Mindy - Rules

[I like Serena]

Here are a few additional statements on bad AI and missing functionality. I only talk about battle AI and we should not mix in non-battle AI:
4. AI spell selection is suboptimal.
5. AI enemy target selection is poor.
6. AI wastes too much mana in battle for no reason: losing battles, winning battles, node battles (AI's spells are dispelled).
7. AI fights out losing battles, which wastes units (in addition to mana).
(snip)
14. AI does not use the move and shoot tactics.
(snip)

That's quite a lot of issues.
I agree that we should not mix non-battle AI.
Your statements seem to refer to areas for improvement and are rather general IMO.
Can we narrow them a bit down?
At this time I do not think we are ready to rewrite everything from scratch and I do not like to become stuck at chipping at a rock that is too large.
So I propose to single out issues in which AI can improve and make them more explicit.

So for 4: which spells are used suboptimal and under which circumstances?
For 5: same thing - when and where?
I propose to postpone 6 and 7 until AI is a lot better than it is now.

I'm jumping ahead and singling out 14 (move and shoot), which I think is a good one, since we can probably do something about it reasonably soon.
Can we zoom in on what AI does wrong exactly, and what it should do?
When should we move in and shoot, and when should we move away and shoot?

A variant is 14b (wait or move and attack).
This is to make suppression counters work for AI as much as they work for us.
We still need to formulate the proper strategy for it though.
Can we catch that in a set of simple rules?

[Asfex]

My conter-opinion: we must mix in-battle ai and overland. Reason:attack tactic and defence tactic is different.
And: use breath and first strike:ai newer see it. So ,hell hounds cost more than its seems to ai. Hell hound always can attack 1move unit first.
3. The battle is complex: for example: how many 1-move units need to catch one hell-hound?can 3move unit catch 4move unit?no one alone can write perfect ai.
Let it be 'personality'.

[FrancoK]

My conter-opinion: we must mix in-battle ai and overland. Reason:attack tactic and defence tactic is different.

We can fix the tactic (the battles) and then change the strategy (AI will not attack 4 hell hounds with the single spearmen, because it knows it will lose). What do you think?

And: use breath and first strike:ai never see it. So ,hell hounds cost more than its seems to ai. Hell hound always can attack 1move unit first.
3. The battle is complex: for example: how many 1-move units need to catch one hell-hound?can 3move unit catch 4move unit?no one alone can write perfect ai.
Let it be 'personality'.

This can be discussed in the AI movement discussion.

[Asfex]

We can fix the tactic (the battles) and then change the strategy (AI will not attack 4 hell hounds with the single spearmen, because it knows it will lose). What do you think?

I think, attack 4 hellhounds by one spearmen, then cast flamestrike is not always bad tactic.
So, tactic ai is depend from strategic and otherwise. Moreover, there is no reason to attack stonewall by ingeneers,if you have disrupt spell.

[Tiltowait]

If I cannot damage the enemy due to unit enchantments (flying, guardian wind, etc) then dispel. This is a last resort after trying all other methods (web, etc.)

How high strength a dispel? Well, that's a subject of its own, depending on the optimal balance between strength of the enchantment, dispel chance, available mana, mana income, casting skill, remaining casting skill in this battle reflecting the number of times dispel can be cast if this one fails, and so on.