Happy day! I've played a little and have liked what I've seen thus far. Thanks for working so hard on this game!!

No workshop

Super impressed with this build. And not a single crash tonight, yay!

0.85 is the beta 2.

The AI is still getting steam-rolled by me. Where are the dreaded multi-army groups attacking my civilization?

This is how my game went against a challenging AI playing Umber.

Turn 90 - umber declares war on me

Turn 130 - still not a single attacker from umber that has come into my territory.

Turn 150 - still no enemy in sight. I finish my stack of doom. It has 2 champions with bows, and 5 spearmen units with leather armor (5 members each). I go on the offensive.

Turn 165 - Game over. I won. The enemy did not kill a single unit from my stack. It was not halted by spellcasting in strategic mode.

The AI's ability to manage its gold seems to be completely ruining it. Why does it have 3 units standing around in every city? That's a horrible drain on gold from wages. So they up their taxes, and lose all research and production. The AI needs to forget about having stationary units in cities. Instead it needs to create armies that go around hunting monsters to level up. It needs to be able to detect and counter invasions from any enemy (monster or faction).

Just like E:wom, the AI is way too passive for such an aggressive game. If the stack of doom is the name of the game, the AI needs to be doing it. And with champions, army bonuses, spells and experience quadrupling tripling hitpoints from a few battles, the stack of doom IS the game. I'm not even kidding about that hitpoint thing by the way.

As long as there are no spells to counter troops in the early game, such as fireball or blizzard, the AI needs to put making stacks of doom above all other things in importance. Lose a city? Don't care. Lose ten resource nodes? Don't care. Just make that stack of doom, and start attacking with it. Nothing else works. Nothing else matters. And don't even get me started on how the entire production, tech and gildar system is set up to favour stacks of doom specifically. Okay, you got me started. Let's describe two possible options. They are not the only options, but I'm going to describe them.

Option 1) Keep 3 units in every city, draining away wages which means I need to lower my production and tech income by a lot.

Option 2) Keep 5 units in a stack of doom, do all the fighting with it. My tech and production stays very high  (taxes = low). My units gain all the experience so they get massive increases to their hitpoints and accuracy.

Which do you think is the best? And which is the AI doing?

Ps. The AI needs to put MASSIVE importance on researching "groups" with 5 men in it. It is the most overpowered tech in the early game.

At this point, you need to consider the direction of the AI and the direction of the gameplay mechanics. They are going in opposite directions. The AI is playing one game, and the player is playing with the mechanics as best he can. Those are two completely different things.

Continuing my post above. In the post above, I detailed how the AI could more appropriately use the game mechanics. In this post, I will be making suggestions for how to adapt the gameplay mechanics to the way the AI plays.

Suggestion 1) Wages for units stationed in cities are reduced by 75%. This creates a high cost of attack, but a low cost of defense. It allows turtling AI to maintain a reasonable army while not giving up all its tech and gildar income to do so.

Suggestion 2) Trained units gain 75%-100% less experience. This will prevent the player from accruing high level troops while the AI does nothing.

Suggestion 3) The Sovereign may cast a single spell in any battle his faction is involved in, inside his own domain - even if he is not physically in the battle. This will put the strongest magical character in the faction in every battle, metaphorically speaking.

Suggestion 4) Cities may not be razed in a single turn. They should take at least citylevel * 5 turns to raze. This will prevent blitzkrieg and scorched earth attacks from players.

Suggestion 5) Trained troops regenerate 0 health per turn outside the faction's domain. In addition, any army that is attacked in strategic mode has its movement reduced by 1 per attack. This allows the AI to whittle down enemy armies with magic and many of the poorly construed armies it is currently using.

Suggestion 6) Increase the strength of militia by giving them appropriate traits. 

Suggestion 7) All attacks dealt by a trained unit in tactical combat ALWAYS cause a counter-attack from any other trained unit. The counter-attack deals 75-100% of its damage. This fixes the issue of the AI always being at a disadvantage due to poor movement in tactical combat and never ever eVER EVER EVER getting the first hit in, which is the largest factor in deciding a battle between such units.

1. Great Idea, I absolutely love that. You can think of the 'In City' Wages as the regular pay,while the added wages troops in the field get are an active duty bonus, or whatever that's called in military speak.

2. Don't like this idea too much. Getting XP is fun, leveling up troops is fun - should get rid of that. Instead offer incentives/better ways for stationed troops to gain XP. After all troops train all the time while in barracks. 

3. I like that in principle - it's also very MoM-like. I'd change it, however, so that your sov only gets to cast spells while he is stationed in his mage tower (Capital City Improvement). He would be able to cast spells up to a certain amount of mana per battle, where the amount of mana depends on the mage tower level (research&building) and character stats. He would also gain XP for helping in battles this way (This way you get to level your Sov while he is at home, like a proper Ruler).

4. Great idea again, I definitely support this.

5. I don't like the 0 health regen - that's just unnatural - you could lower it though, and increase it by having healer-type units around (or supply wagons). The movement reduction is a good idea though. To make it easier, you could just decrease movement in enemy territory across the board (and add a skill that negates this reduction).

6. Of course, that should definitely be in.

7. So far I like that only certain units get counter attacks. However, I don't mind every unit getting ONE counterattack per turn, unless they have skills to increase this. As far as AI being at a disadvantage, I'm sure Brad can solve that through AI magic.

I know this won't get in, but I still like unique units being tied to their faction. Would be cool to have troops that can't be copied.

Yes, by getting a retaliation as default.

Maybe I'm misunderstanding you, but shouldn't this 0.86 feature be a first step in solving the issue in point 7? 

AI:

+ Tactical AI will look at who it can attack when it runs out of moves

If you want the sovereign to cast spells outside his/her domain they can imbue a Champion so that they can cast their spells through the Champion.  That would bring back the flavor of imbuing. It would of course be a maintained enchantment on the Champion and the Sovereign would have to be in a Mage Tower or what have you. 

That would also lead to an interesting choice:  Do you leave the sovereign home well protected and projecting power or send him/her out to gain XP and level up for more spells? 

It is a big adjustment to be playing against the AI again.  I have to forget that I used to be able to roam freely around a mostly empty map because the AIs all got double-killed.  0.77 was more RPG than strategy game as a result.  0.85 is a different game.

My first impression is that the monster AI seems to be much more aggressive against me than it is against the AI players.  I get chased when I get too close, the AI can build cities right beside deadly stacks and be more or less safe.  I did see an Ignys take out an outpost, but I think it just accidentally wandered over it.  Wildling armies sitting right beside AI Pioneers don't move a muscle.  Something ain't right.  This leads to a massive land grab early on that the AI wins because they don't have to defend what they claim.

I also found that the level of the free loot is still too high.  I have 6 weapons and I can only use 2 of them on my level 3 champions.  If I sold them I could run my kingdom without taxes for hundreds of turns.

Dislodging monsters from their lairs still lets you swoop in and loot it without effort.  Not fun.  I still think that only spawns should chase you.  Lairs should be permanently guarded by the best units they have.  And monsters with equal or even greater strategic speed would actually chase you.  The monsters are one of the most interesting parts of this game, but right now they are just too random and unintelligent.

AI players are still not clearing lairs completely.  They will attack the monsters when they can, but don't always take the loot.  I'm more than happy to take it for them, but they might be better off taking it themselves.

What would happen if monsters spawned patrols more frequently, but some types fought other types and leveled as a result, returning to their lairs to heal and reinforce?  Give it some thought.

Which is basically how it "exactly" works in Age of Wonders. If I were Stardock, I would take a really good look at Age of Wonders and see which game mechanics they used (I bet they did take a look but they should take a better look). Age of Wonders had 3 iterations and it got many many things right. To me, it is still the closest thing to Master of Magic and it's its spiritual successor so far.

I am not saying Stardock should copy every idea from Age of Wonders but there is a reason Age of Wonders was a successful franchise and there are good reasons why stuff (magic, game mechanics, etc.) works the way it works in Age of Wonders.