AI overland stack/unit movement strategy

[Seravy]

Other ships, yes.
By recklessly, I mean to go for the usual "equal force", but disregarding the "need at least 5 units for the attack to be effective" modifier. I don't mean to attack 7 ships using 2, but 1 vs 1 should happen instead of having to use at least 3 against 1.
It definitely is effective - either the player is forced to waste mana on these battles, or they lose the ship because the AI won't hesitate to do so. And keeping all ships off the seas is a very powerful strategy for the AI.

The big difference is, ships usually are durable enough to survive the first turn even if the player casts a spell - which is often not the case for land units such as swordsmen or halberdiers. There also aren't any lairs where the stack size difference is a major problem for the AI. So the "avoid using small stacks" tactic is unnecessary on sea.

[Nelphine]

Oh I see. That's fair. But while all those points are true, I would rather see doomstscks acting appropriately rather than naval units. If you can fix it without affecting doomstscks, then yes I agree, but if intercontinental units will have to follow the rule, I'd rather not.

[Seravy]

Already fixed by a simple workaround : The naval procedure simply multiplies stack value back for small stacks to undo the modifier. In theory, the ship stacks cannot split or move and cannot contain nonship units, so their size should not change before the matrix calculation and the attack decision.

[Arnuz]

Leaving less units in garrison for more aggressive personalities I like, mostly because it makes even more extreme the t40 cheating advantage problem and should basically force you to fix it Seravy

Considering we already concluded the cons outweigth the pros on changing the early game, forcing that direction anyway might not be the best thing to do

It's not like we can force anything, even the tests don't hold much weight apparently, so even more extreme t40 conditions despite not being the best might be the only way left



Also quoting this good point left hidden because of last post:

We concluded increasing the production while reducing spellcssting was a bad idea.

We didn't conclude overall cheating shouldn't be reduced.