Proposal (Draft 1): Convert "Withdraw from Melee Bonuses" into "Reduced Melee Damage"

Stalker0

Baller Magnus
Joined
Dec 31, 2005
Messages
10,766
This is a draft of a proposal to get initial feedback.

Rationale: Combat in Civ 5 is highly tactical, at high levels it relies on precision movement and attacks. While there is some randomness in the damage dealt, every other aspect of combat is highly regimented, it is precise and exact. With one exception....withdraw from melee. Such random mechanics do not belong as the core aspects of combat.

Further, often you need units to hold their ground for one reason or another. As such, a withdraw from melee unit is often weaker than a unit without it.

So the idea is:
  • Remove withdraw from melee as a core mechanic. No unit should be "forced" to have this mechanic.
  • Move these bonuses into leaf promotions. If a person really wants the mechanic they can take it through promotions.
Proposal

New Mechanic - Melee Evasion "Take X% less damage when defending from melee" - When you take a melee attack, you take less damage. This is better than a simple CS adjustment for a few reasons.
  • Precise and powerful: Damage reduction from CS is a lot weaker than just raw reduction.
  • Melee units don't take more damage when attacking you. This is more in the spirit of an "evasion", than a Cs bonus that actually makes the melee attacker take more damage.

Unit Updates (this replaces the "withdraw from melee" mechanic the unit currently has)
  • Commando: Take 25% less damage when defending from melee.
  • Destroyer: Removed with no change (weirdly withdraws work primarily against fellow destroyers, creating a weird matchup where destroyers are weaker vs each other. Destroyers are fine units, they don't need a gimmic to make them work, so we are just removing it here).
  • Nau: Take 25% less damage when defending from melee.
  • Commanche Rider: Take 25% less damage when defending from melee.
  • Camel Archer: Take 25% less damage when defending from melee.
Wonders
  • Alhambra: Jinete promotion now gives: Mounted melee units take 15% less damage when defending from melee.
Promotions
  • Blockade: Untouched (as a leaf promotion we leave this alone if someone does want to withdraw from melee.
  • Wolfpack 3: Unit takes +10% less damage from melee (I would like people's thoughts on this one especially. I mean ultimately should subs have this at all? Destroyers are supposed to be their counter, and the sub protects itself mainly by not being there, its not supposed to get hit.
 
I like this approach, but I personally would remove the mechanic entirely (I hate it), not even on leaf promotion. Just my thoughts.
 
I'll make the case for why Withdraw From Melee is a good trait and should exist in the game. I'll preface with agreeing that it belongs on leaf promotions only, as an opt-in mechanic.


Withdraw From Melee is perfect for depicting a unit whose power comes from outmaneuvering and causing overextension by an opponent. It plays to a specific power fantasy of kiting out an opponent, drawing them away from their lines and giving you an opportunity to punish them.

It also plays well mechanically because it lets you have units that are hyper-mobile, without giving them 5 movement points and letting them dash in from out of sight, taking their shots and disappearing between turns. Playing without animations you'd never know where they went. Withdraw solves this by only making the unit move when they are attacked, letting you see where they go. On top of that, it gives you a very clear counter-play: simply use ranged units. They don't withdraw if you shoot them, so archers become a natural counter to skirmishers, for example.

It also adds an element of randomness to combat that, for some, keeps it fresh. I know there are players that prefer combat to be like chess, but I'm in the camp that likes a little bit of chaos that forces you to react to a situation you didn't necessarily foresee. This point is highly subjective, so I won't pretend like randomness trumps consistency, but for those who do like to play with it, the system injects a lot of fun.

I don't think the mechanic is without issue: if there was a way to make it a guaranteed Withdraw 1/turn (2/turn with a higher tier promo), I'd be all for that. It cuts down on giving a unit theoretically unlimited movement, and gives a balance of calculation and randomness.


All that said, this proposal looks good, and I'll support it. As mentioned in the Discord, Jinete really should be a skirmisher promotion, not a knight promotion.
 
Last edited:
Don't see why we can't have both.

I think the ability to withdraw from melee unless cornered is a great bonus for something like the Comanches and should stay there. I think the best use of the ability in that case is for it to be 100% withdraw chance, so both sides are relatively certain that, unless cornered, that unit will withdraw.
The fact that it's on destroyers and can be given to knights etc via a wonder shouldn't salt the earth for more acceptable uses for such an ability.
 
Make withdrawing 100% chance like above but give bonus melee defence when not cornered instead of nullifying damage. Existing withdraw promotions will instead increase this bonus.
 
Make withdrawing 100% chance like above but give bonus melee defence when not cornered instead of nullifying damage. Existing withdraw promotions will instead increase this bonus.
not good enough, still not easily predictable where my unit will go when attacked.

also the withdraw feature is simply not AI friendly.
 
not good enough, still not easily predictable where my unit will go when attacked.

also the withdraw feature is simply not AI friendly.
While not perfect, we could consider adding some consistency to the mechanic. So for example, when a unit hits another with melee withdraw, each hex around the unit has a specific priority. The unit will always withdraw to the highest priority hex, unless that hex is not viable, in which case it goes to the next one. I am not considering this as a replacement to my proprosal but possibly an augment (or probably a seperate proposal as I don't think one is necessary to have the other).

Here's an example:

1687354749288.png


So in this example, the unit attacks from the left. The unit will always withdraw to hex 1 (aka directly away from it) if that hex is viable. If not, than it will always withdraw to hex 2, and then if hex 2 is not available, than hex 3, etc etc.

The logic is:
1) Withdraw directly across from the attack if able.
2) If not, withdraw 1 hex counterclockwise.
3) if not, withdraw 1 hex clockwise.
4) if not, withdraw 2 hexs counterclockwise.
5) if not, withdraw 2 hexs clockwise.
6) if not, no withdraw occurs.

I chose counterclockwise over clockwise purely arbitrarily, we could easily swap them.
 
I'd also offer: 4 and 5 don't even need to be allowed withdraw spots, if part of the gripe is not knowing if you'll actually pull away or not.

I'd be a bit wary of 100% withdraw with no movement cap though, it adds determinism but it also goes from 35% damage reduction to 100% versus melee. I know I argued for the thrill of kiting but literally needing to corner a target against units or mountains/water sounds abysmal. At least (I think) terrain costs factor into the random chance, so hills and trees can become corners too.
 
I'm currently playing against Arabia with Camels and Rome with Jinete- and it is frustrating to have everything set up and then have your enemy withdrawal but it is also kinda of interesting as well because you have to find a solution- as your melee or mounted melee ( which is probably heavily promoted) is no longer valid so you have to have other units like explorers or parthian tactics with exempt ZOC to surround or use range -- so although I was frustrated I was laughing at the same time....

I don't usually noticed it on my own units - as I'm usually packed together or I can't really rely on % chance withdrawal as what if it doesn't occur.- and I agree it is not useful or helpful on my own units.

Although I do play like chess -- on one hand I do like the variability as that is one of the benefits that VOX brings to the table over vanilla to me (i.e. variability or having to find different solutions for different situations) - and as adding additional complexity to warfare that benefits to the AI- but I would not like it on my own units like destroyers which I use to hold ground.

I would be for this proposal.

Also - I think from role play or historical perspective the damage reduction does make sense on the subs- as I think sub would have a good chance to evade or at least some % chance to evade if eventually caught.
 
I like having the effect in the game overall, but I agree its been haphazardly distributed. Not in-love with the proposal, but I don't hate it either.

Alternatively, I wonder if the withdraw bonus can be removed via promo -- if the intention is to have a user-option to either have it or not on a unit, and imagining we have a unit with 50% withdraw chance; if unit gets a -50% withdraw chance via leaf promo to bring chance down to zero, effectively we are achieving same result but with smaller change? Anyone know if this particular bonus can be applied this way?

While not perfect, we could consider adding some consistency to the mechanic. So for example, when a unit hits another with melee withdraw, each hex around the unit has a specific priority. The unit will always withdraw to the highest priority hex, unless that hex is not viable, in which case it goes to the next one. I am not considering this as a replacement to my proprosal but possibly an augment (or probably a seperate proposal as I don't think one is necessary to have the other).

Here's an example:

View attachment 665290

So in this example, the unit attacks from the left. The unit will always withdraw to hex 1 (aka directly away from it) if that hex is viable. If not, than it will always withdraw to hex 2, and then if hex 2 is not available, than hex 3, etc etc.

The logic is:
1) Withdraw directly across from the attack if able.
2) If not, withdraw 1 hex counterclockwise.
3) if not, withdraw 1 hex clockwise.
4) if not, withdraw 2 hexs counterclockwise.
5) if not, withdraw 2 hexs clockwise.
6) if not, no withdraw occurs.

I chose counterclockwise over clockwise purely arbitrarily, we could easily swap them.

I think it'll be preferable to have withdrawing unit prioritize withdraw direction based on what's in the plot -- ie if 1 is outside my territory but 3 is inside, i'd usually want it to go to 3 rather than 1. Or if one tile has a road and others don't, etc. Years ago I created such a function in .lua for MB+ mod, elsewhere on these forums -- it works well enough, better than RNG or just arbitrary direction rules.
 
I think it'll be preferable to have withdrawing unit prioritize withdraw direction based on what's in the plot -- ie if 1 is outside my territory but 3 is inside, i'd usually want it to go to 3 rather than 1. Or if one tile has a road and others don't, etc. Years ago I created such a function in .lua for MB+ mod, elsewhere on these forums -- it works well enough, better than RNG or just arbitrary direction rules.
I think such methods would be way too complex. The user should be able at a glance to tell what direction the unit will withdraw, if they are having to pull out a table or a spreadsheet of conditions its not going to work.
 
I think such methods would be way too complex. The user should be able at a glance to tell what direction the unit will withdraw, if they are having to pull out a table or a spreadsheet of conditions its not going to work.

it can be as complex or simple as desired. The directional rules proposed are an improvement over RNG in any case, however it would be nice to have it responsive to game world: alternatively retreat plots can be evaluated for distance to capital, and unit always retreats to tile that puts closer to capital, and further from attack plot, or it doesn't retreat. ie Simple logic connected to in-game world, that adds the desired predictability

It also adds an element of randomness to combat that, for some, keeps it fresh. I know there are players that prefer combat to be like chess, but I'm in the camp that likes a little bit of chaos that forces you to react to a situation you didn't necessarily foresee. This point is highly subjective, so I won't pretend like randomness trumps consistency, but for those who do like to play with it, the system injects a lot of fun.

The same, or similar, desired "fun" effect can be achieved by layering somewhat orderly rules on top of one another -- eg. allowing the withdraw from melee to exist but mitigating the RNG component in the direction it withdraws to, as has been suggested -- attacking player won't be sure if it will withdraw but can at least plan for the contingency, and similarly smart defending player might leave certain tile available for retreat and not others.

I don't think the mechanic is without issue: if there was a way to make it a guaranteed Withdraw 1/turn (2/turn with a higher tier promo), I'd be all for that. It cuts down on giving a unit theoretically unlimited movement, and gives a balance of calculation and randomness.

This strikes me as a nice restriction to have regardless, even if there were no other changes
 
Last edited:
I don't think the mechanic is without issue: if there was a way to make it a guaranteed Withdraw 1/turn (2/turn with a higher tier promo), I'd be all for that. It cuts down on giving a unit theoretically unlimited movement, and gives a balance of calculation and randomness.
strongly in favor of this. I was even going to suggest it myself until I read over the posts more thoroughly and saw it was already mentioned.

I also think it should be 100% when applicable, and no withdrawing towards the attacker (only 1, 2, and 3 in that diagram above).

Alternatively... different levels of withdraw can affect available withdraw directions as well. Though getting into this (and further, making available directions downgrade after successive withdrawals in the same turn) risk becoming too much to keep track of.
 
strongly in favor of this. I was even going to suggest it myself until I read over the posts more thoroughly and saw it was already mentioned.

I also think it should be 100% when applicable, and no withdrawing towards the attacker (only 1, 2, and 3 in that diagram above).

Alternatively... different levels of withdraw can affect available withdraw directions as well. Though getting into this (and further, making available directions downgrade after successive withdrawals in the same turn) risk becoming too much to keep track of.
So the suggestion is:

1) First attack, 100% withdraw.
2) Subsequent attacks that turn, 0% withdraw.

Not sure how easy that is to code, but seems reasonable
 
I like that. Keep it on a couple units like Comanche, camel archer and subs.

Add the resistance to melee damage thing too. That seems fine. I don’t think they should be considered the same thing; I just think withdraw from melee has been applied too liberally
 
OP expressed some desire for these withdraw units to be able to hold ground in reasoning it should be largely removed, though I prefer the approach of keeping it in but making it more structured; two thoughts:
- withdraw could exclude unit at full HP; ie unit must be already damaged to withdraw
- and/or fortified units do not withdraw

the latter would only apply to a few units, but I imagine a fortified unit is intended by the player to stay put in most cases. Anyway there are definitely some ideas I could vote for here.
 
I disagree with the premise, mainly.

I think withdraw from melee is a good mechanic that should exist. The random chance is not fun, but that can be changed. It’s on too many units, and some in particular are inappropriate places for the bonus. That can also be changed.

We can have a damage reduction from melee. That’s fine. That’s an entirely separate ability though; it’s not a substitute. Maybe some of the places that currently have withdraw could have melee damage reduction instead, sure. That doesn’t make the two abilities equivalent.
 
ok so this is shaping up into 2 proposals, so lets break them up.

Proposal 1: Reduce Withdraw from Melee Mechanic
Unit Updates
(this replaces the "withdraw from melee" mechanic the unit currently has)
  • Commando: Take 25% less damage when defending from melee.
  • Destroyer: Removed.
  • Nau: Take 25% less damage when defending from melee. (should this be unchanged to, should we leave the mechanic in for UUs?)
  • Commanche Rider: Unchanged
  • Camel Archer: Unchanged
Wonders
  • Alhambra: Jinete promotion now gives: Mounted melee units take 15% less damage when defending from melee.
Promotions
  • Blockade: Untouched (as a leaf promotion we leave this alone if someone does want to withdraw from melee.
  • Wolfpack 3: Unit takes +10% less damage from melee

Proposal 2: Updates to Withdraw from Melee
  • Mechanic now works 100% of the time on the first attack (no longer considered based on speed). Works 0% on subsequent attacks.
  • Withdraw always occurs to hex across from the attacker if possible. If not, goes to counterclockwise hex. If not, then clockwise hex. If all 3 hexes are blocked, withdraw drops to 0%.
 
Top Bottom