next »

[baconman]

...and the fact that you can throw enemies against the screen. Now just how badass is that? 

And not only that, they based an entire boss fight on that mechanic! So it was actually USEFUL, too. (But admittedly, a very cheap and quick way to clear foot soldiers that would otherwise take 2-4 knockdowns.) It did make me miss the back-and-forth slam, however.

[laserdracula]

Any good beat 'em up has to have you saving your girlfriend from a gang/ninjas/ninja gang.

[baconman]

Any good beat 'em up has to have you saving your girlfriend from a gang/ninjas/ninja gang.

OR to save your daughter from a gang/ninjas/ninja gang - AND that idiot boyfriend that got her INTO that mess to start with. SPD that shit!

[Seth]

I'm not sure but every beat 'em up I've ever played has paled in comparison to River City Ransom no fancy effects or enemy types I don't even need Dragon Feet or Stone Hands special moves just good old fashioned fist to face

[kenesque]

Good combat design is important. You need to give the characters enough tools to handle any situation without ever dying, but at the same time make sure the enemies are all unique and interesting and have different approaches/gimmicks, or you end up with incredible stale combat. Every setup in the game should be beatable (ideally) without getting hit, though that is clearly not possible, it's best to aim for being unable to die if you are good enough of a player.

I think the most important thing these kinds of games get wrong is hitting buttons doesn't feel fun. When you punch someone, it should feel cool. The sound, the hitstop, the effect, the hitreel of the enemy, you need a nice combo of all of these to make the combat feel good.

SoR2 pretty much nailed all of this besides some of the bosses/enemies being a little too difficult to deal with properly, depending on your character. The only thing I dislike about SoR2 otherwise is the walk speed, too bad SoR3 was such a bad game because the inclusion of universal rolling/running was a really nice touch and pretty much would have made SoR2 that much better.

[the_dannobot]

I think the most important thing these kinds of games get wrong is hitting buttons doesn't feel fun. When you punch someone, it should feel cool. The sound, the hitstop, the effect, the hitreel of the enemy, you need a nice combo of all of these to make the combat feel good.

SoR2 pretty much nailed all of this besides some of the bosses/enemies being a little too difficult to deal with properly, depending on your character. The only thing I dislike about SoR2 otherwise is the walk speed, too bad SoR3 was such a bad game because the inclusion of universal rolling/running was a really nice touch and pretty much would have made SoR2 that much better.

YES, the combat needs to sizzle.  Axel from SoR2 is one of the crunchiest characters in video gaming, which is why that's one of my favorites.

Also, are these games called beatmups yet?

[Derakon]

I think a beatmup would be a rhythm-based beat-em up. Which could be cool.

I just realized why the combat in SoR3 didn't feel as powerful as it did in SoR2 -- they sped it up massively. What used to be punch, punch, crosspunch, kick-kick became pupucrokikick -- at least twice as fast. Which means less anticipation for each incoming hit, less impact after the hit is delivered, less time leaving the grunts stunned before you knock them into the air...SoR3 might have had more room for demanding reactions and quick in-and-out raids, but the act of beating the stuffing out of goons just wasn't as satisfying.

[BlueSweatshirt]

I think in the respect of a successful beat 'em up, you can't go without mentioning Little Fighter 2.

[Kicks]

I'd rather post in the other thread that was linked, but I got some red text yelling at me, so I'll post it here.
I'd like to think that beat 'em ups are my forte. I flippin' love them.
I'm going to talk about a few that are incredible. Post a link to a terrific article written by the programmer of several of those games. Then share a bit on my own philosophies about the genre and lessons learned while making Aces Wild.
I break it all down at the bottom if you wanna skip the game by game breakdown.

http://www.tylerdoak.com/?p=122 <--There's the article. It's really good!

Technos: Double Dragon 2/Double Dragon GBA/River City Ransom/Kunio-Kun
Capcom: Devil May Cry
Ninja: Ninja Gaiden
Clover/Platinum: Godhand/Bayonetta

Technos sorta started it all off with Renegade, but made it a hit with Double Dragon. They also made River City Ransom, which, we all know, is great too.
Double Dragon was basic, but had many elements come together to make it great. You had a nice concise moveset. Enemies could approach you from 360 degrees. Many enemy types. Several weapons. Hazardous environments--which had lots of aesthetic variety as well.
The moveset also increased as you played (RPG LM-entz!). But more importantly its variety meant you had choices to make based on what was going on. The basics of it is slow and strong vs fast and weak. This is simple enough, but many beat 'em ups ignore it. The next part goes along with game design in general: The more actions you have and the more objects you have to act on them, the more variety you have. Because you have several moves, times several enemy types, times the current environment and the actions you take alter all of those at once, you have lots of variety. Most beat em ups boil it down to just actions vs enemies and that's a whole dimension of variety your missing out on.
River City Ransom did the same thing, only it was a console game so the story and rpg elements were ramped up. Also the action was faster, more cartoony and manic. Both games also had weapons which provide a "sanctuary" of sorts that I'll talk about more later. Both of those games had 2p coop. Instant win if it wasn't already. I mention specifically DD2 and DDGBA because they are my favorite. They are incredibly solid and epitomize everything that made Technos great.
After that, we had Streets of Rage, all of Capcom's cool beat 'em ups, and wacky obscure ones like Knuckle Bash. These games were very solid but mainly carried the beat 'em up torch in a pretty consistent direction.

Then we had Devil May Cry and the new Ninja Gaiden. These were gritty, intense, crazy, hard and 3d. The 3d isn't that big of a deal as double dragon basically played the same way spatially, though the immersion was greater.
Instead of having that extra environment element strong, these two games focuses heavily on player ability. You were super human and super badass. Each one took a different philosophy. Devil May Cry gave you tons and tons of moves and lots of enemies and epic bosses. It was your job to use those abilities to create variety. Ninja Gaiden's approach was on the other end. That game wants to kill you. Every enemy is a huge threat and you have to block a lot and really think hard about what you want to do. Which leads to another "innovation" these games had. Defense, you could block, dodge, run and jump dynamically to avoid enemies. This is more important than you might think. In previous games you simply walked around to avoid enemies and some others allowed you to block.

To paraphrase from the article I mentioned, blocking is bad. It is boring and makes you wait for the enemy to act. Dodging is a reaction. When the enemy attacks, you react and dodge accordingly. If you think a bit, you can see how this already adds a new dimension to the combat. You actually have defense and offense to worry about!

So out of the blue we have this crazy game called Godhand. You had TONS of attacks and tons of defensive actions. In Godhand, they did something great. They gave enemies defense. You had to not only worry about attacks in the traditional way, but the enemy might block which means you have to act differently. In Ninja Gaiden, the correct action isn't always clear and enemy blocking is frustrating. In Godhand, to quote designer again, "it's an opportunity." By breaking an enemy's guard, you can stun them, then really go crazy! So now you have 3 basic actions--attack,dodge,break-- that work better or worse depending how the enemy is acting. Then within those design spaces you have different types that have all sorts of different properties. Should I sidestep or duck? Should I punch or kick? What's great about these design spaces is that the game presents you with the variety: enemy block, enemy attack, enemy open: but you get to decide the specifics. So it's a terrific dance with the game and its enemies, while you get to be a bit creative when the tempo changes. Godhand's attack system was a bit lacking despite the number of moves. Basically you just wanted to abuse the best move over and over again and sometimes use one that had different properties. Bayonetta solved this problem and combined two aspects. You have 3 basic buttons in Bayonetta: Quick attack, Strong attack and dodge. To get to more skills you have to work your way through a combo. So you mash on Quick attack, then hit Strong attack to get a REALLY fantastic attack. It takes time, however, to reach that move, so you may opt to go for a lesser attack, but the payoff isn't as great. Further, while you're trying to reach that attack, AN ENEMY ATTACKS! So you have to dodge and risk losing your combo spot(though you can retain it). BUT! If you dodge right at the last second, you get WITCH TIME! This is like the guard break opportunity in Godhand. It gives you free reign to really wail on enemies. Which brings us back to that sanctuary thing. All of those games have limited powerups that allow you to be really powerful for a short time. Godhand and Bayonetta have both special attacks, context sensitive attacks and weapons that do this. Double Dragon had weapons. DMC and NG had special moves and forms.

One of my basic design philosophies is, again, actions vs objects. Just a quick aside, Yoshi's Island really epitomizes this. You can lick, stomp, jump and shoot. All of those things affect both the environment AND enemies IN DIFFERENT WAYS! Really keep that in mind when designing action games and beat 'em ups.

[Kicks]

So we want to design a beat 'em up. You REALLY want to avoid monotony as it is the downfall of many a good beat 'em up. To fight that you want both variety and surprises. It's nice to have both of those in all of your design spaces.

For this, we'll use Player, Enemies and Environments
For Player, let's really go back to our roots, but ramp up those ideals for modern intelligent design.
Let's say...
Attack: the real reason we love this genre
Dodge: because blocking is boring and we wanna be the Matrix in the 80s. (beat 'em ups are in the late 80's by default)
Grab: because we all miss this from old games
Move/Dash/Jump: so we can get around our environment AND enemies

Designing Actions:
Each of these design spaces should try REALLY hard to do something the others don't.
Attack: most direct way of damaging enemies. We'll also let it destroy the environment.
Dodge: how to avoid an enemy's attack.
Grab: lets you interact with enemies and the environment in a more dynamic way.
Move/Dash/Jump: each of these should have a risk/reward that the others don't. Jumping is sorta fast and you can move slightly. You can attack in dynamic ways with this.
Running will have a bit of startup, but then you move more quickly in a straight line. Stopping leaves you vulnerable. Moving is regular. You can't dodge during a jump or run.

What's great about these character states, is that we can mess with them dynamically. What if we could do all of those actions AGAIN while in those states? We'll limit ourselves a bit for balance and calamity, but now you can use those states--and their strengths and weaknesses--to lock more powerful moves.

Designing Attacks.
We'll stick with fighting game basics here. Speed, range, damage, knockdown and area of effect. But never forget these three others: creativity, coolness and immersion. Another design principle is that elegance is often at odds with character. Find the right balance!
Punch Combo: Basic attack. It is fast with a comfortable range. It gets stronger the longer the combo goes on. It ends with a cool knockdown move that can hit multiple enemies.
Kick Combo: It's a short two hit combo. Speed and safety is almost always preferred, so we need lots of reasons to use this over punching. We'll make it a bit slower, but well give it lots of damage and range, plus it moves the character while attacking.
Uppercut: This attack is faster than the kick, but slower than the punch. It has really short range, but ok damage and knocks the enemy into the air. Letting you juggle them with more attacks or simply use it to knock the enemy down. It has poor recovery.
Low Kick: This attack is really really fast, but has very little range. You can do it several times in a row. It's the fastest attack, but its range makes it hard to use and it's damage is bad, so lets give it a bit more oomph. It can stun single enemies pretty well and can also hit knocked down enemies easily and trip running enemies!

Jumping Attack: It will be the classic angled diving jump kick. This is a strong choice because it not only is a cool attack, but is a special movement as well. Since we have to get to the peak of our jump to use it, we'll let it knockdown enemies, but it won't do much damage.

Running Attack: We'll make it a shoulder tackle to add some variety. It can be used to ram groups of enemies. Its the simplest way of damaging multiple enemies. It will also have a stronger knockback effect, but it's difficult to use and leaves you vulnerable after using it.

Let's design the grab!
Not only will it be like a regular attack with damage and range properties, it will operate dynamically based on the enemy state. We'll further design these dynamic grabs to have different uses based on the environment.
Right now, we have mentioned 3 enemy states. Idle, down and running. Let's add jumping, blocking and attacking. WOW! That's six different enemy states giving us at least 6 effects to use with the grab.
The initial grab is about the speed of the punch, but with a shorter range.
Against an idle enemy we have two basic effects which will be the basis of all of our other grabs. One simply pushes the enemy in any direction. The other slams them down quickly. The push does negligible damage, but allows you to dynamically position them and disrupt other enemies he gets pushed in to. If you're lucky you can knock them into a wall or pit for better results. The slam is pretty good damage and knocks them down quickly.
A stunned enemy is one who has been damaged a certain amount without respite. It will change based on the enemy, but a normal enemy would require 3 hits from the punch combo, or 1 hit from the kick combo to get stunned. We'll just follow that template of concentrated damage vs splash damage and ramp up the extreme. We'll also make it so you can tack on some extra damage with knees and headbutts for extra coolness and immersion. The throws themselves will be a strong shoulder throw that makes the thrown enemy knockdown other enemies. Or alternately a strong wrestling move like a suplex for big concentrated damage. Because we've REALLY earned this, we'll make it so you can drag around the enemy slowly and even jump before doing the grabs.
For an attacking enemy, we'll do something special. We'll make the timing a bit tighter for this to happen too. We'll make it a fancier aikido type counter throw. The reason for this is so the animation takes a longer time and we'll make the player invincible during it. This will be like a special dodge with a bigger risk, but gives you more time to think the damage won't be too great either as it is more defensive. We'll just stick to the one grab for attacking enemies since it's more of a reactionary tool.
For a blocking enemy, let's do something else different. It'll be a brutal grab/attack combination that leaves the enemy standing and stunned allowing you to choose your next move>chain grab or uppercut finish?
There are other states like grounded and running. What sort of special grabs could be used there?

Designing the dodge.
Because you have two counter-moves already, you could skip the dodge, but I'll continue to show a way of going about it.
We'll have two types: handspring and ducking. Ducking will allow you to dodge then quickly counterattack, but you won't move at all. The handspring is invincible in the beginning, but leaves you vulnerable at the end. However you can move around while doing it. If you do either of these 3 times in a row. The third one will become an emergency roll which is a way worse version of the handspring with bad recovery.

Designing enemies.
I have a friend that thinks if you had a game with just attack and dodge and tons of enemy variety, it would be incredible. I believe his statement more and more every day. Having lots of enemy variety is of UTMOST importance. Beat 'em ups are basically non-competitive fighting games. Fighting games are great because of a human opponent. This means you need your enemies to have lots of variety and surprises. There are several design direction you can take, but using all of them is best.
Player Actions. Basically this means, make sure all of the player actions find use. This can be a direct utility like breaking through a guard, or simply they take more damage or have a unique reaction to the action.
Attack, that one is easy.
Dodge also easy at first. Making different timings and reasons to use our different dodges is where the work comes.
Range. Make flying enemies, running enemies, keep away enemies, projectile enemies, hiding enemies.
Grab. This changes based on states the player forces on the enemy AND states the enemy presents by itself. Truly our most dynamic action!
Flavor is also important, of course! Not only do we want to test our players' use of actions and strategy, we want surprises and variety that come from immersion and aesthetic. Fast ninja ladies, hulking muscle men in iron underwear, the stern martial artist. Consider their actions, their behavior and how the player can interact with and react to each enemy character.

Designing the environment.
This one is both easy and hard. Thinking up ideas is easy, but balancing it is very tough. As it does indeed add an extra dimension of design, that's a whole dimension of stuff being too good. Pits for example. Very awesome and it feels great to knock enemies into them, but a one hit KO makes things degenerate. Your entire moveset changes dynamically with the environment. Find ways to change the environment. Make the environment change through travel or actual direct altering via player actions and interactions.
I personally made certain things in my game get "used up" like a hanging florescent light or crate. The enemy takes extra damage when knocked into it or it's knocked into the enemy, but it gets destroyed.



This got a little crazy so I'll close by talking about surprises. This is hard to do because people love randomness if it's right and HATE it if it's wrong. One way Godhand did surprises right is with guard breaks. You have to wait for the opportunity, but you can also react when it comes and the payoff from a guard break is great! Surprises can be both beneficial or threatening to the player, but doing it right either way can be positive. One way is to have literal pop/surprise enemy appearances or actions. They could be semi-based off of the players' previous actions. One thing I did is have random items and enemies appear from broken crates. Godhand had certain enemies transform into demons when knocked out. Streets of Rage had certain enemies have long range charging attacks that start off screen. Weapons and items are good because they can be used by the player or enemy making the simple acquisition of them an exciting drama. The idea is to surprise them, but make sure they have all the tools necessary.

To sum up...
Have actions that affect the player, enemy and environment.
Have actions that are used in REaction to the enemy and environment.
Make choices interesting.
Don't have too huge of a moveset, but don't stifle variety or creativity.
Lot's of enemy variety.

[Kicks]

Triple Post. Don't Kill Me Please. I like the organization.

I'd like to talk a bit about actually designing enemy behavior for beat 'em ups.
This is, BY A HUGE MARGIN, the hardest thing I've done in regards to making Aces Wild.

I've heard it a million times but even then I sorta didn't listen, so I'll say it myself. Enemy behavior is an iterative process. JUST LIKE EVERYTHING ELSE ABOUT GAME DESIGN. Step one, get something out there. Even if it's just a dude walking back and forth make it happen.
Sorry to sound so negative, but designing one enemy is a nightmare, so organizing more than one together is a waking hell. It can be really tough. A big part of my troubles, honestly is having my game be 2D sidscroller. 3D makes it easier to organize enemies.
Anyways... I'll share my simple design outline and some specifics along the way.

I'll talk about groups first as the group behavior is a bit simpler.
One article from Gamasutra/GameDevMag, posted here or in that other thread has a nice breakdown of group mechanics and how many different game did it.
Often you have the standby group and the engagement group. Depending on the number of enemies and the game mechanics these groups will change in ratio or even bleed or fuse together.
Basically you have a group that hangs back to give the illusion of threat, while the engagement group actually attacks the player. The engagement group can be static, or they can switch in and out.
Often the enemies don't perform actions on a personal level, but on a group level. So you have a simple timer that counts down, probably goes faster or starts lower based on the total number of enemies and when it runs out, it calls to a random or specific enemy to attack. Some games mentioned have another timer that calls a support enemy to attack from afar or charge in then retreat. I currently use a single group that can have any enemy be called.
So the enemy has been called to action. NOW WHAT?
Actually, there's a couple steps that happened before the actual call.

Here's my basic behavior outline:

Entrance
Standby
Notice
Chase
Approach
Engage/Tell
Attack
Retreat

I've found that it makes battles much more believable, if you have an animation or alteration of another animation to go for every one of those actions. It's surprisingly important.
First the enemy appears when an event is triggered, or the enemy was already there on patrol.
Then the enemy notices based on line of sight or any other senses you want to program.
Many beat 'em ups don't have this as they're action based. But stealth is cool too!
So the enemy notices one way or another then enters Chase mode.
This is a quick way to close distance. I use a run or flying animation or even teleports if they're really far away. I use the Pythagorean Theorem to determine how far away my enemy character is from it's target. The target for Chase Mode is a specific value for the enemy plus the player character's position. This means it will chase until it gets within that certain enemy specific distance.
Once we're here, we have Standby and Approach. I used to treat them the same, but have realized how much "intelligence" it adds by having an Approach animation. For me, that is simply a guarded or threatening walking animation. Certain games don't actually need this. I was looking a lot at Bayonetta and Devil May Cry and first, which didn't really have it. But busted out Double Dragon and Godhand and realized there's some really simple but important behavior going on. This is what I call Approach. It's often just a walk animation but has the enemy move forward or backward, often directly related to player input or velocity at a slightly or greatly reduced rate. This is so important I'm gonna say it again. In the past my enemies would run up, then hang out in an idle position, but wouldn't Chase until I got far away. If they chased constantly it was silly run and stop behavior. By having this buffer of slow movement, you have more engaging behavior. At this stage, you also have some random movement and enemy organizing. Simply, if there are an even number of enemies and one side of the player has less enemies, have the side with more send a random enemy to the other side. This can be done with a special movement, like a jump or hand spring or a simple run. This keeps the player surrounded and adds drama to the battle. A little later I realized that Bayonetta's wasn't simpler, it was more dynamic. Enemies may have several approaches, chases and standby's. The big claw enemies may stalk you or jump around. Either way they just wanted to get close.
Enemies in the outside Standby group or that are close but not moving much, can use an Idle or Standby animation. I often have a special movement or two depending on the enemy as well. My ninjas hop around a bit. My flying mechs dart around. My martial artists sway and juke. Randomly I may have an enemy do a random position switch.

FINALLY, we get to the attack. The chosen enemy then does the Engage. This is often a quick dash or jump to a specific position relative to the player character. Some enemies simply fly to a certain spot. A ninja may teleport above. A fighter may roll or dash. Along with the Engage I have a Tell. You may need only one or both for each attack. When my regular ninjas teleport above, they are going to fall and do a strong punch. When my fighters dash to the side, they may either do a ducking dash then punch or jump then kick. In this case, the engage gets them into fighting position, then they have an additional tell. My flying mechs simply fly to a position then shoot. My dogs have dashing attacks so they will often growl in place then attack. So instead of a positional cue, it's more aural and animation based.
From there I have a retreat. Sometimes this is necessary, sometimes it isn't if the attack is simple like a punch or combo, I have the the enemy backflip away when done. If it's a slower powerful attack the follow-through then a simple run back to standby is often enough. After that you just repeat.
I have a list of attacks for each enemy that is chosen at random. You could make them patterned for bosses or even give them statistical weights. Some of my attacks have random Engage positions. Some Engage actions have random attacks that come out. I have independent enemies that attack outside of the attack timer. They're often projectile enemies that shoot every so often.
It may seem sorta complicated, but I find it works good for me after struggling to have a simple chase and attack system.
You can make them as simple or complex as you need and may opt to skip certain behaviors, like retreat or approach.

Other things of note.
Some games have some enemies with a separate moveset when far away, or will use a certain attack to close the distance.
It can be easy to get caught up in using magic numbers (non-variable values) when coding them. Avoid this if possible.
The line between Approach and Chase is a good way to have enemy variety. Eager enemies will rush in. Smarter enemies will approach with caution.
Some sort of way to predict enemies, either having Tells or predictable/patterned behavior is important, but make sure there is variety or even randomness.

[Makai]

I think a beatmup would be a rhythm-based beat-em up. Which could be cool.

Draglade for the DS had some rythm-based elements to its fighting system.