Modeling fighting stances in medieval swordsmanship

[Blackbeard]

I logged in for the first time to this forum because I believe that Sui Generis has great potential to make new standards in RPG genre and I would like to contribute with my knowledge, suggestions and resources. My modest personal opinion comes from years of playing games, programming sandbox mod for ArmA2, as well as from dreams and ideas about RPG game, and fresh experience in martial arts.

I divided this text in sections and important sentences are highlighted.

1. The problem
2. Personal experience
3. Trends and solutions in games (so far)
4. Possible solution in Sui Generis

1 The problem
Characters lack stances and appropriate counter-stances, as well as transitions between stances to be believable combatants. This opinion I formed based on Pre-Alpha Gameplay video, Combat Demo Beta Release announcement, Update #28 and other people’s opinion on this forum.

The character’s waist doesn't seem to be firmly grounded through feet and legs. Characters still look like tennis players trying to hit the ball that is far away (tennis player horizontal dive). It is noticeable on these images too. They show behavior like pendulum or metronome rather than being firmly grounded and dealing blows that are coming from the waist.

Nevertheless, such combat mechanic could be preserved for novice characters, while veteran characters could evolve their moves to more balanced combat style.

2 Personal experience
I study and practice bokken and staff combat for two years now, and the main postulate, the very first milestone I learned, is that every attack originates from the center (Japanese: hara, tanden)/ waist/ hips (in plain words).

One of the most important elements of historical fencing study is that of basic fighting stances and guard positions. Nothing is more fundamental than these two things. Stances or guards (leger/huten or guardia/posta) are in many ways the very foundation of medieval swordsmanship. The offensive and defensive postures and ready positions from which to deliver all manner of blows lie at the heart of any fighting method. Unquestionably, they represent the beginning of study. All principles and techniques of fighting all are employed in relation to these postures. But they are not “static” postures, but dynamic “ready positions” from which to strike or counter-strike (ref.).

There is also appropriate term in martial arts – zanshin: in the context of kendō, zanshin is the continued state of spirit, mental alertness and physical readiness to meet the situation (such as an opposing attack) that must be maintained when one returns to kamae (fighting stance) after attacking. It is one of the essential elements that define a good attack [ref: Wikipedia – “Complete Kendo” by John J. Donohue; Tuttle Publishing, 1999, page 89]. That includes being balanced after finishing the attack.

A friend of mine (together with his team) is doing medieval sword fighting for over 10 years. In this video taken with GoPro camera (only to test the camera and settings) you can see that there is no losing of balance at all. Stepping and sword-work are synchronized. On this other video that I found by searching ‘longsword triangular movement’ you can see that every step is precise regarding speed, angle and distance.

The utility vs. elegance discussion concerning melee combat is one that pops up a lot here.

A common misconception about medieval combat is that it's completely choreographed. Yes, knights were taught and trained in medieval warfare, and possessed the knowledge and the techniques used in what people teach about combat nowadays.

The combat, in its current state, looks chaotic. It makes Sui Generis look like a "swing & pray" kind of game, which it isn't. Behind the chaos, however, are certain patterns, perhaps invisible to a lot of people that do not know a lot about/ interest themselves in combat in any way.

In that sense driving F1 is also choreographed... It looks like that, but once one starts practicing, one realizes how many thoughts and decisions go through beginners mind. On the contrary, masters focus on strategy, and actions are performed accordingly and automatically.

3 Trends and solutions in games (so far)
Sui Generis should stay committed to the original combat concept. Clichés are hard to break but Sui Generis has huge potential to do it! So what should be avoided?

In my opinion, there are two major classes, so far, when it comes to (let’s call it) modeling ‘medieval-sort of’ combat (action RPGs only; games with significant dice mechanics and thus enemy scaling behind combat are excluded here):

1. ‘Anti-button mashing’, skill oriented games – Mount and Blade (series and its derivatives), Chivalry: Medieval Warfare, Dark Messiah of Might and Magic – that give the player full control over the movements and strikes but require timing and fast reflexes (with the mouse and keyboard/ controller).

The problem with the class I was nicely described by one member at the Paradox Interactive forum:

Characters seem to be (and indeed are) composed of two distinct and independent pieces: the torso and the legs. To make a visual analogy: it's like a tank and its turret. You have the upper half that can face any direction while the lower half can move freely. Characters couldn’t make a 360° turn with their torso, thankfully, but it's still pretty unnatural and unrealistic. Nobody could fight like that in real life, moving as if its leg are independent from the body.

I'm going to say it once and for all: melee combat in was horrible to look at. It wasn't a duel between two human beings, it was some kind of robot fight where two weapon platforms on legs tap-danced around each other in jerky motions while swinging unrealistic blows at strange angles and spazzing out feints.

What this game needs is a movement stance and a combat stance. The first one quite obviously allows you to move freely but you can't attack. The second one forces you to take small steps while facing a single direction (or maybe even locking down on a specific opponent) but allows you to attack freely, add the possibility of taking a small but fast hop once in a while to dodge or make a surprise attack and you have a pretty solid combat mechanic in my opinion. No more absurd backpedalling spearmen, no more ridiculous claymore swinging clowns that run around in circles, wouldn't that be good?

And of course this means no more head bobbing and ‘found footage’-like camera movements.​

2. ‘Hack&slash’-like cinematic oriented games that rely heavily on MoCap and inverse kinematics – Assassin’s Creed series, The Witcher series, The Cursed Crusade, Ryse: Son of Rome, Blade Symphony, etc. – that mainly require from the player timing and reacting according to rock-paper-scissors mechanics = light and heavy attacks (guard breaks) vs. block, evade/dodge, parry, deflect… bla, bla, bla…

The problem with the class II is that they tend to oversimplify the gameplay and reduce it to two possibilities: to spam attacks/combos or use ‘fast-kill-counter’ button/combo. In other words, once you understand the mechanics, gameplay becomes ridiculously easy and, although sometimes being fun, extremely repetitive.​

There is, however, the class III of games that fall somewhere in-between. Games like Sui Generis, Kingdom come: Deliverance, Clang and legendary jewel of all – Blade of Darkness. Especially at the beginning of Blade of Darkness, simple moves (without “thousands of combos”) are effective and precise blows are enough to win the fight. Movement, timing and striking together were one swift thought – not just one-button-press, complicated yes, but intuitive and with enough control – and such game mechanics led to spectacular fights.

4 Possible solution in Sui Generis
When I summarize everything said above, what would be solution that might lead to better and believable medieval gameplay in Sui Generis?

Stay faithful to the original concept where combat was model-oriented and not "strafing&chambering"-key-combo-oriented. Players should observe the fight, be patient, adapt to situations and be prepared for the opportunity to strike (opportunity that arises from the combat model). I remember you conceived the combat in the beginning as following: one button dedicated to attack and while not attacking character performs blocking and parrying. Mouse position to control line of sight, movement keys to control character's position. Thrust/stab should be default straight-line attack. Overhead attack should be straight-line attack chained to the right or left cut. Directions are picked up randomly.

If concept diverges from this plan, Sui Generis will just emulate already existing "strafing&chambering" concepts or, in other words, "turrent-tank" combat behavior (class I), with "tennis-play vibe". If you then add: press this to do that, press combination to do something else, then you mix it with class II. And then uniqueness is definitely gone forever… Here is a glimpse of that:

Actually, I think your both slightly off with the stabbing. I was under the impression that you hold to attack normally (slashes) (hence why you can stop a swing to do a "fakey"). And to stab, you double click, then hold the second click. So click, then click again and hold.

You are right. To perform the 'normal' horizontal swings it seems we indeed have to hold the attack button. My bad!
About stabbing/overheads, I agree it sounds a a tiny bit less straightforward than what I had originally expected. I'm sure I can get used to it though.

Distinct synchronization of the movements gives rise to the stances. I searched images for ‘Capo Ferro’. On this particular image you can see that the front, right leg gives the support to the center of the body, while the left leg and left arm provide balance in order to have the most reach with the sword. Again I will refer back to choreography -- this is not a pose but a stance. It has a purpose in combat.

Here is a video where you can see the attack in slow motion. Notice two time points: 1) at 322.655ms the whole strength of the attack, from arms and torso, is grounded through the left foot, 2) at 383.653ms there is a gliding movement over the ground while the right foot is preparing to “land” and left foot is sliding a little bit, thus maintaining solid supportive structure. However, the end of the attack is when the right foot landed, although the opponent stopped the strike. In this other example from the same development team, one can nicely notice stances, movements and how strikes are connected.

Experience combatant should never overcommit to the attack in such way to lose balance and move like being drunk (*from the previous video about overcommiting). Therefore, combat should not be modeled like tennis play where one can have luxury of compromising his/her own balance and follow the tennis racquet with whole body in order to reach the ball. It should be modeled like casting a fishing pole/ rod. If attack misses – it misses – balance is preserved. Distance was too far, that’s ok – one is still in control and ready for the next movement.

Their physics engine is as "correct" as it gets, especially for an RPG. Weapons, objects and players have accurate and realistic collisions, weights and C.O.G.s (Centre Of Gravity) based on all of their properties, rather than a simplified hitbox and an assigned value. Weight and momentum play a big part in the game, especially the combat, and it's handled in the way it would be in real life.

Almost every game, however, uses animations. Whether they are static (canned) or active (uncanned) is fairly unimportant if you look at it in black & white. Those that don't, however, are often games that don't include any substantial gameplay. Sui Generis makes use of active animations, meaning that any possible interference from terrain or objects within the game affects the animation in some way. If it didn't, for example, your feet would clip through the terrain when walking up and down slopes, because the standard animation is that for walking on a straight surface.

The stance is the final product of the synchronization between arms, torso and legs [example]. The whole body should be attacking, not just the arms (or sword like racquet). Therefore, game doesn’t actually need separate movement and combat animations, MoCap references or (semi-)canned attacking animations. What physics here need is just an adjustment of the animation speed of separate parts to form a stance at particular time point (namely at the end of the move) i.e. to be more synchronized. A combatant should not swing a sword while taking three steps, but rather take one step and synchronize the strike with footwork while keeping waist centered.

It’s like when beginners learn kata or flourishing – “a solo free-flowing series of techniques (both attacks and covers) performed in a non-pre-arranged order that simulates fighting against an opponent or multiple opponents and encourages economy of motion, footwork and good bio-mechanics” [quote from The Association for Renaissance Martial Arts Forum]. I believe gamers would really like to see single character connecting blows without losing balance and walk like “drunken robot”! Other aspects of the combat – flow, responsiveness, intuitiveness, simplicity are gorgeous already from the beginning!

 

[Fawz]

Looks like you really put a lot of effort into this complex post about doing melee combat in games. Nice to see such passion from a new forums member; hopefully your lurker days are over as it really seems you have a lot to offer

A lot of the things you mention seem to be very sound, and hopefully it's of some help to the devs (though I suspect they already knew most of this). The aspect about balanced stances rings especially true to me, it does indeed feel like the current swing system has character overextend a lot and not aim to keep a balanced stance when swinging.

I'm just worried about what is technologically possible and what can be done by such a small team that doesn't have a dedicated animator. There's obviously things that can be improved about the current combat system, but it's already a lot of fun and amazingly unique. I'm pretty sure that right now the devs are mostly focused on other parts of the game, but I'm sure they'll come back to touch up the combat aspect of the game throughout development.

 

[lvk]

Hello and welcome to the forums! I hope you'll stick around, well-formatted input is always welcome.

Forum member Elric von Rabenfels also had an issue with how the characters seem to be controlled, that the center of physics appeared to be the shoulders rather than the waist. Since it's on the insider forums I cannot actually copy the full posts, but Madoc's reply was in line with what Fawz said here; since a lot of work already went into the combat system (and indeed because they also lack a dedicated animator), they're focusing on developing more of the game itself, first.

From early Kickstarter days on, a primary criticism has been that it looked more like engine components than an actual game. As Bare Mettle was planning to work on Sui Generis for an indefinite period of time, I'm sure they'll revisit it eventually.

Another important reason is that Sui Generis actually isn't all that combat-oriented. The final game will feature a lot of taking your time to explore the world and uncover all kinds of plot details. Combat is a part of it, but given the role it plays in the full game there's already an unreal amount of time dedicated to its current state.

I'm not sure how much information of the insider forums I'm actually allowed to share, but Bare Mettle has been looking at stances regardless. If I'm not allowed to share information from there, I'll promptly quit. But do consider pledging for insider forums access, as that's where a lot of feedback and developer replies regarding early versions of the game currently are.

Regarding the main problem, one final thing I'd like to add - as videos can't show it - the AI does a very good job at making combatants believable. Opponents step back from your swings and parry and counterstrike accordingly, and it's very much like playing against another human player, which I felt was impressive and worth mentioning.

In the end, Sui Generis is still a game, and one BM intends to make as fun as possible, and in its current state it actually is a lot of fun to play! Even with the current lack of combat stances, people have reported feeling as if they're truly improving at the game, something lacking in a lot of games that are actually very combat-oriented, which are more like the press-button-to-flourish games.

I think we'll see improvement eventually, but given the other priorities and that Bare Mettle only has one programmer, not on the short term.

(as a post summary: stances are being considered, BM's currently focusing on the prelude and broadening gameplay, combat gameplay/enemy AI makes up for perceived animation issues in my opinion, and pledging for insider forum access is very much worth it)

 

[Empire²]

In that sense driving F1 is also choreographed... It looks like that, but once one starts practicing, one realizes how many thoughts and decisions go through beginners mind. On the contrary, masters focus on strategy, and actions are performed accordingly and automatically.

I think you understood my statement wrong. The misconception people make is that a duel between two knights is straight from a textbook. While this is not wrong for swordfighting in general, it's something more related to fencing as opposed to things taught in, e.g., the German School of Swordsmanship.

Even in F1, there are procedures and rules set in stone for drivers to follow to the best of their abilities while in a race. These include the manners of passing other drivers, making way for others, going to the pits, etc. Failing to abide by these ground rules will have consequences (penalties or even disqualification), but it doesn't mean the entire sport is built upon said procedures. The same goes for fighting. Even though there have been general conceptions about combat since about the start of recorded history, regarding offense, defense, stance, movement, general tactics & strategies and more, these are merely a framework through which a fight develops.

A back-and-forth between you and your adversary is not merely a thing of fiction. However, between knowing how to attack and defend, there is space open for an unaccountable factor; the chaos. Rarely will you come across someone so disciplined and trained that they take not a moment's notice to assess the fact that you have the upper hand. As such, the human survival instinct comes into frame. It's been proven time and time again that, when cornered, people are the most unpredictable they'll ever be. This means that your textbooks, years of training and studies, in a moment of utmost danger, are irrelevant. It means that you'll literally risk life to avoid death.

Sure, you can say there's modern military outfits that train their units in a psychological manner, but "The Horrors of War" are more than good titles for mini-documentaries. It's about the aftershock that comes with the mental pummelling, that, for example, Navy SEAL members have to sustain during their extensive training. They are devoid of morality and human instinct in a way that echoes through until they are dismissed. That's when their primal instincts, most notably fear, set in. A lot of war veterans end up with severe psychological damage, done to them through having their superiors literally drill their humanity out of them.

As long-winded as that is for an explanation of my statement, it still doesn't capture the point I was trying to make, which is one that you, being new on the forums, may not be as familiar with as some of us here; the discussions regarding realism & gameplay, and sacrificing one for the other.

There are a handful of people here that are quite knowledgeable about history and combat. Elric has spent a lot of time sharing with us his wisdom and inside knowledge on medieval combat, answering many of our, as well as some of the developers', questions about certain matters.

I don't have that knowledge, as much as I would like to tell you that I do. I can tell you that Herostratus reduced the Temple of Artemis to ashes in an attempt to seek his fame, thus coining the term "Herostratic Fame", but I don't know a damn thing about blacksmithing, let alone swinging a sword. I bought a katana once, because it looks awesome. What I do know a lot about, though, is game design. The most important thing about a game, especially an RPG, is that you need to realise you're going to sway closely to the line between a game and a simulator. They might not seem far apart, and the differences between the two may seem like they're insignificant enough that they're synonymous, but the problem sets in with realism.

A sizeable amount of people have had their criticisms at how Sui Generis' combat and game world were being handled. Some of these people thought it wasn't realistic enough. "The combat feels like a flash ragdoll game.", and statements along those lines. They were convinced the game needed more control. When you hand over more and more control to a game, the rules become linear and rigid, however open-ended the game may be.

By placing control in the hands of a player, you turn it from being a simulation into a game, because a player can enforce one thing that a computer cannot; chaos. Imagine fighting a tough opponent. You're almost at the end of that damn dungeon you've been at for a few hours now, dying over and over to its denizens. But now you think you've got it. You got a few good hits on your opponent, and decide to take the risk of going for a big swing to finish them off. You expect the opponent to defend, but they don't. As opposed to moving to protect their life, they move to take yours. And it's a big hit. You're down to about 20% health now. No potions, you're in for it now.

At this point, you're going to be weighing your decisions. Do you parry this blow and go for your opponent? Or do you take the risk to evade their attack entirely, and create an opening in their defence?

That's where the game literally throws all control into your lap. You might die. You might live. But at least the game left it up to you. Perhaps you didn't realise that. You had no clue the game was going to make you scramble for the right key to press. And miss. That's what separates a simulation from a game. A simulation has a beginning and an end. It goes from A to Z. A game does not. Instead, a game takes its own path. Not because of the mechanics, the AI or the world, but because they left it up to you. You were in charge. That makes failures your failures, but it also makes accomplishments your accomplishments, and that's what makes games memorable.

And in the end, I'd rather have one memorable experience than ten realistic ones.

 

[Chris.J0621]

So many places to post, I think I'll just post my thoughts here: while I absolutely love the combat here, may I request a little tuning to the swinging control? What I mean is, in order to swing from right to left (typical swing) I can click anywhere to the right of, or even on the enemy to make this swing. In order to swing from left to right, however, I must move the mouse pointer so far to the left of my opponent (his right, my left) that my character turns to face the pointer. The end result is me missing my target more often than because I'm turning so far to make this swing, when I turn back to face my opponent I end up looking like a drunken baboon. If this has been discussed already, I apologize.

Now, before you jump the gun, yes I know you can swing from left to right through holding down the LMB but that starts the combo with a right to left swing. I'm talking about starting the first swing from left to right. All I'm proposing is tuning the mouse pointer so I can don't have to completely turn my character in order to swing over my left shoulder.

Also, if anyone has any tips as far as each weapons strategy (namely, the zweihander and the sword and board combo) feel free to school me. I managed to get to expert arena, and if I beat the first guy, the second fight (sword and board lady in red) almost always whoops my ass. If I beat her, the third guy always kills me. (The first guy in plate you fight) Thanks in advance!

Also, in response to the first gentleman's post at the very end about taking one step in synch with a swing, as opposed to say three steps: I very much agree this should be implemented, because it is the logical thing to do, however implementing this fluidly without sacrificing too much player control could be tricky.

 

[lvk]

@Chris.J0621 With the two-handed swords against sword and shield opponents, initiate an overhead strike barely in their range. When they step in to attack you, step back and you should be hitting their heads pretty hard. Crouch mid-swing for maximum momentum.

 

[Chris.J0621]

@lvk What about parrying in general? I seem to be able to block attacks all day long with a shield, especially the bigger ones (for obvious reasons) but it seems two handed weapons, namely the swords, are hit or miss. Regardless of how hard I try to position myself properly, attacks from the enemy's left to right seem to give no f's about my attempts to parry,lol. Do you record a lot of videos or know someone who does? I'd like to observe how y'all fight. Also, thanks! I'll try that!

 

[Crayfish]

Also, if anyone has any tips as far as each weapons strategy (namely, the zweihander and the sword and board combo) feel free to school me.

I find the halberd the easiest weapon to master for the expert arena, it does very good damage against armour compared to swords. Keep out of range of your opponents weapon and land as many blows as possible. Three quarters of your attacks will be blocked but the other quarter will do the business. It can be very hard to keep out of range of the shield users as they will push you back to a wall, just try to keep circling them and out of range of their lunges.

 

[lvk]

@Chris.J0621 The biggest thing about parrying with a weapon alone is that your weapon must be directly in the way of the attack to stop it. If you're facing your opponent directly, chances are the attack will glance off your weapon and carry enough force to deal damage to you (though even failed parry attempts will net you reduced damage, as long as you weren't attacking). If the attack's coming in from the left, you'll have to turn a bit to the left as well. Additionally, you could step into the swing so there's less room for the opponent to make a full swing.

The cool thing about strategizing in Sui Generis is how combat is really about timing, swings and the force behind them. You can really focus on what's happening in the moment, and how to react to it.

Gameplay videos are all over the forums, and difficult to locate due to them being unlisted on Youtube by NDA requirements. I'm pretty sure @Tony uploaded a few at least, I hope he can link them.

 

[Chris.J0621]

I had noticed "attempted parries" dealing reduced damage, I will have to be more deliberate in the future attempting to parry. I've been trying out all the weapons I come across and so far the zweihander is my favorite, however sword and board, followed by the mace/morningstar are a close second. I find the halberd to be very useful, but situationally so, for me at least. I had noticed as well bumping into your opponent stops most of the force from their swing/creates an awkward angle for striking, however my only issue with this is then i create no opportunity for counter-strike.

I've been practicing overheads too and , as mentioned, they are incredibly effective against sword and shield wielders. Has anyone mastered the art of the riposte? I can nail a few ripostes here are there, but ultimately more often than not the combo swing usually nails me before I can land the riposte. Also, footwork, how do you deal with the more aggressive opponents constantly coming at you, and are there any tips you can share (obviously I'm trying to stay within my weapons range, and out of theirs)? At first I was having trouble with the shield users, but overheads make quick work of them (when I can land them). My other issue then lies with the expert two handed warhammer. That guy... ugh. lol

Thank you all for your input!

 

[Tony]

@Chris.J0621, I made a short video to demonstrate some different tactics which can be used in combat. Watch my mouse cursor to see how I perform different attacks; it doesn't require much movement to parry or to perform different attacks and it is quite possible to execute precise attacks and movements.

In the video I'm using the starter two-handed sword with no armor or clothing to show that what I'm doing is not gear dependent. I spent additional time fighting the fourth opponent (guy with a hammer) to demonstrate how to parry. Since his hammer has such a long reach it helps to dodge either backwards or to the side to get out of range of his attacks while attempting to parry. I also tried to demonstrate how to use ducking to attack underneath an opponent's shield. I use this technique on both opponents who are wielding shields in the video.

I hope this helps! If you have any additional questions feel free to ask. I have many videos of me playing Sui Generis that I can upload if you want me to demonstrate anything else.

(Edit: Video removed since this is not Insider forum and game is still under NDA. I didn't realize this was in the general forum)

 

[Blackbeard]

Thank you all for warm welcome and your replies
To be honest, I expected fair arguments like these (all true):

• I'm new here and I'm not familiar with animator's job
• developers and insiders had already covered these topics (and discussed these issues) before I showed up here
• there are people in the team / on the forum that know a lot about medieval combat (presumably more than me)

From that point of view, anything I say is a disadvantage for me, right from the start, I admit... In the worst scenario, with my suggestions, I could be tacitly labeled as ignorant. I am aware of that.


On the other hand, as a person observing from neutral point of view and with fresh brainstorming, I made a conclusion (systematically elaborated above) that what combat modeling in Sui Generis needs is (sketched now in one succinct sentence):

Synchronization between moving parts in such way that, at certain time point, there is a stance as a result of coordinated movement of the moving parts.​

More than a year of programming of a game mod taught me that a small change in point of view could at first seem like a leap of faith for the concept. However, after implementation, such changes were usually a major breakthrough.


@Tony thank you so much for posting that video, because it is the most fresh illustration of my point. I made an excerpt from "The Terminology and Execution of Steps" (for further reading check spoiler below), but I will summarize it in my own words:

Spoiler

Passing Step
The rear foot passes to completely in front of the leading foot in the manner similar to a walking step.

Chasing Step
This step consists of two motions. In the first motion, bring the rear foot forward until it is even (or nearly so) with the front foot; in the second motion, step forward with the front foot. [...] Since the Chasing Step covers roughly the same distance as the Passing Step, it is often used in a situation where you would use a Passing Step but the ‘wrong’ foot is forward.

Half Step
The half step is actually the name for two different but related steps:

• Increasing Step or Accrescimento -- alternatively, this could be the second motion of a Chasing Step or Changing Step
• Gathering Step -- alternatively, this is the first motion of the Chasing Step and the Changing Step

If one is attacking with right-to-left diagonal cut, than :

• forward step with right foot, or
• backward step with left foot

is needed to have a stable posture.​

If multiple steps are noticeable in gameplay videos during one single swing of the weapon, than some issues are still not discussed in proper manner...

The most important thing about a game, especially an RPG, is that you need to realise you're going to sway closely to the line between a game and a simulator. They might not seem far apart, and the differences between the two may seem like they're insignificant enough that they're synonymous, but the problem sets in with realism. [...]
"The combat feels like a flash ragdoll game.", and statements along those lines. They were convinced the game needed more control. When you hand over more and more control to a game, the rules become linear and rigid, however open-ended the game may be.

With my bold statement I am really at risk again to sound ignorant and overweening, but I firmly and deeply believe that combat in Sui Generis needs only that one small tweak to be more realistic. Of course not to become a simulator (separate control for every single action), but to remain a fun game as it already is.



=========================================================
@Empire² thank you for such a rich reply. I understand your point of view and why you felt misunderstood after my comment. I cannot agree with some of your arguments since I have different point of view. Here it is...

Yes, there are rules in F1 and you stated them quite correctly. Those rules serve to avoid chaos. However I speak about technique. A driving technique in this case. I was professional swimmer for over 20 years and now I am doing martial arts for fun. So my experience tells me that training serves to minimize the chaos, exercise the movements till they become automatic / autonomous, exercise the mind and the spirit. It is a road to perfection. It's the same for any work, art or sport -- training. You'd agree, wouldn't you?

Actually, what is presented in any swordsmanship textbook is paradoxically both a starting position and a goal! A starting position for training and a goal for sparring/ combat. During training one does it slowly and systematically, and during sparring one simulates combat.

This means that your textbooks, years of training and studies, in a moment of utmost danger, are irrelevant. It means that you'll literally risk life to avoid death.

It's quite the opposite! Yes, life is at risk. Therefore, there should be no space for chaos and unpredictable events! That's the purpose of the training.

With everything else I agree and you wrote wonderful reply. I especially like how you described trilling fighting moments!

=========================================================

Another important reason is that Sui Generis actually isn't all that combat-oriented.

It's very much like playing against another human player.

Sui Generis is still a game, and one BM intends to make as fun as possible, and in its current state it actually is a lot of fun to play! Even with the current lack of combat stances, people have reported feeling as if they're truly improving at the game, something lacking in a lot of games that are actually very combat-oriented, which are more like the press-button-to-flourish games.

That's completely in line with my thoughts. With such a small tweak both human characters and AI characters would be more believable in combat, which consequently increases the fun!

=========================================================

... and hopefully it's of some help to the devs (though I suspect they already knew most of this).

I'm just worried about what is technologically possible and what can be done by such a small team that doesn't have a dedicated animator.

And since developers accomplished so much so far, that's the reason why I believe such improvements are possible!

 

[Tony]

@Blackbeard, I removed the video since this thread is not in the Insider forum and I didn't realize that when I originally posted it. I really need to stop posting things on the forum when I haven't slept because this is the second time I've made that mistake in a 24 hour period . I suppose you got a sneak peak of something you weren't supposed to see since the alpha and beta versions are still under NDA.

 

[Blackbeard]

According to what I wrote above

...multiple steps are noticeable in gameplay videos during one single swing of the weapon...

two simple solutions would be to:

1. Prevent character from being on the wrong foot while attacking
2. Prevent character from taking multiple steps while attacking/ swinging weapon

These two rules should go on top of existing rules/ combat system i.e. to override other footwork rules.

Thus, programming solution might look like this:

while {event_LMB == true} do //LMB is pressed
{
    //Get/set values
    _attack = 1; //sets flag (handle) to signal that attack has started; values: 0- no attack, 1- attack, 2- attack chain
    swing_direction = getSwing player; //getSwing function converts attack direction to 1 of 3 categories - "straight", "left", "right"
    movement_direction = getMov player; //getMov function takes movement vectors and covert them to 1 of 3 categories "standing", "forward", "backward"
   
    if (swing_direction == "right") then
        {
            if (movement_direction == "backward") then
                {
                    _moveFoot setStep = "leftBackward"; //setStep function sets flag for the 'stepping animation' i.e. footwork
                }
            else    //character is already moving forward or standing still
                {
                    _moveFoot setStep = "rightForward";
                };
        }
    elseif (swing_direction == "left") then
        {
            if (movement_direction == "backward") then
                {
                    _moveFoot setStep = "rightBackward";
                }
            else
                {
                    _moveFoot setStep = "leftForward";
                };
        }   
    else
        {
            chooseFoot = random 10; //random real (floating point) value from 0 (inclusive) to some X (not inclusive) 
                                    //randomly choose foot if character is performing overhead strike or thrust
                                    //move forward and cut or thrust
            if (chooseFoot <= 4) then {_moveFoot setStep = "rightForward"} else {_moveFoot setStep = "leftForward"};
        };
       
    waitUntil {_attack == 2}; //waits for the next attack in the sequence (external flag/ signal from other script or function)
                              //notice that script is now again acquiring values for swing_direction and movement_direction providing LMB is still pressed
};

_attack = 0; //sets flag back to 'no attack' -- LMB is released

Since character progression system should reflect experience, a likelihood for using these two rules might be implemented like this: master fighters should be more likely to follow these rules (95% chance) during fight than novice fighters (5% chance).

If I come up with more elegant/ different solution, I'll post it.

@Tony maybe I was not supposed to get that sneak peek, but it surely helped me more clearly draft ideas I have maybe that short sneak peek will serve some greater good, who knows...

 

[Madoc]

The reason why this sort of stuff is difficult to do in practice is because all leg/foot movement is entirely procedural and under the player's control. No action besides an explicit step input from the player should cause the character to move and as it is just shuffling the feet would prevent the player from taking a step because the feet need to physically reposition themselves in a way that allows the character to move, the character does not slide around independently of what the feet are doing. As it is the system deals with the character stepping and turning at the same time, this works because no "animations" interfere with the positioning of the feet. The system could possibly be extended to include some repositioning of the feet based on actions but it's by no means simple to do, as I think I've mentioned before, the footwork, though apparently simple, is the single thing I've spent the most development time on, it's surprisingly difficult to get consistently functional and fluid behaviour from it. Either way because of how the game is played we would not want the character to move when performing a swing but only as a result of an explicit player input to move. We've always prioritised very explicit and direct control over appearance, it's a very different type of game from one where you execute precanned actions at the press of a button.

Just giving some context here, improving the footwork is something we're keen on (and we do have specific plans to improve it already) but there are important constraints and goals that need to be considered.

 

[Blackbeard]

@Madoc thank you for your reply.
A friend of mine from MIT informs me from time to time about programming robots, which is done there, so I am familiar with procedural programming. In context of leg movement, mathematical formulas and algorithms are used in real-time with no a priori motion data, but utilizing empirical and bio-mechanical concepts, to give rise to flexible and adaptive movement. That's a huge beast (Boston BigDog) you are "hunting down" and I admire you a lot!!!

What I suggested involves no "animation interference" or "precanned actions at the press of a button" (quoting to avoid retelling).

We've always prioritised they very explicit and direct control over appearance...

I agree and my suggestions are concordant with yours:

• restriction of the extensive footwork while swinging, and
• avoiding being on the wrong foot

are actually sings of good, direct control, not appearance ("cosmetics"). Please allow me to illustrate, in systematic way, through this example:

• my character has right foot in front of left
• if I position mouse cursor so that my character attacks with cut from the left to the right, and I press LMB,
• should feet wait for the separate input from keyboard in order to shuffle/ reposition (formal term Changing Step)

or​

• such mouse input should also influence footwork ?

So we agree, don't we?

This video is one year old, however that's the best example how swings do not cause the movement of the feet and my major point in this entire thread is that they should! Even small steps would be better and they would resemble combat stance, rather than feet being nailed to the ground in the absence of keyboard input...

What could be a feasible solution? ... The simplest one would be checking which foot is in front and doing forward (default option) or backward step (if there is overriding input from user) to put appropriate foot in front and then stopping the footwork until next attack comes (like chaining attacks). This should generate believable fighting movement.

In above example, upon pressing LMB, from parallel feet stance, it would be like player briefly pressed W to move forward (or S to move backward), but you have deliberately chosen which foot, left or right, will make a step -- Madoc, I have a feeling you have all components and tools to do it...​

...the footwork, though apparently simple, is the single thing I've spent the most development time on, it's surprisingly difficult to get consistently functional and fluid behaviour from it..

Therefore, I will try to help as much as I can. If I come up with some fresh ideas or solution for this, I'll post it here, you can be sure of that. It is really an amazing work you are doing!!!

 

[Vold]

Personally, I think that the leg/footwork is beyond good enough. It feels very responsive to external input (WASD) which adds to the combat an additional layer of complexity (in a good way). The responsiveness would be largely affected by additional movements triggered with internal input, and to what end?. Also, this probably would demand plenty of time and as the devs have stated in this thread (Status Update), they are now focused on higher priority aspects of the game; so I'd let it go for now.

 

[Blackbeard]

so I'd let it go for now.

Oh, it's so nice from you to formulate it in such a polite way. Thank you. I only try to help...

Your reply implies that I am suggesting something that is complicated, time consuming, that will ruin responsiveness and that I mock already good footwork. After everything I wrote before to systematically and clearly elaborate my suggestions, and brainstorm about solution -- that's kind of immature reply...

I only noticed that solely attacking/ swinging weapon is not associated with footwork, and that, when standing still, character looks like "driving away flies" and not like actually, physically cutting/ striking. And I think it is visible through all layers of complexity that you mentioned.

 

[Vold]

Oh, it's so nice from you to formulate it in such a polite way. Thank you. I only try to help...

Your reply implies that I am suggesting something that is complicated, time consuming, that will ruin responsiveness and that I mock already good footwork. After everything I wrote before to systematically and clearly elaborate my suggestions, and brainstorm about solution -- that's kind of immature reply...

I only noticed that solely attacking/ swinging weapon is not associated with footwork, and that, when standing still, character looks like "driving away flies" and not like actually, physically cutting/ striking. And I think it is visible through all layers of complexity that you mentioned.

If my reply made you feel that way, I apologize. I think that your observation is valid.

Actually, I am also looking forward to future improvements of the animation in general. Nevertheless, there are other aspects of the game that draw my attention even more.

 

[calithlin]

I am intrigued by your thorough and detailed breakdown of your suggested changes, @Blackbeard. At least with my own programming background I was able to follow the logic of your suggestion quite readily given your pseudocode example. It seemed to possibly function similar to some node based AI I had done at one point to program robots in formations for a soccer-like game simulation, where given any starting position multiple robots could manuever themselves over time into a predefined grouping shape to carry out a 'play' - only in this case the nodes would be joints in a body, rather than robots on a soccer field.

What I am most fascinated by is the possibility of phasing in and out the likelihood of properly flowing into stances based on the level of skill. On one hand it may go against an idea of BareMettle to have the player always have a clear understanding of how their attacks will act, but on the other hand, players will be still performing attacks only on input, and will slowly come to expect how their attacks will fall into the proper stance given higher levels of their character. (This would also be a huge extra factor for creating a wider range of enemy difficulty - such as having a powerful, but sloppy barbarian, or a weak, but precise aged swordsmaster).

While I think these improvements may need to be rightly shelved until other aspects of the game such as character interaction and abilities, progression, etc have been further tackled, I am hoping Madoc might take this concept into consideration once again if or when player movement mechanics comes back into the scope of feature tweaks/updates.