Notes on top-down ARPG combat in 2026

Combat in Aftercode is the third pillar, not the first. That changes everything about how it should feel.

This post is the design-rationale for combat shape. I read three reference games hard before locking in the spec: Riftbreaker, Helldivers 2, and Torchlight 3 (and a bit of Path of Exile 2). Each gets one thing brilliantly right and one thing wrong-for-our-purposes. The synthesis is what’s now in the build.

The constraint we’re solving for

Aftercode is a terraforming game first. Combat exists to:

Defend the work. Outposts, beacons, machines under construction. Combat is protective, not advancing.
Gate biomes. Each biome’s apex is a tier-up moment. Drop the apex, deploy the seed pods, watch the biome flip.
Punctuate sessions. A 30-minute terraforming session needs something that’s not “watch a meter fill.” Native fauna provides the punctuation.

What combat should not do: dominate the loop, demand mastery curves comparable to a dedicated ARPG, or require the player to grind for stat-jumps.

That spec rules out a lot. Let’s look at what it rules in.

Riftbreaker — the top-down extraction feel

Riftbreaker nails the top-down camera + mouse-aim combination. The character and the targeting reticle are two separate things, and that separation is what makes the genre playable on a top-down camera. (Compare a twin-stick where aim is locked to a stick — fine for arcade-feel, wrong for placement-heavy gameplay.)

What Riftbreaker gets right:

Camera height + pitch (~65–70°) keeps the world legible without losing personality.
Mouse-cursor as the aim source means placing a turret and shooting at a fauna pack are spatially identical actions. Place + shoot share a vocabulary.
Tools and weapons live in the same hotbar. There is no “weapon mode vs. tool mode” — you swap fluidly.

What Riftbreaker does that doesn’t fit Aftercode:

Hordes-against-tower-defense as the dominant loop. We’re not doing that — fauna in Aftercode is sparse, not swarms.
Heavy emphasis on damage-per-second tuning. Aftercode’s combat is too brief to support DPS optimisation.

Adopted: camera framing, mouse-aim, agnostic hotbar. Rejected: wave defense as the central loop.

Helldivers 2 — the “every fight is a story” pacing

Helldivers 2 is masterclass-level top-down combat. The thing that’s stayed with me reading it is the brevity of individual fights. You crest a hill. You see four chargers. You burn three stratagems. 25 seconds later they’re a smoking pile and you’re moving on.

That brevity is precious for Aftercode because of the third constraint above: combat should punctuate, not dominate. A Helldivers fight is 30 seconds, not 3 minutes. That’s the right shape.

What Helldivers 2 gets right:

Each fight has an arc — set-up, escalation, payoff — but the whole arc fits in under a minute.
Tools matter more than stats. The orbital strike and the resupply pod are the answer to the fight, not your DPS.
Ammo and resources are scarce. Every shot is a decision.

What Helldivers does that doesn’t fit:

Mission-as-encounter structure. Aftercode’s “missions” are atmospheric stages and biome restoration, not 30-minute drop-ins.
Stratagem call-down theatre. Beautiful, but tonally too loud for a contemplative terraforming game.

Adopted: fight brevity, tools-over-stats, scarce ammo. Rejected: mission structure, stratagem theatre.

Torchlight 3 — the “lots of tools, fast hits” loop

Torchlight 3 is divisive but I’m specifically interested in its hit feel. Torchlight games hit fast. 2–5 hits per enemy. The combat moment-to-moment is whippy — you’re always doing something, always seeing damage numbers, always feeding the dopamine loop.

What Torchlight gets right:

Hit cadence — fast enough to never feel sluggish, slow enough to register weight.
Damage numbers as feedback. (We have these — DamageNumber.cs and DamageNumberSpawner.cs already shipped under Phase 0 work.)
Skill modules as plug-and-play. Add a support gem, the skill behaves differently, the player learns by combination.

What Torchlight does that doesn’t fit:

Item churn. Torchlight is built on the loot loop. Aftercode is built on the terraforming loop. Items in Aftercode should be occasional and meaningful, not constant.
Class fantasy. Torchlight has classes; Aftercode has loadouts.

Adopted: hit cadence, damage feedback, skill+support composition. Rejected: loot churn, class systems.

What Aftercode’s combat shape ends up being

Synthesising the three:

Top-down camera (Riftbreaker pitch).
Mouse-aim (Riftbreaker fluidity).
Agnostic 8 + 8 hotbar (Riftbreaker tools-and-weapons together).
Fights last 2–5 hits, ~30 seconds total (Torchlight cadence + Helldivers brevity).
Five sci-fi damage types — Kinetic, Plasma, Cryo, EMP, Radiation. The R3 rename completed last week. Each has a status counterpart (PlasmaBurn, Cryo slow, EMP amp, etc.).
Tools-over-stats. Status effects, area denial, deployable turrets, cone weapons, projectile + AoE — not stat-stacking.
Scarce ammo. Most weapons consume resources. Most placeables have a budget.
Defense via building, not via blocking. No active block button. You build a wall, you place a turret, you stand behind it.

What this looks like at the ground level

Imagine a mid-session moment, ~15 minutes into the playtest scene:

You’re scanning a Biotech Crystal node. Three Frostshades approach from the south — tracked by the minimap, not by hearing them, because storms drown out audio.

You drop scanning, swap from your scanner to your Sporeshot (slot 4). Three cone-burst shots deal Plasma + 60% Burn. Two die instantly; the third is on fire. You finish it with a melee swing while it’s still ticking damage.

Total elapsed: 18 seconds. You go back to scanning.

That’s the loop. That’s the rhythm. Brief, decisive, tool-shaped.

What’s not designed yet

A few things still on paper:

Apex fights. The biome bosses. They need a 60-second arc with telegraphed phases — not the 5-minute fights from the old dungeon era. R7’s defenses work informs this.
Storm-coupled combat. A Local storm rolls in mid-fight. Should fauna behavior change? Visibility drops, ammo doesn’t. Worth testing.
Co-op fight density. Solo combat is sparse. 4-player co-op probably needs 3× the spawn rate to feel right. Tuning lives in R9 + R10.

What this means for content scope

Combat for Aftercode is fewer enemy archetypes, fewer damage curves, fewer class systems — but each one has to be designed with care. Quality over quantity. We have ~12 archetypes planned across all biomes; 1 apex per biome; 8 + 8 hotbar weapons mapping to the 5 damage types.

That’s a content scope a solo dev can ship. It would not work as the headline pillar of the game; it works as the third pillar feeding into terraforming + building.

The next post

Combat-shape decisions feed directly into Three loops, one world — the design rationale for the three pillars overall. If you’ve read this far, that’s the post that ties the rest together.

Or — the newsletter, if you’d rather catch the next devlog as it drops.