As a suggestion to the OP; try increasing the concentration of monsters and shards and stuff when creating the map; things will be closer together, and there will be more things for your heroes to do, before you and the AI players are close together and in a position to war on each other.

Do the factions AI's behave differently?

If so, is it possible to adjust custom factions behaviour? (eff you spellcheck, behavious is a perfectly cromulent word)

It's kind of interesting you should say this. I Slash and Burn, EVERYTHING. Partly to force the AI back without leaving exposed cities in my wake (that they could theoretically reclaim with a mere Pioneer), partly because the 50% unrest penalty on taking a city from an enemy can be insanely prohibitive, especially if you haven't gotten your Fort cities up to level 4/5 yet and base unrest in your current cities is at, near, or above 50%.

On top of that, I find the AI tends to make fairly poor location choices (or at least, poor compared to a human strategist), which often leads me to want to re-settle their former lands anyway.

I'm at the tactical point where my city type decision tree is as follows:

If tile has 2+ essence, Conclave.

If tile has 4+ food, Fort.

If tile has 2 or 3 food, Town.

Ignore tiles with 1 food, unless grain is in reach.

Prioritize tiles with access to rivers for Forts and Conclaves, access to forests for Towns. Access to both is epic and trumps all other concerns except in the case of:

I need more mana/research production, ergo, prioritize Conclaves.

or

I need more unrest reduction, ergo, prioritize Forts.

In all cases, two lesser build sites are preferable to one awesome site, except where a 3 essence site can be garnered, which quite literally trumps ALL other concerns.

It kind of sounds like much of the complaint is that the bots always "play to win", and some people want to play in a more RP setting, where at least most of the bots have some kind of desire for peace, rather than being constantly focused on achieving one of the win parameters.

Currently, if you're a small country with little army, bots get so mad at you for it that they're willing to commit suicide by sending most of their army across the map at you when they have incoming armies from larger countries headed their way.

I also wish bots understood when they were losing a war, not just power score. It would be nice if after losing 10 battles and 3 cities to me, the bot would say "okay, this war is not at all profitable, lets ask for peace", and then if I accepted, they didn't declare war again for a very long time. However, currently they just keep smashing their head into the wall until my power score is above theirs.

I have managed to make one of the two AI's into a friend. This is very unusual for one of my games, and the only thing I can put it down to is having the world difficulty on a higher level than the AI's.  I now often see the AI's scouts getting squashed by monsters.

Just waiting to see what the second AI is going to do.  I'm fully expecting it to go through the same old routine of asking for money then WAR.  It would be refreshing if it did something else as did my first friendly AI.

I would dearly love to be able to get on top of the AI and punish it for the many games it has ruined for me.

I decided to play my current game on easy/easy, and though I haven't officially made contact with any of the 7 AIs yet, I did notice that Altar conquered the Shroud city before turn 50.  Surprising, since on challenging/challenging the AI took over 150 turns to start to truly expand.  This should be interesting.

This is a really good AI thread. I hope the title doesn't prevent Frogboy from reading it. It has just the kind of feedback he likes to hear (as far as I can tell).

Happened during the FE beta, too.  Problem is, most testers never play the lower difficulty levels, so they end up being way out of whack with what they should be.  That's why threads like this are important; casual/new players exist too, and the worst thing SD could do is have an AI that's so tough on low levels that those players are never given a chance.

Like I said before, if you are not that good and still have problems on easy, then maybe you should try one of the lower difficulty-levels. However, if you really don't want to do that, then maybe you should take a look at some of the Let's Plays out there. The ones from Eddy and Das123 are pretty good. They are based on 0.50, so certain things have changed since then, but they should still provide enough information for you to improve your game.

Yes, two in fact.

As I remember I use to play FE on beginner, but that was a little too easy, the monsters were pussy's.

I have been watching Das123 lets play videos, but they are for FE.  I didn't know he had made some for LH, when I done a search that came up blank.

Now I know where the LH episodes are I will be watching those.  Thanks for the link, I like his videos, I watch them in bed on my tablet.

@lelovelady

We are told that the AI doesn't cheat, but I can't imagine how it can expand at the rate it does without cheating.  I have 2 spear-men and one and a half towns while the AI owns half of the map and fields legions.

One of the problems is the sheer amount of luck involved at the beginning.

Either you get very lucky, with several good city spots in close proximity, with weak opponents you can easily take out with basic units you start with, letting you get a fast start to developing your realm.

Or you don't get lucky, have only one shitty spot to settle, are surrounded by enemies ranging between strong and epic, and have your capital razed by an angry dragon in the first 5 rounds.

Those are, naturally, rather extreme ends of the starting spectrum, but the fact that those are both things that happen (latter more oftenthan the former in my case) is an issue on random maps.

I play on normal, and i don't use cheats (mainly because the "give 1000 everything" cheat breaks the cities so i got 1k populace, and can't get past level 2, and not allowed to pick specialization), and most games end with me getting my ass handed to me (if i don't just give up soon after starting), but my current one has me 8 hours in with power score slightly above other 3 factions combined, 3 armies roaming around exterminating monsters (main army beiing my level 17 sovereign, 6 level 10 warg riding pikemen, and 2 fell dragons), i got 8 cities and high taxation (unrest in cities ranging between 0% and 19%), got all the towers and the forge done, have all other factions at war with each other (because i bribed them to declare war), and only reason i haven't won yet is because i'm too busy exploring shit to actually go and cast the spell, or exterminate the other factions.

And all because i got a sweet starting spot with the option of immediately starting 2 other cities, with only few weak monsters to take care of to secure them.

It is likely that whenever they release a new patch they will overwrite that file, so you may have to change those 3 numbers again later to get the same effect. It is also possible, though I think it would be bad form of them to do so in most cases, that they may have hardcoded some automatic triggers for war. I know there is an event that causes all civs to declare war on each other, and that number will almost assuredly have no effect on that (though it may make peace treaties easier to get). I have never personally had that event, so it may be pretty rare.

If I were to guess how the war tag works, I would say it adds up all the numbers in the diplomacy history (like -2 for military history), and multiplies them by the priority. The faction rating may be part of the equation too (as in the ratio of there power to yours). Maybe the formula is (dipNum1 + dipNum2 ... + dipNumN) * (their rating/ your rating) * warPriority (I just included this for fun speculation). This would make it a negative number in all cases where war was a possibility (by setting warPriority to 0 the number will always be 0), and if that number is more negative than a certain hardcoded threshold there is a random chance every turn for a war declaration. The reason for making a war priority would be to alter the behavior in data, and not in the code where it is run. That is just speculation though.