Melee is not one combat system

The label melee covers Golden Axe, Double Dragon, Bayonetta, God of War, Devil May Cry, Dark Souls, and Monster Hunter. Close range is all they share. A combo game asks for a sequence you already know. A duel asks you to wait. A spectacle asks you to look. Treating them as one system is how a fight gets a move list and no pulse.

The question that sorts them is what decision you make in the next second, and what information you have when you make it. Dark Souls, FromSoftware, 2011, gives you a tell and a commitment. Bayonetta, PlatinumGames, 2009, gives you a cancel and a reward for not dropping the sequence. Both can be excellent. A game that borrows the tells from one and the cancels from the other without deciding which second matters will feel like neither. Build the second first. The weapons are costumes for it.

The label "melee combat game" stretches across Golden Axe, Double Dragon, Bayonetta, God of War, Devil May Cry, Dark Souls, Yakuza, Monster Hunter, and Castlevania. Close-range violence is all they share. Nothing says they should be built the same way.

Devil May Cry wants speed and flow. Dark Souls wants weight and patience. Yakuza wants a scrappy brawl in a side street. Principles carry across all three, but every detail has to bend to the feel the game is chasing.

So begin from the thing the player is doing, which is thinking. Who dies first? How close should I stand? Which of my moves suits this gap? Is it safe to swing yet? And when the answer comes back at me, do I block it, roll it, parry it, or hit through it?

The pace varies wildly. Some games hand those questions over slowly; others fire them at you faster than you can name them. The good ones make each second a run of decisions you can read, that cost something, and that pay off when you get them right.

An action hero is largely described by what their attacks do. The oversized blade in Monster Hunter, the launchers and air juggles of Devil May Cry, the axe you throw and recall in God of War, the whip in Castlevania.

Most heroes carry a small handful of base attacks, and each one should be the best answer to something. Raw damage. Speed. A wide arc that catches a crowd. Extra reach. Shoving enemies back, popping them upward, dragging them in, stunning them, cracking a guard, or letting you hurt something without standing next to it.

Every advantage needs a price attached. The heavy swing is slow. The crowd clearer barely scratches anyone. The ranged option burns a resource. The showpiece special sits on a cooldown. Whiff any of them and you are standing there, open.

The aim is not a longer menu of damage. It is a set of tactical answers that barely overlap, so the interesting question stays "which one, right now?"

Fast and slow attacks are really a conversation about animation. The button press does not land the hit. It launches an animation, and the hit arrives later, if it arrives at all.

Three phases of that animation do the work. Anticipation is the wind-up. Contact is the frame where the blow lands and damage happens. Recovery is the climb back to neutral.

Anticipation is what a player feels as speed. Transformers: Devastation can land a light attack in 14 frames and take 36 for a heavy one. Dark Souls pushes it further: a heavy strike from a huge weapon can spend close to two seconds winding up before anything connects.

All of that time is time you cannot defend. The target walks away, raises a guard, or simply hits you first with something quicker. That exposure is the reason a heavy attack reads as a commitment instead of a larger number.

More on this: Godot 3D game templates and the work they do not remove

Recovery deserves the same attention. Start a heavy attack in Dark Souls and you are usually locked in: no dodge, no block, no second swing until the animation is done. Broad animation canceling is simply not on offer.

Bayonetta takes the opposite line. An attack can often be cut short by dropping straight into a dodge, which is a large part of why the game feels loose, expressive, and cheap to change your mind in.

None of which makes Bayonetta the softer game. When the player can bail out of nearly any animation, the enemies are free to crowd and harass, because they no longer have to wait politely for a swing to finish.

Canceling is therefore a dial, not a virtue. Lock the animations down and every attack becomes deliberate and dangerous. Open them up and combat turns fluid and reactive. Whichever way you turn the dial, enemy design and player planning move with it.

Range is inseparable from stickiness, meaning how far your character slides toward a nearby target the instant you press attack.

Batman: Arkham Asylum leans hard on it, letting Batman fly across a gap into a distant thug, which makes precise range mostly irrelevant. Bloodborne snaps far less, so distance, angle, and the shape of the swing are all yours to manage.

Cut the snapping and positioning becomes a skill in itself. Players start to learn reach by feel: how far the whip carries mid-jump, how wide a sweep covers, whether the sword tip will graze or the big swing will sail past nothing.

A quick, busy game may need some snapping to stay readable and forgiving. But making the player think about where they stand adds a whole dimension. Super Smash Bros. commits to it by making Marth's blade hit hardest at the very tip, turning spacing into the point of the character.

Extra moves can mean extra tactics, though mapping an attack to every face button is not the only route there. Devil May Cry builds an enormous move set around one main attack button by changing what that button means.

Hold it and you charge. Hold a direction as you press and the same button becomes an uppercut or a lunging stinger. Elsewhere the trigger is context: environmental finishers in Yakuza, backstabs in Dark Souls, moves that only exist from a particular position.

Then there are combos, the button sequences borrowed from fighting games, which pay out for memory and timing. Some do nothing but raise damage. Better ones produce different tactical results, launching an enemy, knocking it flat, breaking its guard, or moving you somewhere new.

Long strings can also bury a newcomer. Friendlier games often reward timing instead, adding damage when you press attack at the exact frame the last hit lands. That asks you to read animation and hold a rhythm rather than memorize a string.

Half the system is what you do when you are not attacking. A strong combat game gives defense real options, normally some mix of blocking, dodging, and parrying or countering.

Those options should differ as sharply as the attacks do. A Dark Souls block is simple and can absorb the damage outright, but it eats stamina and some enemy attacks smash through it anyway. A roll can void the damage completely thanks to invincibility frames, and it costs timing, stamina, a recovery window, and your current position.

That is where the decisions come from. Blocking is the safe expensive answer. Dodging is the clean risky one. Giving up your spot on the ground might save you or wreck you, depending on the arena and who else is swinging.

When two defensive tools solve the same problem the same way, players pick one and forget the other. Good defense offers several answers and charges for each.

More on this: Celeste's Assist Mode is a tone of voice

Few things in an action game feel better than a counter. Hit the button on the exact frame an enemy attack arrives and you collect a prize: a killing blow, a stun, a stretch of slow motion, or some other edge.

The risk is that the prize gets too good. A parry that is easy to hit, safe to attempt, and lethal on success turns the whole game into standing still, waiting to be attacked, and deleting enemies one at a time. Early Assassin's Creed games drifted into exactly that.

Designers have several levers. Tighten the window, the way Nioh does. Raise the cost of missing, as Breath of the Wild does by leaving a mistimed dodge stuck in recovery. Or ration the attempts, which is what Bloodborne does by spending quicksilver bullets on gun parries.

The other approach is to keep the parry useful without letting it end fights. Bayonetta's last-second dodge opens Witch Time. God of War's counter leaves an enemy stunned. Both hand you an opening, and you still need the rest of the move set to cash it in.

More on this: Health systems in games tell you where to stand

The failure runs in the other direction as well, when a player can stay aggressive forever and blocking, dodging, and parrying stop mattering.

Stun-locking is the usual culprit. If a hit during an enemy's anticipation frames wipes out its attack and leaves it unable to act, you can simply keep swinging and never think about what it wanted to do.

Plenty of fixes exist. Give enemies poise or super armor so they shrug off hits until enough damage accumulates. Put the player on stamina so the swinging has to stop. Send several enemies at once, so pressure on one target means exposure to another. Make certain enemy moves impossible to interrupt. Let enemies break out of a stun on their own.

What you want is a rally between attack and defense. Nobody should be able to grind enemies down on autopilot, and nobody should be able to camp behind a shield waiting for a perfect counter. A healthy system keeps shoving the player between pressure, caution, reaction, and commitment.

Enemy design is the main instrument for examining what a player has learned. Enemies should travel differently, swing at different speeds, defend by different means, and demand different priorities.

One might fight from range and have to be closed down fast. Another might heal or raise its fallen friends. Some block, dodge, or parry you back. Some punish you for charging in, and others punish you for waiting.

They also have to be legible. Distinct silhouettes, colors, animations, and sounds let a player identify a threat in the middle of a crowd. Enemy anticipation should be exaggerated more than the player's, because reading it is the only way to answer it in time.

Bayonetta handles this with visual and audio tells, glinting weapons among them, that announce an incoming hit. In a game packed with effects, motion, and noise, that kind of signal is doing serious work.

A tempting mistake in enemy design is the creature that only one move can beat. It looks like depth and it is the opposite: it takes every option the player has and collapses them into one required input.

Resistance works far better than exclusivity. The God of War brute you cannot parry pushes you off a favorite habit without reducing the fight to a single button.

Enemies worth fighting drag the player through more of the move set. They make certain answers better in certain moments without erasing everything else.

The point is to break the habit of coasting through an entire game on three attacks and one combo you happen to like.

Enemies are not the only lever for variety. What the game pays out for can steer a player toward the parts of the system they would otherwise skip.

Furi hands back health for a hard parry. Transformers: Devastation can turn a finished combo into a heavy vehicle attack. Rewards like that make the risky, skillful option worth reaching for.

Scoring makes the pressure explicit. Clover and Platinum-style action games grade you on combos, variety, how fast a room clears, and whether you took a hit. Someone can still mash their way through with a poor rank, while a player chasing the top grade has a reason to use everything.

Enemies that adapt work too. Middle-earth: Shadow of War fields foes who learn from repetition and start deflecting the move you have leaned on, which leaves you looking for a different one.

You can get the distinct attacks, the layered defense, the enemy roster, and the incentives all correct, and the game can still fail because the hits do not land with any force.

Impact is built from art, animation, camera, effects, and sound. Studios that specialize in action spend a startling amount of their time making a single strike feel like it weighs something.

The animation phases return here. A satisfying hit usually has a big anticipation, a sudden snap into contact, a brief freeze on the moment of impact, and a recovery you can see. Resident Evil 6 has a punch finisher that runs the whole pattern: an enormous wind-up, a fast snap into the target, a held pose on contact, and an arm left hanging afterward to sell the effort.

Other games push the same idea harder. God of War can drop into slow motion as a heavy attack begins. Ryse can drop frames so a blade covers a huge distance in an instant. Transformers can freeze both fighters in the impact pose for a few frames.

Around the animation sits everything the player sees and hears. Colored trails trace the arc of a swing. Blur sells velocity. Sparks, flashes, hit reactions, red damage flickers, knockback, hit pause, and a solid sound all say the same thing: that connected.

None of this is decoration. It is how a player reads what just happened, and how the game confirms that the timing and the spacing were right.

A parry with great numbers and no punch behind it will not feel like anything. A heavy attack with no wind-up, no freeze, and no recovery will not feel heavy however much damage it deals. The feedback has to agree with the rules.

No recipe covers every game. Each decision should be measured against the feeling you want the player to walk away with, and the only reliable way to check is to get the build running and fight something yourself.

Pull the camera back for a flowing fighter that hands you a crowd, or push it close for a one-on-one brawl. Snap hard in a fast cinematic scrap, or barely at all when spacing is the point. Allow generous canceling for expressive speed, or almost none when you want commitment and danger. Widen the parry window for a power fantasy, or shrink it for a duel that means to hurt.

Underneath it all sits one question: what should the player be thinking about? Combat can be slow and considered or frantic and improvised. Mastery can live in long combo strings, in precise spacing, in stamina, in parry timing, in choosing targets, in chasing a score, or purely in how good the hits feel.

The systems that work know which decisions carry weight, make those decisions easy to read, and make getting them right feel enormous.

Good combat gives you attacks worth choosing between, defensive answers that are genuinely different, enemies that interrogate both, and reasons to keep digging into the move set.

It stops any single strategy from flattening the game, whether that is endless parrying, endless aggression, stun-locking, or one move that outclasses the rest. It sets positioning at the right level of importance. It tunes commitment and canceling to the tempo it wants.

Then it sells every correct decision through animation, sound, hit pause, effects, camera, and the way enemies react. Depth alone is not enough; the game has to feel good in the hands.

Combat is not really about damaging enemies. It is about being made to choose under pressure, living with the choice, reading what comes back, answering it, and feeling the weight of the moment you got it right.

In the Flockbay app, write the decision in one sentence: wait for a tell, or keep a sequence alive. Build one enemy that punishes the wrong decision and ignores the other. Play it. If both styles work, you have not chosen.

An AI soulslike maker and a character-action template are the two answers. Pick one sentence.