Test games played

[Bahgtru]

Once again the AI struggles with mana costs a single turn into the war, while having off the chart army strength (like, 8 times mine) and tons of stacks.
Maybe we should lower the AI gold bonus (do they even need ANY gold bonus? They might not be as good as the human but they at least know to spend more on adamantium etc cities, and they already have a production bonus for generic production...) and raise the power bonus (but this risks all the extra turning into research and skill if they weren't at war, like the Myrran).

Again, before changing AI bonuses please make them progressive. AIs need to have more later in the game, while they have no problems at the start. 
If you raised mana as you mention from the start it'd lead to ridiculous situations for some AI wizard/retort combinations.

[Sapher]

Part 3, part 4 and part 5 (final) of lunatic walthrough with heroes is uploaded.

[Sapher]

New 5.43 lunatic walkthrough. part 1 and part 2.
In general my impression is that the game became a bit harder because of 3 wizards on myrran. But if played right winning is easy(not my case i still make a lot of mistakes).

[Seravy]

Glad to hear 3 wizards helped. I'll watch these two now.

[zitro1987]

Nice black + sorcery combo early in game - vertigo + weakness on hydra outside chaos node over and over till it works, then summon a couple of zombies. You get a free undead regenerating hydra that probably can be granted 'flight'

[Sapher]

Nice black + sorcery combo early in game - vertigo + weakness on hydra outside chaos node over and over till it works, then summon a couple of zombies. You get a free undead regenerating hydra that probably can be granted 'flight'

I thought about the same thing. But i imagined it like warp creature into 0resist then black sleep and zombie. 
Behemoth would also work. But in general that is just too unreliable. Massproducing draconian bowmen is much better. You get many free items, heroes, spells,mana/gold from that. And if during that mass production you see a lonely hydra - good. If not - whatever...you will gather enough gold to buy armoured guild in a troll town and buy lots of wartrolls to crush everyone. Better always have a plan B.

By the way parts 3-5 are uploaded.

[Sapher]

Finished uploading last walthrtough and also added a new troll walkthrough. It has 3 parts first is here.
I guess it was the easiest one as it has only 3 parts. Even werewolves had 4. So for now trolls might be the fastest way to kill lunatic AI. 
Didn't pay attention to lairs/nodes much as before wartrolls i couldn't conquer them and after wartrolls i didnt need to. It would be a waste of turns.

[Seravy]

So, after watching all four games, some thoughts :

-At first I was worried as Death was used 3 times in a row but in the last one you used Chaos/Sorcery so I guess early realm balance isn't that bad. (would be nice to see some Nature runs though, as that's the other "rush" realm besides Death. I always worry undead creation might be too much of an advantage for Death.)
-Myrran in all four. The first two still had low opponent count and trolls require it so can't say it has balance problems just for that. I hope I can finish testing Arcanus races this year and then start playing Myrran races again.
-Maybe regeneration is too good for werewolves now that it restores 2 health and they have 3 armor? The two changes made sense separately, but it certainly has quite a bit of synergy and makes those regenerated health points much more valuable than expected.
-A single troll swordsmen (or 2) seems to be as good as an entire army of other swordsmen/bowmen. Maybe it's too powerful? Not that it really matters in a game where you could do the same with a common spell. (cloak of fear, fire elemental, wraith form come into mind, but there is always the "hit and retreat" usable with any common damage spell too. Using a troll costs no mana to win though.)
-Regen restoring 2 health is too abusable. Yes, it "only" allows you to kill an entire stack you could otherwise kill with the same unit in 3-4 battles - but doing it this way is a lot faster and speed is the key to early game balance. (it also allows you to split up a stack and use fewer units for the same thing, again speeding up things a lot.)
So yes, trolls don't do anything they shouldn't be doing, but they do it way faster than they should.
Of course this only works that way if the enemy isn't actively damaging your unit, but most early game units are melee with 2 moves so they won't. Even ranged units seem to be no threat - Large Shield on swordsmen reduces the damage well enough for it to survive due to the regen and high health pool.
Maybe the ideal solution would be to make Regeneration an ability with a variable amount of effect (regen 0,1,2,3, etc) and make sure early game units don't have a high amount (0-1) while later units can have the higher numbers (2-3).
-Are there too much prisoners? I see plenty of them appear each game even though the games end very early.

[Bahgtru]

-Myrran in all four. The first two still had low opponent count and trolls require it so can't say it has balance problems just for that. I hope I can finish testing Arcanus races this year and then start playing Myrran races again.

Dwarves almost ok with lower minerals - my fail-starts in lunatic have raised in percentage. Draconians have however become more attractive due to increased relative importance of treasure, for the same reason. I agree with Sapher's observation that it's possible to go trolls by starting draconians, it's what I do as well. (except my trolls are undead, but that's another story)

-Maybe regeneration is too good for werewolves now that it restores 2 health and they have 3 armor? The two changes made sense separately, but it certainly has quite a bit of synergy and makes those regenerated health points much more valuable than expected.

Definitely OP. Especially with adamantium. Shouldn't have either. You like it but it makes them unbalanced.

-A single troll swordsmen (or 2) seems to be as good as an entire army of other swordsmen/bowmen. Maybe it's too powerful? Not that it really matters in a game where you could do the same with a common spell. (cloak of fear, fire elemental, wraith form come into mind, but there is always the "hit and retreat" usable with any common damage spell too. Using a troll costs no mana to win though.)

Trolls are horrendously slow to produce anything and if the swordsman phase gives them a slight advantage, they become obsolete fast.

Maybe the ideal solution would be to make Regeneration an ability with a variable amount of effect (regen 0,1,2,3, etc) and make sure early game units don't have a high amount (0-1) while later units can have the higher numbers (2-3).

Wouldn't change much... The damaged units usually are just escaping till the end of the game anyway. Rather, reduce armor on regenerating units, so they risk more at the beginning and make it a real possibility that the stack will be lost due to lost battle.
What was the reasoning behind giving armor to werewolves rather than just more HPs if needed?

-Are there too much prisoners? I see plenty of them appear each game even though the games end very early.

That varies a lot from game to game, I've had none (or just the stupid barbarian/dwarf) in the last 5 games.

[Seravy]

What was the reasoning behind giving armor to werewolves rather than just more HPs if needed?

If I remember well, it was strategic combat strength.