[WOM - Varied AI Strategies]

from Elemental Forums

A proposal for varied AI strategies.

At game start the kingdom/empire strategies for each AI fraction would be randomly selected. This strategy would give their research, build and expansion and diplomacy strategies for the entire game.

Example

  1. 01-15: Steamroller Strategy - Build one big strong army and proceed on a campaign of contest (15%)
  2. 16-30: Pillage, Burn, Destroy - Attack weakly guarded cities and burn them to the ground (15%)
  3. 31-45: Slow Expansion via Strongly Garrisoned Cities (15%)
  4. 46-80: Standard AI Strategy (35%)
  5. 81-90: Magic Strategy - All stat bonuses for sovereign go into Intelligence, focus on Magic Research (10%)
  6. 91-00: Magic & Archery Strategy - Research Archery and build bow units for an army that is 50% Archers  (10%)
8,736 views 7 replies
Reply #1 from Elemental Forums Top

I like where your going with this but I think one competent AI that adepts to the situation is a better idea. The sole exception being victory conditions, different AIs should go after different conditions. Like Magic, Conquest, Military, Diplomatic, or Quest.

What is a standard AI anyway?

Reply #2 from Elemental Forums Top

I would definitely like to see the AI use some measure of coherent strategy. If that means working on the AI as a whole, great. If it means designing separate packages like this, fine. I don't care as long as the AI plays smartly and provides a challenge.

 

I would like to see AI with personality, especially personality that is lore-friendly. This would bring more life to the game and the lore. Right now they all basically play the same. It would be nice is each Kingdom / Empire or Sovereign had a unique personality.

Offworld Trading Company
Reply #3 from Elemental Forums Top

Standard AI - is the generalist AI strategy that the current game's AI follow.

Agreed, I would like to see an AI with personality that is lore-friendly. For example - I would like to see the Wraiths tend to follow a magic focused strategy and the Trogs a melee focused conquest (50%) or pillage & burn strategy (50%).

Reply #4 from Elemental Forums Top

I don't want path dependent AI, this is a horrible idea. I want an AI that adapts to what I build, so that he can defeat me better. If I build archers, he should be building mounted heavy knights. If I spam spear units, he should be using AOE magic.

Personality is one thing, being locked into a strategy that the user can EASILY exploit is another.

Reply #5 from Elemental Forums Top

I disagree. An AI without a strategic focus is a lost AI. It will not know what to do. If faced with Fraction A fielding archers. Fraction B fielding spear untis and Fraction C pursuing a diplomatic strategy its research results will be minimal as it bounces back and forth researching a little in one area and a little in another area as its threatended by each fraction in turn.

Is it better to have an army that does everything in moderation or an army that excels in one area?

Example: I played a V1.39 where I put every Sovereign Stat point into Intelligence - that game me a sovereign with 45 points in Intelligence - enough for spell blast that does 15 points of damage AOE or Arcane Arrow 22 points of damage. That's an example of focus. I could have placed those points into Strength and achieved one heavy hitting damage dealing sovereign. If I had split the pointsbetween strength and intelligence the results would be much less effective and a leader that does not excel in one area.

I do agree that the AI needs to adapt to what their enemy does; however, it also needs a strategic focus. IMO the current AI is weak because it lacks this focus.

I agree that the AI needs to better counter the opponent's strategy. I never see the AI casting Wind Shield or Spell Immunity when faced by an archer army or my spell casting sovereign. If the AI is taking a melee focused approach the AI should research wind shield to protect his units from archers. If the AI is taking an archer approach he might be best served by researching spells that slow down attacking units (or cause them to lose their turn).

Reply #6 from Elemental Forums Top

I think we're discussing such flexible variables here that words just don't cut it. Obviously we agree that there should be some adaptability, but on the other hand we also want flavour. We need to come up with concrete examples of how much the AI should adapt, and how much they should stay the path, otherwise we will just get nowhere. For example, an AI produces 100 units. 20 are adapted to suit the enemy they are at war with and wish to destroy soonest, 20 are decided by the path chosen, and 60 are just chaff. Or should it be 80 adapted units, and 20 path-dependent units? Or 80 path-dependent units and 20 adapted units? I would personally aim for no more than 40 path-dependent units, simply because once the player sees that path, he can counter it way too easily, but it should still be enough to make them appear distinct. On the other hand, I wouldn't make the other 60 adapted units either, mainly because it will just force the player and the AI to build general jack-of-all-trades units which are fairly boring. So maybe 30 units should be adapted, and the last 30 chaff.

Reply #7 from Elemental Forums Top

Ai is the hardest part of a game to get right, and we do need concrete examples.

So what do you consider the opitmal priorities for the early game?

01) AI focused on Melee Combat with combat oriented sovereign
----------- Sovereign puts 100% stat points into Strength
----------- Arcane Research Priority:
02) AI focsued on Melee Combat with Magic Casting Sovereign
----------- Sovereign puts 100% stat points into Intelligence
----------- Arcane Research priority: Spell Blast
03) AI focused on building army of invulnerable Champions equippred with Legendary Armor and Weapons