Stupid things you have seen the enemies do?

[ping]

FE is not a symmetric game by design - player and AI have neither similar army composition, nor unit placement, nor victory conditions. The AI isn't meant to "win the game", it's just meant to put obstacles on the player's path, but it's also meant to make sure that the player has a way to overcome them. Imagine an AI that wouldn't fall for the ol' "place a unit in range to draw them out" ruse, which would be completely transparent to any human opponent. I've seen people complain about the "group aggro" mechanics in the DS games, where it's often impossible to pull individual units out of an enemy formation. Imagine a group of wyvern riders that will not only stay stationary until a unit is in range of all of them, but that will also make sure to stay just outside the players' range. In the same vein, the suicidal tendencies of the enemy are a good thing, as well, or at least better than the alternative of enemies that will always run away when their HP gets low. That would probably be a really annoying mechanism, wouldn't it? ;) @Raqueswiss' pic is another example where the smart (well, smarter than going for the 0% hitrate) AI behaviour doesn't really do the game any favours; those bloody pegasi running for the forts in Erik's FE7 map is another one.

TL;DR: An AI that plays smart by player standards would worsen the game by a lot.

[Julius Nepos]

The NES games are fun because you can often predict who the AI will go for (thanks, prioritization). I found that out in FE1 when I was afraid the enemy was going to kill my wounded Castor, but instead decided to attack Marth for no damage.
I gave Silque the speed ring in Gaiden and it was just so much fun seeing every single enemy attacking her with 0% hit.

33 minutes ago, ping said:

FE is not a symmetric game by design - player and AI have neither similar army composition, nor unit placement, nor victory conditions. The AI isn't meant to "win the game", it's just meant to put obstacles on the player's path, but it's also meant to make sure that the player has a way to overcome them. Imagine an AI that wouldn't fall for the ol' "place a unit in range to draw them out" ruse, which would be completely transparent to any human opponent. I've seen people complain about the "group aggro" mechanics in the DS games, where it's often impossible to pull individual units out of an enemy formation. Imagine a group of wyvern riders that will not only stay stationary until a unit is in range of all of them, but that will also make sure to stay just outside the players' range. In the same vein, the suicidal tendencies of the enemy are a good thing, as well, or at least better than the alternative of enemies that will always run away when their HP gets low. That would probably be a really annoying mechanism, wouldn't it? ;) @Raqueswiss' pic is another example where the smart (well, smarter than going for the 0% hitrate) AI behaviour doesn't really do the game any favours; those bloody pegasi running for the forts in Erik's FE7 map is another one.

TL;DR: An AI that plays smart by player standards would worsen the game by a lot.

This is very true. It depends on the game and on the map in question, but in a lot of maps the enemy is stronger (at least in numbers) than the player. It is also much easier for the AI to achieve a "victory" than the player because the player needs to think long-term, while the AI only has to think short-term. If I miss a bunch of villages in one chapter, that is a long-term loss for me and therefore a victory for the AI; if one of my characters dies, that's a long-term loss for me and therefore a victory for the AI. The AI doesn't have to plan any further than the chapter it's in: destroy that village, kill that character; the player, on the other hand, has to plan ahead for the rest of the game.

I'd still like to see them make the AI smarter in future games, but it also shouldn't be too smart.

[Hawkwing]

On 12/25/2017 at 5:14 AM, Claudius I said:

This is very true. It depends on the game and on the map in question, but in a lot of maps the enemy is stronger (at least in numbers) than the player. It is also much easier for the AI to achieve a "victory" than the player because the player needs to think long-term, while the AI only has to think short-term. If I miss a bunch of villages in one chapter, that is a long-term loss for me and therefore a victory for the AI; if one of my characters dies, that's a long-term loss for me and therefore a victory for the AI. The AI doesn't have to plan any further than the chapter it's in: destroy that village, kill that character; the player, on the other hand, has to plan ahead for the rest of the game.

I'd still like to see them make the AI smarter in future games, but it also shouldn't be too smart.

This is a very interesting insight, and it most certainly explains why the AI isn't hesitant to sacrifice its units just to spite the player.

I'm actually very interested now in seeing a Fire Emblem game with an AI that thinks in the long term, being more careful with it's units and having your foes actually level up and even promote if it has fought in enough battles. It would also be interesting if the player could actually LOSE the battle on a map without causing a game over, with the player having to recuperate and also think in the long term. I could see this working if IS expanded upon the concept and mechanics of the world map that was used in Gaiden/SoV.

[Marco]

I think this was in Path of Radiance. I had not properly determined where the enemy could attack, so when I moved a unit, I accidentally left a very weak and almost dead unit now open for attack. I was far into the chapter so I was very pissed but I passed the turn anyway to see what would happen. Instead of attacking the weak unit, it goes for my knight and just... *dink* 'no damage'.

I was just stunned, looking at the screen and blinking a few times. After a bit, I just said out loud "This guy is packing fish AI."