SP - Battles & Sieges Suggestion to improve Combat: Momentum, Friction, and Attractor-Stickiness

Users who are viewing this thread

TL;DR: By treating each unit as a particle with momentum and friction, as well as reacting to attractors (and repulsors), you can resolve bottlenecks and dumb behavior, as well as adding realism and more tactical depth to the game.

1. Momentum: When riding into a crowd, the units on foot slide to the side of a mounted unit, when in reality you should see the impact throw bodies back as momentum is transferred from the mounted unit to the ones hit. The more bodies the mounted unit hits, the more it loses momentum. The bodies that are thrown may collide with someone nearby, imparting their momentum on others, etc etc, making a sort of ripple effect. Since these impacts are not perfectly inelastic, the "wave" should die out after the second or third person in the chain of collisions.

These impacts, should interrupt weapon swings or even knock people down. A simple calculation could be if the momentum transferred (∆mv) overcame a threshold for knockdown, which could be calculated by the unit's mass, and their movement state (bracing would give a bonus, running would be a penalty), and maybe even their athletics skill.

This brings up the important, very important concept of bracing - when shieldbearers in a shieldwall interlock shields, and soldiers behind them help reinforce by pushing or putting additional shields on top of the first (kneeling rank). Currently in Bannerlord (as Warband, as well) you'll see a mass of infantry run into a mass of stationary infantry, but instead of bouncing back, the charging infantry mass compacts into a tightly overlapping mass of soldiers bodies clipping into one another. What should happen instead is that the first wave of attackers will hit the shieldwall (with their bodyweight behind their own shields, in an attempt to break the wall) and we should see them bounce backwards like a ball bouncing off a wall. The attackers right behind them will collide with this first rank, bouncing the first rank against the wall again, and the friction (more on that later) that reduces the momentum transferred from all these reactions means that within a few seconds, the attacking mass stops vibrating.

This is then a good point to talk about....

2. Friction: There's two kinds of friction, static and kinetic. But rather than a complex calculation of every surface of a unit or weapon or shield, you can simply treat the entire unit as a particle, and you can think of the unit's stability (ability to stay in place) as the coefficient of static friction. It suffices to say that once you overcome a certain threshold, you'll move that unit. Think of two wrestlers with interlocked arms that are simply pushing on one another. If one wrestler is stronger and pushes the other back, we can say that the weaker wrestler's stability was overcome, as in physics when a force overcomes an object's static friction and starts moving it.

When an attacker charges and hits shield-to-shield with a defender, unless that defender is sufficiently braced or massive, they will be thrown back (and the attacker will end up not moving, having transferred their momentum). If instead you have two defenders pushing on a single shield, it is more likely that it will be the attacker that bounces backwards. Since humans are not infinitely hard objects not all the momentum is perfectly transferred or reflected - think of arms and legs as dampening shock-absorbers. This in effect becomes the kinetic friction between moving objects. Soldiers moving in tandem in a formation (not bumping against each) other minimize this kinetic friction. How well soldiers move in tandem is matter of drills and practice - and modeling this self-organizing behavior is where attractors and repulsors come in.

3. Attractors / Repulsors: The game already uses some form of them, and it's easy to see in the way units move or go into formation. What I'm suggesting is having additional attractors/respulsers on a smaller scale, and to implement some features that may require more animations to be made:

- Sticky Spots: Any time someone grapples another, locks shields, pulls or pushes; these two units are effectively stuck together. In-game, it should work like magnets coming together. For a shield wall, each unit's shield should have a "Friendly-Shield-Attractor" point to the left, right, and above (if the game wishes to implement testudo formations). That means a friendly unit nearby have a magnetic-like force attracting their neighbors's shield, but only opposite sides; in other words, left sides of shields only attract right sides, and vice versa.

Another use for stickiness is at geographic features or structures. For examples, the edges of crenels of a battlement should attract nearby ranged units to stand just slightly aside the crenel, and pop out to shoot intermittently. (This highlights the need for a "shooting from cover" animation in game, by the way). Trees should generate an attractor on the side opposite the opposing force, so a dispersed troop of archers in a copse of trees will automatically station themselves behind those trees, minimizing their exposure. The stickiness of that position keeps them from moving around constantly (unrealistic and immersion-breaking) and they are more likely to stay in place when faced with a charge.

An infantry unit in the first rank of a shieldwall should also have an attractor for second rank units to come up behind and hold their shields upward and provide greater stability for the first rank. The second rank units should generate an attractor for spear-wielding units to come up behind and thrust weapons over the first two ranks.

There's lots of positions around siege equipment for units to hold up shields protecting the units moving the towers or battering ram that need to be addressed, as many others have pointed out.

- Personal Space: without changing the size of the collision mesh, each unit should have a respulsor acting at close-by distances. That way, units can get squished together momentarily (like from two crowds colliding), but they will naturally spread out after that (unless they are sticking together, as outlined above). Horses tend to "pull" each other along, in a sort of flocking behavior where sudden changes of direction from leading horses will quickly influence the direction of following horses

- Hazards: The edges of walls and stairs without safety railings should also exert a repulsor force, which is only come when there's a lot of shoving that makes units fall over the edge (not simply thru accident of overcrowding). The falling animation should happen for any unit that falls over an edge, for that matter. Trees and walls should be strong repulsors to horses (horses will not willingly run into a tree or wall, you know) which makes a gentle way of soft collision-avoidance, which is something horses do naturally. I'm super tired of my horse coming to dead stop at a tree while I'm busy skirmishing, and it just simply looks stupid -- all horses in the game should naturally swerve around trees.

---------- EDIT: More-------
- Objectives: The scripting in sieges are pretty broken, I know. But instead of simple scripting to tell soldiers to climb ladders or to crew ballista, why not make certain interaction items be attractors in the sense that if a unit is close to an unused ballista, they will feel a strong compulsion to use that ballista. The same with ladders -- make the pull of an empty ladder stronger than the pull of one that is full. Make the pull of the broken front gate pretty strong, but if a soldier is right next to an empty ladder, the empty ladder will exert a stronger pull.
 
Last edited:
That almost all makes sense and I agree with 90% especially horses serving to avoid trees and momentum.

That part about attractors and repulsors I am trying to envision and not sure how that would look. So nearby shields in a shield wall would 'snap' together when the men move close for the formation 'stance' that allows less personal space?

Opposing attackers would have their shields drawn to the defenders shields and create a big shoving match? The defenders would 'win' this match because they are in shield wall and there are no attacking formation for infantry? Or this would give spears a chance to be effective because the attackers could 'advance' into stabbing range of the spearwall and not have to worry as much about getting blocked by the person next to them and able to thrust freely- it would make shieldwall something more for short range weapons while spearmen and polearms would be in their natural formation when doing an advance?
 
Last edited:
Updating and reowkinr physics would be best update but that might be impossible. As i say, they know whats worth their time and in this case it may be not, then we have to live with it.
More over, they might need to scrap and begin developing a code from scratch. From my gaming XP, core mechanics are never reworked after they release it in Alpha version of the game.
 
It really gets me when I'm full speed on an armored horse charging at a single Aserai peasant unit in cloth robes with nothing but a small spear, and while they are moving their feet manage to get in front of my horse and make an attack, which does almost no damage, but stops my horse directly in it's tracks and forces it to rear up. What an immersion breaker. All laws of physics completely ignored. All momentum vanishes.

I would like it to take more to stop a horse dead in it's tracks. A successful hit on charging horse should not completely interrupt it's movement.Cavalry charges need more impact. Also it would be nice to tell my cavalry to equip their lances and couch them.
 
That almost all makes sense and I agree with 90% especially horses serving to avoid trees and momentum.

That part about attractors and repulsors I am trying to envision and not sure how that would look. So nearby shields in a shield wall would 'snap' together when the men move close for the formation 'stance' that allows less personal space?
It's actually quite simple in practice. Imagine an attractor simply exerts a pulling force and a repulsor exerts a pushing force. As the horse gets closer to the tree, the repulsor force gets stronger in the direction perpendicular to the vector between horse and tree.

The snapping of shields together is exactly how I would envision it.

Opposing attackers would have their shields drawn to the defenders shields and create a big shoving match?

If you watch The Last Kingdom or the HBO series Rome (just the first episode), you'll get an idea of how shieldwall tactics work. This tactic goes back further than the Roman Empire, with Greek Phalanxes and such - it was the de facto way wars were fought for thousands of years. I'm no historical expert but watching two waves on infantry come together and idiotically chop each other to bits in seconds just doesn't look right to me.Tactics were tests of endurance, discipline, and clever ways of maximizing resources, not just numbers.
 
Updating and reowkinr physics would be best update but that might be impossible. As i say, they know whats worth their time and in this case it may be not, then we have to live with it.
More over, they might need to scrap and begin developing a code from scratch. From my gaming XP, core mechanics are never reworked after they release it in Alpha version of the game.
Unfortunately, you're right!

I was not part of the Beta test, otherwise I would have noticed this and made the suggestion earlier. In games, it becomes readily apparent what games utilize momentum in their animations and which do not -- if you press a direction and your character instantly is moving at top velocity in that direction, then you have a game without momentum-based physics. (when a keypress = velocity, rather than keypress = force) Multiplayer online action games that don't have momentum annoy me, because it allows "twitchy" players to spam the direction/jump buttons and negates the whole point of tactics.
 
The AI also needs to be able to fight in a shield wall. Perhaps the second rank could use the "up" attack with their spears and stab over their comrades' shoulders? As it stands the shield wall is only useful as an arrow sponge and actually leaves your infantry vulnerable to the enemy's infantry and cavalry as they can just envelop or even permeate into the shield wall and attack your troops, who are too busy looking at the backs of their shields to answer their attackers.

That's not even to mention that soldiers are pressed so close to each other the combat doesn't even work properly; they're constantly chamber-blocked by teammates behind them and most attacks are whiffed off teammates beside and behind them. This is compounded by the fact that the AI has only one fighting style, which is face-hugging. The AI is far too relentless in how it never moves backwards, regardless of their weapon. They'll walk up to you with a spear and use it like it's a dagger. In reality fights are constant in/out movement, getting in and out of range for attack and defense. This has been an issue since warband.
 
"...This is compounded by the fact that the AI has only one fighting style, which is face-hugging. The AI is far too relentless in how it never moves backwards, regardless of their weapon. They'll walk up to you with a spear and use it like it's a dagger. In reality fights are constant in/out movement, getting in and out of range for attack and defense. This has been an issue since warband."

This is one of the weakest points in warband and DEVs should understand that it's not worth making same mistake again. On the other hand it might be very limited how much they can do regarding to AI. There are lots of good games with problematic AI, i cant mention exact alternatives to M&B series, but games like KCD and Total war facing same exact issues.
 
As someone who has spent a lifetime of fighting in various martial arts, with and without weapons-- please no. The whole stickiness/attractor/repuslor thing is not realistic at individual level combat, although it can be a good model for getting the correct outcome for unit behaviors. They did this in Kingdom Come: Deliverance and eventually got it optimized, but it did not feel like a real fight. Modeling primitive behaviors like Taleworlds does is an unusual, novel approach and I think it is by far the closest to actually fighting. You can dodge, shift, block, rapidly change direction to attack or shift attack focus. You are never "locked" into a focus with one combatant, and unless you are fighting as part of a tight formation, you shouldn't be. This may make battles LOOK good because you are forcing the outcome to look like real tactical unit behaviors for a shieldwall, but will not FEEL good as a participant. That Kingdom Come feeling of getting magnetically pulled into a fight with participants you did NOT want to engage was horribly annoying and broken; I eventually modded it myself and removed this crap and the game played much better.
 
As someone who has spent a lifetime of fighting in various martial arts, with and without weapons-- please no. The whole stickiness/attractor/repuslor thing is not realistic at individual level combat, although it can be a good model for getting the correct outcome for unit behaviors. They did this in Kingdom Come: Deliverance and eventually got it optimized, but it did not feel like a real fight. Modeling primitive behaviors like Taleworlds does is an unusual, novel approach and I think it is by far the closest to actually fighting. You can dodge, shift, block, rapidly change direction to attack or shift attack focus. You are never "locked" into a focus with one combatant, and unless you are fighting as part of a tight formation, you shouldn't be. This may make battles LOOK good because you are forcing the outcome to look like real tactical unit behaviors for a shieldwall, but will not FEEL good as a participant. That Kingdom Come feeling of getting magnetically pulled into a fight with participants you did NOT want to engage was horribly annoying and broken; I eventually modded it myself and removed this crap and the game played much better.
I'm 99% sure this idea is for the AI's movement, I don't think anyone wants M&B to have lock on style combat.
 
Also want to add that I think your points on friction and momentum are very good suggestions; it's just the attractor/repulsor part I don't agree with for this game. If it were for a game like total war Rome, this would be a great idea, just not for THIS game.
 
I'm 99% sure this idea is for the AI's movement, I don't think anyone wants M&B to have lock on style combat.
That would not be so bad then, although many of the combatants in Bannerlord are modeled after groups who did not fight with tight unit formations; the Celt shield wall (Battanian) was very loose and quickly broke into a charge or individual smaller units trying to flank and individual combat... nothing like the roman shieldwall. The Khuzait and Aserai would not fight in tight formations; maybe to an extent Vlandians and Sturgians would. For the Empire though, having tighter formations and staying locked up would be a better representation, but you can't lock their opponents in as well or you lose the essence of their fighting styles.
 
That would not be so bad then, although many of the combatants in Bannerlord are modeled after groups who did not fight with tight unit formations; the Celt shield wall (Battanian) was very loose and quickly broke into a charge or individual smaller units trying to flank and individual combat... nothing like the roman shieldwall. The Khuzait and Aserai would not fight in tight formations; maybe to an extent Vlandians and Sturgians would. For the Empire though, having tighter formations and staying locked up would be a better representation, but you can't lock their opponents in as well or you lose the essence of their fighting styles.
I understand that but the way melee battles happen in game now feels broken. Two blobs of men just merge into each other and whiff attacks off each other, making no attempt to use their weapons' range or position themselves. They just constantly move forward attacking in random directions. An individual soldier will walk forward into a blob of enemies, get himself surrounded and be whiffed to death, rather than either harass from range or fall into formation for protection. It's ridiculous

The formation system in Bannerlord is so amazingly cool, it's disappointing that it falls apart in actual combat.
 
That would not be so bad then, although many of the combatants in Bannerlord are modeled after groups who did not fight with tight unit formations; the Celt shield wall (Battanian) was very loose and quickly broke into a charge or individual smaller units trying to flank and individual combat... nothing like the roman shieldwall. The Khuzait and Aserai would not fight in tight formations; maybe to an extent Vlandians and Sturgians would. For the Empire though, having tighter formations and staying locked up would be a better representation, but you can't lock their opponents in as well or you lose the essence of their fighting styles.
Excellent point, and I agree that attractor/repulsors should not affect the player-controlled character at all. EXCEPT -- I feel that there's certain "cheats" that should be made to make up for the player's limitations, like aiming a couched lance. I can never seem to line up the lance correctly in 3rd person, and I don't want to switch to 1st all the time, so I think that at higher polearm skill level, there should be a little bit of auto-aiming taking place so it's not so reliant on player skill.

As far as different fighting styles -- absolutely; the position/strength and flavor of attractor-repulsors could be set per culture, and modified by the unit's skill level and current state of morale. It would give reasons to mix and match or separate troops from different cultures. You could even create a whole new game mechanic where you drill your troops in certain formations until they become much better at maintaining that formation.
 
Excellent point, and I agree that attractor/repulsors should not affect the player-controlled character at all. EXCEPT -- I feel that there's certain "cheats" that should be made to make up for the player's limitations, like aiming a couched lance. I can never seem to line up the lance correctly in 3rd person, and I don't want to switch to 1st all the time, so I think that at higher polearm skill level, there should be a little bit of auto-aiming taking place so it's not so reliant on player skill.

As far as different fighting styles -- absolutely; the position/strength and flavor of attractor-repulsors could be set per culture, and modified by the unit's skill level and current state of morale. It would give reasons to mix and match or separate troops from different cultures. You could even create a whole new game mechanic where you drill your troops in certain formations until they become much better at maintaining that formation.
I think every weapon has things you can really do that don't map well into mouse and keyboard; explosive motions and subtle feints with two handed weapons that are really how you avoid getting hit (not blocking), using terrain by crouching/hiding behind cover if you are a skirmisher. Sword and board basic mechanics works amazingly well though (of course it could be tweaked).
Okay, gotta admit that tailored attracto-repoulsors sounds really cool! If the parameters were culture specific and dynamically adapted base on unit experience (and leader style/preference?) that could be extremely interesting. Every strength can become a vulnerability though, even the ability to fight in tight formations (e.g. European field armies vs. Mongol Horse Archers, British redcoats vs. American militias, etc.). I could def see a Battanian leader who broke his infantry dynamically into groups who each had a role (one fixes front, others flank/encircle and use ranged into exposed) could be a major headache for a shieldwall. For a lot of these complex questions I think the best approach is for the dev to offer a general model framework and make it easily accessible to the player (modder) to work out different concepts of implementation, then the end user (pure player) can just grab a mod that matches their preferences.
 
I understand that but the way melee battles happen in game now feels broken. Two blobs of men just merge into each other and whiff attacks off each other, making no attempt to use their weapons' range or position themselves. They just constantly move forward attacking in random directions. An individual soldier will walk forward into a blob of enemies, get himself surrounded and be whiffed to death, rather than either harass from range or fall into formation for protection. It's ridiculous

The formation system in Bannerlord is so amazingly cool, it's disappointing that it falls apart in actual combat.
I agree, but I guess as an eternal optimist I just assume this is the basic combat mechanics framework and now they can work on optimizing it over time to get it to where it's full release quality.
 
Back
Top Bottom