[Vote] (5-47) Unit Capture Consistency

Approval Vote for Proposal #47


  • Total voters
    53
  • Poll closed .
Status
Not open for further replies.

azum4roll

Lost the game
Joined
Jul 17, 2018
Messages
4,819
Location
Somewhere
VP Congress: Session 5, Proposal 47

Current unit capture mechanic:
If a melee attack kills an enemy of the same domain, and the attacker can capture units, and the defending plot only has one enemy unit:

If defender cannot be captured (there's a promotion column for that), chance is 0.
If attacker has a set capture chance (CaptureDefeatedEnemyChance exists in DLL, but the table doesn't have this column), use that.
Otherwise, chance = (10 + (attacker base CS / defender base CS) * 40)%, capped at 80%.
Proposal:
Change the chance to 100% if attacker final CS / defender final CS > 1.2 (after all combat modifiers), and 0% otherwise.

Implementation details:
CvUnit::GetCaptureChance currently only takes the enemy unit as parameter. To maintain Lua compatibility, the implementation should calculate both combat strengths inside the GetCaptureChance function, instead of passing them as parameters, or completely bypassing this function.

Rationale:
Chance always sucks, especially one that caps at 10%-80%, which gives a Warrior a chance to capture a 1HP GDR. This guaranteed capture mechanic allows the player (both human and AI) to know exactly where your attacking unit will end up after the attack.
Attacker/defender CS ratio is set to higher than 1, since attacker has the advantage of preying on the weak and picking the killing unit. Very rarely will you choose to attack into a losing battle even if you end up killing the target unit.
 
Last edited by a moderator:
as you are the person who likes to talk about AI-friendliness proposals how would this consistency be AI-friendly? They attack and kill when needed. if you can't teach this, I can't see supporting it.

modmodders definitely use this mechanic more than VP does.
 
@ilteroi, is it viable to teach the AI to consider capturing if this is implemented? Capturing is beneficial to using melee to last hit since the unit doesn't advance.

The knockback and withdraw abilities too, they don't seem to consider those.
 
Last edited:
I sponsor this.
 
Would flanking and unit promotions count towards this? That would make skirmishers give a guaranteed capture in most cases.
 
Would flanking and unit promotions count towards this? That would make skirmishers give a guaranteed capture in most cases.
The proposal does use the final combat modifiers, which would include everything....flanking, defense modifiers, Anti-War Monger Fervor, etc etc
 
I will admit while I really like ideas that remove randomness, I am worried about this one. It seems like something a human could really optimize and then just amass a huge number of units very very quickly.

This would especially be true with Dutch Sea beggars that don't have to worry about terrain and can often do large surrounds of ships to get big flanking bonuses
 
I agree this seems like a crusade to hunt down random numbers, without much concern for if having consistent conversions would be more fun.

This is framed as being more AI-friendly but... is it? It seems to supercharge maxing promotions and flanking, both things that humans are better at executing on than the AI.
 
This is done with an eventual tactical AI support in mind. Currently the AI does NOT factor in capture at all, but when they do, it's better that it's consistent so they know where their killing unit will end up at. Remember that capturing also serves as a "no follow up".
 
This is done with an eventual tactical AI support in mind. Currently the AI does NOT factor in capture at all, but when they do, it's better that it's consistent so they know where their killing unit will end up at. Remember that capturing also serves as a "no follow up".
If you're not going to do the AI work for it, don't assume that it will be done at all.

A nebulous "oh, I'm sure sometime will code the AI to be able to use this mechanic sometime in the future" is not a good reason to make this change, when in the interim, the AI will perform comparatively worse .
 
Last edited:
Tactical AI is exclusively ilteroi's thing. How do you even get them to sponsor?
 
Timestamp post to arrange all the threads in a neat order.
 
Status
Not open for further replies.
Back
Top Bottom