with the war goal at 93% and all that is required is one last battle to 100% and take the entire petty Kingdom, your ally West Francia would prioritize that last decisive battle over an independence war just started by some minor noble in the south of France. But no. Immediately pulls off the war that is almost won and goes to fight some asshole with 350 troops an entire year away. Scuppering 8 years of work in a war that was all but over. (if that sounds oddly specific its why I am pounding my desk atm)
Its rather frustrating honestly. Paradox needs to hire a game theory mathematician to teach their AI how to prioritize importance on win conditions. It could be as simple as telling the damn thing to prioritize winning a war when its at 90%+ before haring off to deal with the one that just started.
And in the opposite direction if you are trying an overly complicated start (like say Cornwall), you would think that after an 8 year long war with Wessex, with the war goal at 93% and all that is required is one last battle to 100% and take the entire petty Kingdom, your ally West Francia would prioritize that last decisive battle over an independence war just started by some minor noble in the south of France. But no. Immediately pulls off the war that is almost won and goes to fight some asshole with 350 troops an entire year away. Scuppering 8 years of work in a war that was all but over. (if that sounds oddly specific its why I am pounding my desk atm)
Its rather frustrating honestly. Paradox needs to hire a game theory mathematician to teach their AI how to prioritize importance on win conditions. It could be as simple as telling the damn thing to prioritize winning a war when its at 90%+ before haring off to deal with the one that just started.
Last edited: