Designing for melee

I'm currently messing with a prototype for a new game that is very focused on medieval melee combat. I'm curious what sorts of systems people have seen, and what they like and don't like about them.

In my original design brainstorm I quickly ran out of inputs to set up an accurate melee system. Actively dealing with movement and targeting strikes and shield positioning doesn't seem feasible with the average number of human fingers.

Systems that come to mind for me include:

• Diablo, simply click on an enemy and let the game do the rest
• Assassins Creed, melee combat is based primarily on accurate timing of hitting the action buttons.
• Dark Souls, combat consists of a series of moves. Each weapon allows for different move animations. Movement and positioning is important. Shield use is significant.

In all of these games combat is still fairly simplistic. I'm trying for a system that had the player actively positioning their shield, performing feints to hide a real attack, choosing target location and so forth.

Thoughts.

 

I have not seen a 3D game using this approach although there may be some out there. Not sure if you are looking at a Diablo viewpoint, FPS viewpoint, fighting game or something else. That may help for coming up with a control system.

The first game that came to my mind was the old Gladiator I used to play in the arcades long ago. It has the targeting for attacks and blocks you mentioned along with a cool idea of applying that to the armor of player and enemies.

 

More questions first:
- first or third person?
- input device options?
- desired aesthetic? E.g. closer to Minecraft or Dark Souls?
- how high do you want the skill-ceiling to be?
- singleplayer or possibly also multiplayer at some point?

 

It sounds like you should be looking at at fighting games rather than RPGs, for the level of detail you want. These of course are based on rather complex button combos, with timing and distance being important factors on how they play out.

I was a big fan of Soul Calibur at one time; Bayonetta and God of War are more adventurish combat games that I haven't played but which, I hear, do a pretty good job of this. Perhaps they can provide some inspiration.

 

War of the Roses is kinda like that. the animations were sluggish though. It could defiantly be improved upon. if you could make the game play smoother I think you got a winner.

 

You may want to look into some martial arts / fencing material to get a fix on what kind of portrayal of combat you want.


There's plenty of theory you can dig into. For instance, there are only 8 cuts in using a sword, and 4 of them are reversals of the opposing 4. (9 total if you count the straight stab.) You could use this for shield positioning. Choosing a defensive posture along the right line would be a perfect block.

You might go the elegant fencing route and emphasize "flow" where each cut leads into a choice of a couple others. This would allow for an "attack" to evolve over several steps, resulting in the shifting of "advantage" from one opponent to the other based on their choices in cuts and defense posture. When the attacker over-extends himself, he leaves himself open to having the advantage taken from him. If the defender fails to monopolize on a successful defense, he will forever be defending until he is overwhelmed.

You could look at boxing too and take elements from there- I understand it can get really technical. There are punches to limit mobility, there are punches to disorient, and there are big haymaker punches to go for the win, and there are series of punches to pin an opponent into a corner and overwhelm defenses. Striking at the body changes the fight differently than striking the head, keeping the ability to move around and maintaining defense properly are critical elements- heck, breathing is a critical element of combat.

 

Sounds obvious now that you say it. I think I need to go do some research on what the modern fighting game looks like.

Still very early in the design stage, so the answers are somewhat nebulous.

• Currently first person. I'm open to change if third person suits better.
• For input I'm initially looking at mouse and keyboard. The other option would be an xbox 360 controller.
• Its going to be closer to minecraft then dark souls.
• I'm looking for a medium skill level. I want the game to be beatable by your average gamer. I would expect players to have to learn the system, but once they have learned the system it should be fairly straight forward.
• Single player. Maybe if I get really ambitious I'll go to split screen multiplayer.

The overall picture will be something like StarWars Battle front in medieval times. Half a dozen weapon combinations/roles that players can choose. A large number of agents on the field. Something resembling a medieval pitched battle.

But that's the large scale ridiculous dream. At this stage I'm simply trying to build an engaging combat system. I want to be able to get reasonable fighting between two sword and board characters. If that works then the rest of the game is theoretically viable.

 

Get yourself out of theory and into practise as soon as you possibly can with a movement prototype and target dummy. Fun isn't found on paper.

 

Working on it as we speak. I'll throw up an early prototype as its done. It will make my ideas much clearer.

 

Sounds good. Arkham-style combat is a pretty proven (maybe overused) approach to melee combat. But I haven't seen it done in first person perspective so far. I'd find that approach interesting since it theoretically allows for pretty cool stuff to happen on screen without overly complicated input schemes. And it would be less inventing the wheel and more trying to make it work with first person perspective. However it's always a possibility it has never been done because it just doesn't work well. In that case I'd rather stick with first person, because 3rd person melee has been done more often. Dying Light's melee combat wasn't bad, but probably a bit too simple for what you are looking for. If I remember correctly it did convey the feeling of impact and struggle pretty well. Imho getting that feeling right and e.g. not let swords clip through armor etc. would be important to stand out from the crowd, because barely any game managed that. I only played one of the medieval FPS games very briefly on a free weekend (the one on source engine, forgot the name) and I didn't like the feeling of it. It's probably good to try all of those out and see what you like and dislike. You should check this out too: http://store.steampowered.com/app/362490/
I haven't played it, but afaik they where kind of going for "physically simulated Diablo". Sounded unique and might be worth a look. Don't think the concept really worked out too well though.

What kind of animations can you get your hands on, might be one of the most important questions because it could limit the viable approaches.

 

Did I miss it, or has no one mentioned gestural combat like Die by the Sword? Chivalry and Mount & Blade also do something similar, but they're not quite as all-in with the 1:1 player:character interface.

If you're willing to entertain alternative interfaces, there's always Leap Motion, Kinect, Wii controllers, etc.

 

I would definitely check out the absolutely excellent Mount & Blade. It's got a pretty fluid mouse and keyboard combat system with a surprising amount of depth and skill, while also not requiring absurd amounts of animation and IK.

It's a little dated, but probably one of the best games ever.

In my experience, the major problem with melee combat is that once you up the scale even remotely, you start to have major problems with clumping. Melee combat is not really fun if its just like a mob of 30 guys bumping into each other while clubbing each other to death.

I think this 'clumping' problem is one of the real challenges with melee combat.

 

I have some vague ideas on the clumping issue. But it is something to keep in mind. I've participated in battles where two shield lines simply stand and look at each other for the duration. Not the most exciting activity.

It's something I'll tackle after I've got satisfying single combat working.

 

In Mad Max the enemies pretty much take turns attacking you. It's like you have one AI directing the battle, not a dozen isolated AIs that don't communicate. Makes for a bit of awkwardness at times (as much as in most movie fight scenes with one hero and many more opponents), but solves the clumping issue pretty well imho. Shadow of Mordor was similar, afair.

 

I'm expecting to end up with about even numbers on each side for most scenarios, so no uber strong hero taking on the world single handedly.

I haven't gone through AI yet. I'd love an agent based one. But a overlord of some sort is probably a good idea.

 

This video shows a pretty good example of "the clumping problem" in the first minute:

I'm really not sure how you would go about designing around this if you want good scale melee.

This kind of thing is the main challenge of dealing with melee. In terms of 'strategic gameplay' you counter balance the strength of clumped units with stuff like AoE to give incentive to spread out. There are other difficulties though when you start dealing with just larger scale stuff.

I'd love to hear what some of your thoughts are in terms of how to deal with this kind of thing. I'm not entirely pleased with my ideas for it in my own project.

 

I'm going to make shields very significant. My hope is that this will lead to the formation of lines rather then clumps. Worst case scenario I simply force the AI to take this behaviour. It's cheating, but hey, we are making games.

I'm thinking a paper scissors rock thing. Spears beat shields, shields beat archers, archer beat spears. There will be a couple of other roles to mix thing up a little.

I also want the player to have some control of the strategic layer, while at the same time simulating the fact that they are leading a bunch of disparate forces with their own captains and motives. No idea how I'm going to do that. Might end up beyond my abilities.

 

I'm excited man, this sounds like a big step up from pond wars! Good for you!

 

Another source of inspiration for your game could be Skyward Sword. Its combat system isn't overly complicated, but the reliance it has on attack direction does add an element of strategy I haven't seen in other games.

Of course, you then would face the problem of how to bring those mechanics into your game. After all, most computers don't come with a Wiimote.

One solution could be to use mouse movement. The player holds down a button, jerks the mouse up, and wala! The character also swings up. If the mouse moves diagonally right, the character also swings diagonally right.

If you didn't want to use the mouse for combat, you also could use arrow keys. Hold down the button to ready an attack, then press one of the arrow keys.

 

Yeah. At my current rate it's probably going to take me years. And I won't get half of what I've said in there. Still it's been fun so far.

 

Two suggestions:

#1 - Constant play testing. Never go for more than a month without having people play your game, even if it's a mess.
#2 - Don't worry about the strategy layer at all until the tactical layer is fun.

 

Wow! You're designing a melee game?
Makes sense now.
That would explain, why so many of your avatar pics, have you and others, dressed like Norman warriors,
holding many different weapons.

I'll try and help, as best as I could.

PS: Are weapon wielding Goblins, and evil nasty stinky smelly Orcs etc. in your game too?

 

I'm rooting for BM's melee game.

 

Its the other way around. I and my friends currently dress as warriors, and fight one another. The design was an attempt to capture the combat experience in a video game. Its been peculating in my head for a while, but I've just recently started to put pen to paper and open up the editor.

The current pic is Siconze, it supposed to be Italian renaissance themed. I suppose you could call my the Guardians (previous warband) approximately Norman.

Assuming I can get the IP rights. I play board games with this guy, so I might approach him. I've even though about doing some motion capture with a few of the guys. But characters and story and strategy is way beyond my current scope.

Thanks!

At this stage I've gone with mouse and keyboard controls. The keyboard will control movement in a similar style to early 3D games like Doom. The mouse will control striking location of the sword. I loose a few things here, no looking up and down, strafing, and the other conveniences from modern FPS games. I think I can live without them. I'm probably just going to do a couple of button presses to control the shield.

One advantage is the same control scheme will work for archers and spears and mages and healers. Duel wielding is just going to have to suck up being nerfed.

There will be a playable prototype soon*.

*You can currently play ten pin bowling by throwing your sword at the other guy. Its fun, but not the game I'm designing.

 

I think you can get away without looking up/down and jumping, but I really wouldn't mess with the WASD default for walking and strafing. Imho that would be a guarantee for angry reviews and unhappy players. "Conveniences" is something people get used to und feel entitled to.

What I could see working is something like this:


default mode:
wasd: forward, backward, strafe left, strafe right
mouse: look around/turn
space (hold): raise shield to block (and optional auto face enemy?)
ctrl: quickly drop to crouched position for lower strikes or to duck head-high swings, but getting back up should be slower (even more so the more armor you wear)


q: mode toggle between default and defensive + lock faced target

shift: sprint
shift + space: sprinting shield charge to ram enemies, but blocks most of your forward vision (maybe look a bit to the bottom right while doing that) and limits turn speed
ctrl+space: cower behind shield. An attacker with shield charge would trip over you and it is a good defense against arrows, but you see nothing in front of you.

defensive mode:
w: slow advance
s: jump back
a: circle strafe locked target left
d: circle strafe locked target right
space + w: shield bash upwards
space + a: shield bash left
space + d: shield bash front (maybe with the edge of the shield?)
space + s: ?


attacks:
left click: quick strike/thrust attack, contextual based on position of center of screen to engaged enemy
left click (hold): medium slash attack, contextual based on position of center of screen to engaged enemy
right click (hold) + move mouse right/left/up/down: prepare heavy strike, and either execute it by moving the mouse back in the opposite direction or release right mouse button to abort the attack. Range and timing influence force of attack.
right click (tap): I don't know that many ways of attacking with a sword, maybe a quick strike with the hilt of the sword? Or a contextual attack with a jump forward?

f: kick (to counter shield defense/charges and stagger enemies. Kicking a shield charge should stagger both.)
e: use/pickup
r: could switch between different attack styles (see slow/medium/fast combat styles in old Jedi Knight games after the first one)


Just some thoughts, use them as you like . I hope this scheme should map well to a controller, offers enough complexity and isn't too crazy to learn.

 

I'll tell you what I like, the last one, Dark Souls. Nobody wants to click on enemies and let the game do the rest, because after a while they get bored. Really bored. Believe me, I tried it.
The Assasind Creed type game sounds good, but if allyou are going to do is click on the buttons that pop on the screen, you get too addicted, or too bored. You become bad at playing other games because they are not like AC. Then you scrap the games that made you a hero, and play one single game, for your entire life.
The best option is the Dark Souls type, it keeps your mind occupied, and you can also think in a 360 degree way.

Anyway, good luck.

 

@BoredMormon: I think you need to watch this:

 

I actually spent a lot of time looking at authentic sword fighting. HEMA and stuff, swordfish, etc.

Trying to do anything remotely realistic is way, way outside of reach. That video sounds pretty awesome, and just looking at the game play footage it seems like they really did a good job. That said, doing anything remotely like this is sooo far out of scope for any kind of small indie it's almost not really worth thinking about.

Especially the fact that fights should really generally take like 2 seconds ;D

 

That was pretty cool. I might just buy their game instead.

There approach was interesting. There are some design areas I want to do differently. They mentioned providing for two different basic attacks, then allowing combining with further well timed attacks. I'm hoping to go for a more free form approach, my combat experience leads to more oppurtunistic strikes, rather then set moves and combos.

On the other hand my graphics requirements are much lower. Something blocky like Minecraft or Crossy Road will do the job for me.

My physics will also be simpler, though it's still pretty complex. I'm only going to use two unique hit zones. But I do want to be able to do things like push a spear out of the way with a shield.

 

Too much effort will be put for a mid-scale game BoredMoron is making according to the vedio. But, always, you can try it if you are going to turn the game into full-scale.

 

Update

So swinging swords around worked. It doesn't quite feel natural yet, but it does feels like it will work with a bunch of tweaks to all of the numbers.

On the other hand:

This is pretty much spot on. Nerfing strafing makes the movement feel really clunky. To the point of being unplayable. You can only approach a target from straight on, simple maneuvers like circling around an enemy are impossible without turning your back. Right or wrong years of FPS training won't let me ditch it.

So as much as I like the idea of using the mouse to target individual spots on an enemy to target, I'm probably going to tie it back to the more traditional FPS job of looking around.

Its not much to look at yet, but you can see what I've been playing with here.

 

Glad I was right ^^. However I don't quite understand what you think you'll lose by going classic wasd + mouselook. Can't you simply use the center of the screen for your aiming point. As it currently is implemented, I don't fully understand how the swinging/aiming system works.

 

I'll give it a go. You are probably right. I had this vision in my head of being able to choose exactly where to strike independent of view. Now that I've prototyped it out, going with the standard FPS view might be better. I'll make a branch and try it out.

The idea is that the sword returns to a resting position when you aren't actively swinging it. When you click the mouse button the sword swings towards where you are aiming, then returns to rest. I'm going to tweak the values a bit and put in several more rest positions. The general idea is that strikes need to be timed.

There is a game in there somewhere. I just haven't found it yet.

 

Getting this to feel good with physics based swing animations sounds like quite the nightmare to me. If timing is important to you, how about mecanim based animations where you have small time windows to chain swings in a certain way? And when you land a hit you can switch to physics based bounce back animations or one of several premade ones based on parameters of the impact. You could make the sword animation position relative to the view direction (within sensible restrictions) and let the player use the mouselook to aim the strike that way. But more like "this strike has a fixed arc that it follows, and I'll position my view to hit the opponents head" instead of "the strike always goes to center of screen, so I'll click on the opponents head". Imho that also works much better than trying to click on a moving target at close range. And you can still make some adjustments while the strike is being executed to compensate for the moving target.

 

hmmmm, I'll admit I don't have that much experience with mouse and keyboard, much prefering a controller, and am talking as a lefty who generally uses the right number pad for such game controls, thus probably talking out my ass when it comes to whats comfortable for your standard mouse/keyboard player, but would it be possible to use the surrounding keys to WASD to perform different strikes? for example q being a left shoulder to right hip slash, e right shoulder to left? or go a little mmo and map them to the number keys?

you could also consider switching to a pseudo-aiming mode when preparing an attack - basic controls would be mouse-look and wasd move, but when preparing the attack you stop using the mouse to look around and re-implement the mouse track to aim strikes? it'd loose a little mobility, but it might help keep your original vision and combine it with usual video game
convenience?

I quite like the idea of timing based strikes, it gives a nice tactical element compared to skyrims and the like's 'mash attack till it dead' approach

 

Its just very hard fighting against gamer instincts.

I think there is also value in connecting the mouse to the motion of the weapon, but I would suggest something a bit more direct. As in, the weapon itself sort of follows the path of the mouse, so you sorta gesture with the mouse. So a slash from top right to bottom left would be 'mouse down' - mouse up and right, 'mouse release' - mouse down and left.

Holding the button is the 'round up', release the button is the swing.

Camera angle would still follow the mouse during both the round up and swing, but maybe with reduced sensitivity.

Having the mouse choose a target disconnects the player from the swing. Which is just a missed opportunity. Since you're using a first person camera, you're banking on a connected feel. If you were using a top down camera or something then being a bit more disconnected would make more sense.

The more you can directly connect the player to the swing and make him feel like he's in control of it, the better it'll feel. It doesn't need to be precise or super crazy but this kind of project is 100% about the controls. If you get the controls right, then even just that empty little demo scene will end up being fun (at least the first 5 or so times, you as developer will totally hate it after a while )

 

This is an interesting combat system from a game shows an effective use of timing

This anime apparently depicts medieval combat realistically

 

Yay! Glad I finally found this post again.

Check out this cool vid.
It might not have a shield, pushing away a spear. But.....
They're deflecting attacks, with hard plasticy, replica montante type weapons.
Hopefully, this extra short video might help, with your combat system.

 

Aye. I hope things are going well, for you and your game.
Because I know these types of games, are sometimes difficult to make.

Oh, I forgot to ask:
Will the player and allies, be able to throw their pommels, at their opponents
as a distraction, so that they can move in and end their opponents rightly?

 

Speaking of melee system, here is a break down of mechanics:

Spoiler: Beat'm analysis

BEAT'm up gameplay analysis
http://aztez.com/blog/2010/12/03/evaluating-beat-em-ups/
http://aztez.com/blog/2010/12/07/meaningful-and-elegant-defense/
http://aztez.com/blog/2011/01/24/difficulty-in-beat-em-ups/
http://aztez.com/blog/2011/01/29/the-mash-flow/
http://aztez.com/blog/2011/03/19/beat-em-up-sheet-music/
http://aztez.com/blog/2012/01/24/action-game-scenario-design/
http://aztez.com/blog/2012/05/24/the-amalur-problem/
http://aztez.com/blog/2012/07/23/watch-em-ups-sucking-the-meaning-out-of-action-games/
http://aztez.com/blog/2012/08/23/the-history-of-beat-em-ups/
http://aztez.com/blog/2012/12/13/a-combat-accessibility-fork/
http://aztez.com/blog/2013/01/03/on-genre-distinctions/
http://aztez.com/blog/2013/03/14/walking-and-running-implementation/
http://aztez.com/blog/2013/11/26/enemy-study-castlevania-dawn-of-sorrow/
http://aztez.com/blog/2012/12/26/mechanic-analysis-now-in-video-form
http://aztez.com/blog/2012/12/26/links-to-great-articles-re-combat

http://smashboards.com/threads/stratocasters-hitbox-system-complete-or-is-it.283973/
http://ensabahnur.free.fr/BastonNew/
http://www.actionbutton.net/?p=999
http://www.gamasutra.com/view/news/37356/Opinion_What_Makes_Combat_Fun.php combat

edit:

Spoiler: more of that

http://www.gamasutra.com/blogs/Matt...ding_Asymmetrical_Balance_with_Frame_Data.php
http://gamasutra.com/blogs/EduardoG...nderstanding_the_Street_Fighter_Tier_List.php

https://breadcrumbsinteractive.com/about-combat-system-design/
Mike Stout (Ratchet and Clank, Skylanders) has a great talk available for free online at the GDC Vault called “Reaching into the Toy-Chest“. Additionally, he wrote a series of articles on game and combat design focusing on a methodology he calls Trinity ( Part 1 through Part 9). These together were a huge help in getting me to think better about the player’s toolbox and how to design enemies. The combat design is framed as the game asking questions to the player (enemy behaviors), giving him the tools to answer them (player’s attacks) and changing the questions slightly in each instance of the combat (level design, enemy combinations), while keeping everything neatly organized in archetypes, dimensions and choices fields.

Sébastien Lambottin (Assassin’s Creed, Spiderman, Alone in the Dark) covers The Fundamental Pillars of a Combat System in his article on Gamasutra. The same approach of designing in terms of complementary tools and enemies that support those tools is seen here, plus he goes into more details on risk vs reward of using each player ability and designing complex enemies. On his blog he goes into more details on some topics like interesting enemies, groups of enemies and a few prototypes related to combat design. Update: looks like Sebastien’s blog ( http://immersivegameplay.net ) is no longer online. I’ll leave the links in the text in case you want to read them via the Wayback Machine.

Enemy Design

• How To Design Enemies, by Mike Birkhead, only available through the Waybackmachine.
• Writing this list, I found Mike’s blog, with a few other articles I plan to read next week. Don’t know how I missed these earlier.
• Enemy Design and AI, by Bart Vossen
• Enemy Attacks and Telegraphing + Post Attack Vulnerability, by Mike Stout
• Enemy Design in Link to the Past, by Nathan Savant (yes, it’s a list… didn’t say they’re not useful, just don’t use them as recipes)
• What we can learn from Doom, by Mark Brown
• Boss Battle Design and Structure, by Mike Stout (yes, him again… call me a fanboy if you must)

Level and Encounter Design

• Designing Combat Encounters in Uncharted 2, by Benson Russel
• A Beginner’s Guide to Designing Video Game Levels, by Mike Stout
• Level design for melee combat systems, by Bart Vossen
• Incorporating level design in Melee combat systems, by Bart Vossen
• Creating Satisfying Combat Experiences, by Joel Goodsell
• The Art of War: Gears of War, by Zi Peters
• Anatomy of a Combat Zone, By Josh Bridge

Misc articles that don’t quite fit the other categories

• 7 combat systems that every game designer should study, by Bo Moore
• Controlling Chaos Designing Compelling Combat in Action Games, by Adam Bormann
• Combat Design, by Mark Acero
• How combos in Kingdom of Amalur work, gamefaqs thread. Also this. Not an article, but a good inspiration if you want to design a combo system
• Aiding Asymmetrical Balance with Frame Data, by Matthew DeLucs
• UFGTX: How to Make Fighting Games, by Mike Zaimont
• Animation in Action Games, by Turbo Button
• Depth vs Breadth in Combat Design, by Mike Birkhead (so apparently I did reach Mike’s blog in the past, just missed some of his articles)
• Frame by Frame: The Sword’s Basics, by Ben Ruiz
• Combat Analysis, by Ben Ruiz. A set of articles looking at combat design in various games.
• Assassin’s Creed Combat System, by Alex Mandryka
• Fibonacci game design, by Mike Birkhead (oh come on! so apparently I found some of his articles in other places, and didn’t know they had the same author)
• A discussion on reddit about types of enemies

edit2

Spoiler: bonus

it's a shooter but they are explaining how they took inspiration in melee combat design

 

I flit between projects like a butterfly. That's the hobbyist life for you. I started this project off with melee combat, knowing that would be the hardest portion to solve. I never did solve it in a way that met my design intent, so I jumped to something else. Right now I'm working on polishing a mobile game.

I was going to do ranged weapons at some point, which would certainly have allowed throwing something just before engaging.

 

That's alot of good info.
I noticed, that one was for Mike Z, aka Skull Girls guy.

 

Would you possibly share the scripts and other stuff related to your combat system.

 

But @InstinctDevs...
I didn't think you needed any code...
https://forum.unity.com/threads/creating-games-with-unity-without-coding-for-free.539469/

 

Not possible. I didn't solve the design side satisfactorily. So the scripts never got written.

 

@verybinary thats how i make games without code by using other peoples code. stupid.

 

@verybinary wtf is it to you huh i just told that games without doing scripting yourself can be made jerk.
Why are there dicks like you.

 

@InstinctDevs
shhhh... this thread isn't about you
@Kiwasi
have you considered a "Fallout type aiming system"
select what your aiming for, then mouse and keyboard to determine what kind of attack?