Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[iceball3]

The aim here is simple, allow patterns of combat behavior to be expressed as an extension of organized groups and their leader's tactics skill, or the militia commander's (player's) orders. This specifically concerns specific squads, or nearby groups of allies who happen to be rallying together. Many of these should have elements in the military menu.
Combat-arts would be actions on an individual basis. For instance, wrestling and charging both count as combat arts by this metric.
Situation Tactics
Assess enemy: an action that can be taken by those with high military-tactics skills, this allows the respective squad (or fortress) they're in assess the skill level of hostile and neutral units. Those with very high skill can determine with more specificity, and sometimes before the target even attacks.

Scouting: This tactic can be assigned to small squads, particularly for the purpose of the "assess enemy" action. They will try to use their ambush skill, attempt to avoid the enemy's bearing as opposed to simply remaining a distance from them (project where the enemy is going, and avoid that path), and potentially take pot shots at them with ranged weapons. If spotted, they will sprint away, usually to a "place of retreat". Scouting dwarves will use their observer skill to look for hostiles who are likewise sneaking.

Place of retreat: defined areas in which squads will attempt to retreat to when called for. These can have different levels of defined "security", and squads will tend to prioritize higher security retreat locations if they have several to chose from within similar distance. Security levels and locations are defined in Fortress Mode like burrows, and in worldgen towns exist as their associated regions in towns (pubs, barracks, inside buildings, quarters, castle, etc) Security ratings are arbitrary, and set by the player in fortress mode.

Staggered retreat: dwarves will seek the closest shelter available and defend it. If their current shelter is being overrun, or after a quick look around a corner they determined they can retreat to the nearest shelter with higher security rating, they will retreat further to that location.

Withdraw: The squad will hold a tight formation, (careful not to push their allies prone with pathing unless it's down a 1 tile hallway), relatively slowly retreating to designated place of retreat. When engaged, the squad will prioritize defensive maneuvers and continuing to move further back, unless they are confident they can make quick work of the unit they're being engaged by, or have a confident crippling shot against that unit available.
Combat-art: Dwarves with high shield skill can attempt to block hostile projectiles that are passing through squares adjacent to them. Dwarves who are adjacent to hostile archers can directly "parry" their shots, that is to say, strike their weapons so that any hope of a good aim is lost. Archery weapons ought to be relatively fragile if struck when strung, possibly disabling it for archery purposes outright.

Rout: A squad has completely lost morale and stopped listening to their leader. Members will individually continue fighting as if in swarm, or run away in attempt to break line of sight with hostiles as soon as possible. Once line of sight is broken, units will move towards the nearest place of retreat until their morale restores. Leaders use the related social and leadership skills to restore this morale and can be ordered to seek and rally disorganized troops.


Engagement Tactics
Low-risk tough target: This tactic can be called for when there is a particularly durable, dangerous, but unskilled target. Dwarves (or whoever is using this tactic) will attempt to approximately surround the unit to prevent easy escape. There, they will attempt to attack the unit as normal. Dwarves suffering from very low stamina, nausea, or other debilitation will withdraw temporarily, or retreat outright for significant injuries (pulping, bone break).
The intent of this militia art is to help dwarves who are poorly equipped take on large animals with little risk, assuming the animal is basically crippled by it's low stamina. Specifically, it's to prevent dwarves from attacking a unit that isn't a major threat until they themselves fall unconscious, where they are vulnerable to killing-blows in the brief moments the hostile wakes up. This also applies to the likes of highly skilled squads trying to face off with megabeasts.

Intervention Arts: When a dwarf isn't being directly targeted and grappled, their highest priority is to attack the limbs of nearby targets being used to injure their allies, assuming there are somewhat confident hits to be made. This tactic is a call against wrestlers and large creatures with bites alike. If an enemy is standing over a prone ally, a dwarf can either try forcing the hostile off of them, damaging the hostile's legs, or "dragging" their ally out from under the hostile as a combat-art.

Swarm: Current normal combat behavior, requires little training or morale to maintain.

Subdue: Dwarves will attempt to subdue the target while keeping them alive. Edged attacks are avoided, and dwarves will attempt to wrestle and pummel the target into submission. As a combat-art, a dwarf can be equipped with ropes, chains, or a cage in which to trap the target when they are unconscious or over-exerted, but will be instantly interrupted by incoming attacks when attempting to do so. Hostiles can request quarter which will allow them to be captured earlier. Dwarves against dramatically powerful units will be too intimidated to subdue and capture them in this manner, unless they request quarter or are already unconscious. Will basically not be used against groups of enemies.

Hold Ground: Dwarves will attempt to defend a tight location, refusing to advance beyond where they are assigned to to engage enemies. No acrobatics through fortifications allowed either, though dwarves may attempt to seek cover in obstacles in their defensive area to help against ranged attackers if they can't shoot back.
Cover: Like Hold Ground, but dwarves will go prone until there is an enemy unit within a few tiles range. If they have not been engaged yet, they will be in ambush mode. This is specifically to reduce the chances of being hit by oncoming archery fire before engagement, but can be used by archers and soldiers alike to ambush enemy squads in the forest.
Those are my ideas so far. What do you think? Do you have any ideas yourself? Make sure to take into account how Dwarf Fortress works so far when you ponder your idea.

[Sarmatian123]

If this tactics would stop archers rushing to front to bash enemies with their crossbows, then everyone playing this game would be for this improvement. Unless...

However this would be entire new expansion, military expansion, and would probably further slow an already dying to low fps game. Now, this I don't think many would be for, especially this improvement would impact mostly adventure mode.

Next expansion is magic. I think I read somewhere on forum, there was in development thought of expansion dealing with low fps death of fortress once upon time... Long time forgotten maybe. Whatever cures the weight of 20k+ quartz blocks spent on constructions from impacting my fps I am all for it.

[Shonai_Dweller]

If this tactics would stop archers rushing to front to bash enemies with their crossbows, then everyone playing this game would be for this improvement. Unless...

However this would be entire new expansion, military expansion, and would probably further slow an already dying to low fps game. Now, this I don't think many would be for, especially this improvement would impact mostly adventure mode.

Next expansion is magic. I think I read somewhere on forum, there was in development thought of expansion dealing with low fps death of fortress once upon time... Long time forgotten maybe. Whatever cures the weight of 20k+ quartz blocks spent on constructions from impacting my fps I am all for it.

Have you seen the dev notes? Shall we just forget it all because of fps? If so what are we all doing here?

Better tactics and martial arts are in the notes somewhere. It's great to get suggestions on how they might be looked at. In the mean time tech-minded folk can debate fps solutions. (Re-write the game, whatever, doesn't really matter how it's done in the long run, doesn't negate the rest of the game in any way).

And Toady has never mentioned an "expansion dealing with low fps". What people say on a forum has nothing to do with development of Dwarf Fortress for the most part. He deals with optimization as the game is developed.

[SixOfSpades]

A lot of these, particularly those that involve retreating, will be very difficult for the game to compute on the fly. You, the player, would be able to manually designate strategic points like bottlenecks, and assign each a "security rating", but going beyond that is probably asking too much, IMO.

Assess enemy: . . . this allows the respective squad (or fortress) they're in assess the skill level of hostile and neutral units.

Threat level, not skill level. I can plainly see that you're carrying a bow, for example, but I have no idea how skilled you  are with it until I see you in action. This would make more sense if it were based on the type of unit viewed as a threat: e.g., flying enemies (usually) cannot be harmed by melee attacks, so when the user is selecting squads to go confront the flying threat, any squad composed entirely of melee dwarves would be shaded red, to indicate they would not be effective. If there's a dragon or bronze colossus knocking on your door, the raw-recruit unarmored wrestlers would be similarly red-shaded, due to their lack of training and effective equipment. The accuracy of the shading could be dependent on the intelligence/experience of the spotting dwarf--so if the sentry doesn't know that zombies should only be handled with blunt weapons, he might suggest you give The Valiant Axes the go-ahead.

Subdue: . . . a dwarf can be equipped with ropes, chains, or a cage in which to trap the target . . .

Include nets, once fishing nets become a thing. If your civilization has learned that rope nets can be cut by enemies with bladed weapons, someone will get the idea to make War Nets, using chains. Can be used against groups of enemies.

[GoblinCookie]

If this tactics would stop archers rushing to front to bash enemies with their crossbows, then everyone playing this game would be for this improvement. Unless...

However this would be entire new expansion, military expansion, and would probably further slow an already dying to low fps game. Now, this I don't think many would be for, especially this improvement would impact mostly adventure mode.

Next expansion is magic. I think I read somewhere on forum, there was in development thought of expansion dealing with low fps death of fortress once upon time... Long time forgotten maybe. Whatever cures the weight of 20k+ quartz blocks spent on constructions from impacting my fps I am all for it.

Have you seen the dev notes? Shall we just forget it all because of fps? If so what are we all doing here?

Better tactics and martial arts are in the notes somewhere. It's great to get suggestions on how they might be looked at. In the mean time tech-minded folk can debate fps solutions. (Re-write the game, whatever, doesn't really matter how it's done in the long run, doesn't negate the rest of the game in any way).

And Toady has never mentioned an "expansion dealing with low fps". What people say on a forum has nothing to do with development of Dwarf Fortress for the most part. He deals with optimization as the game is developed.

FPS is something that the devs indeed need to take into account. 

But really I don't agree that it would be harder on the FPS to have a proper tactic system.  The main FPS hit is pathfinding and having the whole squad act in a unified manner would be good for FPS inherantly since less pathfinding needed. 

[iceball3]

If this tactics would stop archers rushing to front to bash enemies with their crossbows, then everyone playing this game would be for this improvement. Unless...

However this would be entire new expansion, military expansion, and would probably further slow an already dying to low fps game. Now, this I don't think many would be for, especially this improvement would impact mostly adventure mode.

To be honest, allowing your squads to do something besides "run directly into the enemy and desperately bludgeon them to death" would totally be of use to dwarf fortress players.
That, and if you simply just set their behaviors to default, your squads won't have much of an impact on fps taking on certain targets.
Have you looked into optimizing ram latency and frequency in your computer upgrades?

[iceball3]

Assess enemy: . . . this allows the respective squad (or fortress) they're in assess the skill level of hostile and neutral units.

Threat level, not skill level. I can plainly see that you're carrying a bow, for example, but I have no idea how skilled you  are with it until I see you in action. This would make more sense if it were based on the type of unit viewed as a threat: e.g., flying enemies (usually) cannot be harmed by melee attacks, so when the user is selecting squads to go confront the flying threat, any squad composed entirely of melee dwarves would be shaded red, to indicate they would not be effective. If there's a dragon or bronze colossus knocking on your door, the raw-recruit unarmored wrestlers would be similarly red-shaded, due to their lack of training and effective equipment. The accuracy of the shading could be dependent on the intelligence/experience of the spotting dwarf--so if the sentry doesn't know that zombies should only be handled with blunt weapons, he might suggest you give The Valiant Axes the go-ahead.

I figure that dwarves with the proper skills can make approximate assessments on enemy skill levels by noting their visibly displayed practical and weapon handling discipline. It would be a dramatic difference between a haggard untrained slave army with little training and a professional army with high skill levels, at least, even with mostly the same gear.

[Sarmatian123]

FPS is something that the devs indeed need to take into account. 

But really I don't agree that it would be harder on the FPS to have a proper tactic system.  The main FPS hit is pathfinding and having the whole squad act in a unified manner would be good for FPS inherantly since less pathfinding needed.

I did 10+ embarks, just because pathfinding as main issue hitting FPS was suggested. So I tried to wiggle my fortresses around different architectures. I found out that 20k blocks is more efficient fps killer, then any pathfinding ever could. More bricks I put in, the slower pace game run at INHERENTLY no matter how I put them.

Furthermore walling around embarks (just 6 tiles from map's edge) and leaving 4 3-tiles wide entrances and building a lot of roads around at edges and to center of map to my fortress usually hit my fps at worst extremes. With just 1-3 land animals pathing their way around to half enclosed forest areas and that was on my large 220 Dwarves colony embark. Just 1-3 animals? Not all those bricks I put in into walls and roads? It just doesn't inherently make sense.

It is not pathfinding that kills inherently FPS, but amount of bricks put into construction. 10k bricks hello 20fps, 20k bricks hello 10 fps, 40k bricks hello 1-2 fps. Do you get it? Even if you excavate by channeling and pave with bricks and roads everything to be on same level. Constructions inherently kill FPS, not pathfinding issues.

[Sarmatian123]

To be honest, allowing your squads to do something besides "run directly into the enemy and desperately bludgeon them to death" would totally be of use to dwarf fortress players.
That, and if you simply just set their behaviors to default, your squads won't have much of an impact on fps taking on certain targets.
Have you looked into optimizing ram latency and frequency in your computer upgrades?

Maybe you are right there. Since I upgraded from 32bit os to 64bit os, I need more RAM indeed.

[Lesurous]

All I want is to be able to tell squads to engage melee or range, and to equip back up weapons for marksdwarves, so they have something for melee combat other than a crossbow butt.

[iceball3]

To be honest, allowing your squads to do something besides "run directly into the enemy and desperately bludgeon them to death" would totally be of use to dwarf fortress players.
That, and if you simply just set their behaviors to default, your squads won't have much of an impact on fps taking on certain targets.
Have you looked into optimizing ram latency and frequency in your computer upgrades?

Maybe you are right there. Since I upgraded from 32bit os to 64bit os, I need more RAM indeed.

Not just "more ram", though it might be helpful if you're running low, but more speed, higher Mhz and lower CAS latency.

All I want is to be able to tell squads to engage melee or range, and to equip back up weapons for marksdwarves, so they have something for melee combat other than a crossbow butt.

Backup weapons for ranged combatants would be useful, yes. If you're looking for a current workaround, you can edit the crossbow raws so they come with bayonets pre-installed. Might be ahistorical, but makes sense if it's the only weapon they'll want to hold anyway.

[Sarmatian123]

Backup weapons for ranged combatants would be useful, yes. If you're looking for a current workaround, you can edit the crossbow raws so they come with bayonets pre-installed. Might be ahistorical, but makes sense if it's the only weapon they'll want to hold anyway.

I wish you could edit crossbow raws, so Dwarves could use the same military issued crossbow for hunting too, without constantly throwing it at the floor. What kind of ahistoric bayonet suspenders I would have to put on crossbows to enable that? eh...

[GoblinCookie]

I did 10+ embarks, just because pathfinding as main issue hitting FPS was suggested. So I tried to wiggle my fortresses around different architectures. I found out that 20k blocks is more efficient fps killer, then any pathfinding ever could. More bricks I put in, the slower pace game run at INHERENTLY no matter how I put them.

Furthermore walling around embarks (just 6 tiles from map's edge) and leaving 4 3-tiles wide entrances and building a lot of roads around at edges and to center of map to my fortress usually hit my fps at worst extremes. With just 1-3 land animals pathing their way around to half enclosed forest areas and that was on my large 220 Dwarves colony embark. Just 1-3 animals? Not all those bricks I put in into walls and roads? It just doesn't inherently make sense.

It is not pathfinding that kills inherently FPS, but amount of bricks put into construction. 10k bricks hello 20fps, 20k bricks hello 10 fps, 40k bricks hello 1-2 fps. Do you get it? Even if you excavate by channeling and pave with bricks and roads everything to be on same level. Constructions inherently kill FPS, not pathfinding issues.

Just because pathfinding is the main load on FPS does not mean that if you throw in enough things that are not pathfinding you will not end up killing FPS.

If you build enough blocks it will kill FPS.  That is because every block is a tracked object, the game has to continually count every single brick in order that it can then tell you how much the total value of your fortress is.  The same does not apply to the regular rocks because the game does not have to count them up for anything.  But why do even need to know what the total value of your fortress is, are you intending to sell it anytime soon? 

This is a problem with the game.  It is totally blocked up with redundant mechanics that are not needed for anything at all.  If we actually were to sell our fortress, then it would make sense to then count the total amount of blocks in your fortress because the total value of your fortress now means something.  We could also force-pause the game for a while, in order to reuse processing-power for said calculations.  I suppose not to be too harsh on the devs, back when they were less experienced Toady One just threw in mechanics without much through for the cost/necessity ratio until they ran into the limitations of the engine; I find no evidence that things are being developed inefficiently nowadays. 

This is why adding proper squad mechanics would improve FPS.  By removing active squad members from the list of regular pathfinding creatures which is not presently what happens, only the squad now has to pathfind not all it's individual members.

[Bumber]

But why do even need to know what the total value of your fortress is, are you intending to sell it anytime soon?

Score?

[iceball3]

But why do even need to know what the total value of your fortress is, are you intending to sell it anytime soon?

Score?

No idea if it's actually putting a dent in FPS, but if it is, making it so value is only periodically calculated would be nice.