Trapmancer

Inspiration

One trap is nothing. The chain is everything.

Most tower defense asks where to put your towers. I wanted one that asks how cleanly you can end the wave.

TRAPMANCER began with a single feeling: the rush of a perfect chain, a lane frozen, then struck, the whole pack shattering at once, and the game paying you back for it. It treats defense as an act of authorship, where the reward comes from how you kill, not just that you killed.

What it does

TRAPMANCER is a real-time, top-down tower defense built on one rule: a stylish kill refunds mana, and that mana funds the next, bigger chain. You do not build fixed towers. You hold a hand of four magic-trap cards, drawn from your deck and paid for with mana, and you place them on the floor lanes and the side walls as the horde pours toward your Relic.

Traps are elemental, and elements react. The game is won by chaining them:

❄️⚡ Frost then Lightning · SHATTER · a frozen pack explodes, and the mana floods back.
🛢️🔥 Tar then Flame · INFERNO · a soaked lane ignites into a wall of fire.
🔮 Lure, then anything · pull the horde into a cluster, then end them as one.

Between waves you draft a roguelite boon, so every run compounds differently. You are graded on the kill, Brutal, Elegant, or chained, and your style is both the score and the fuel. Difficulty comes from your timing and your deck, not from random spikes.

How we built it

This is a design-first submission: a Game Design Document, a Player Journey Map, a Visual Concept Package, and a Production Plan, backed by a playable browser prototype of the core loop.

It is designed for Meta Horizon's mobile client as a real-time, top-down 3D experience that is solo at its core and scales to co-op, with every system mapped to a mechanism the engine provides:

A small 3D vault scene with stone lanes and two trap surfaces, floor and wall, under a fixed top-down camera.
A real-time card hand and mana economy, where mana regenerates over time and jumps back up on a stylish kill.
A deterministic simulation on a fixed tick for the horde, with waypoint pathing, fixed-timer waves (so there is no stalling to farm mana), and the elemental reaction lookup.
A Custom UI overlay for the hand, the mana bar, the style meter, and the between-wave boon draft.

The Production Plan lays out build sequencing, MVP scope, dependencies, and how the core loop is tested before full production.

Challenges we ran into

The hardest balance was the flywheel itself: rewarding good play with mana without ever making mana free. Tune the style refund too high and the loop runs away; too low and the magic of a chain paying you back disappears. The answer was a careful refund curve paired with fixed-timer waves, so a player cannot stall on a single straggler to bank resources.

As the game grew to a fifteen-trap roster, an elemental reaction matrix, roguelite boons, and co-op, the real challenge became adding depth without diluting the core. Every new trap had to give the player one more reason to chain, never a separate game to manage. Keeping the reactions readable at top-down phone scale, and the card hand playable with one thumb, took the most iteration.

Accomplishments that we're proud of

A premise that fits in one line, backed by a single mechanic that ties the resource economy to an act of skill.

The combination is new to mobile tower defense: the familiar draw-and-pay of a card battler, the chain-and-style soul of a trap game, and a mana-refund flywheel that makes playing well literally pay for itself. Restraint did the heavy lifting, one deck, one mana bar, one style score, so the depth comes from how systems combine, not from a wall of separate features. The documents and the prototype share one rule, one palette, and one voice, so the whole submission reads as a single game rather than a stack of files.

What we learned

Designing TRAPMANCER taught us more about how to design a game than about any single feature:

🎯 Start from a verb, not a feature list. "Chain everything" guided every decision and kept the scope honest.
🔁 One rule can carry a whole game. Style refunds mana grew the economy, the skill ceiling, and the replay loop on its own, with no extra systems bolted on.
✂️ Design by subtraction. We judged each new idea by one test: does it give the player one more reason to chain? If not, we cut it.
💡 Decide the feeling first, then the numbers. We chose how a clean chain should feel (it should pay you back instantly), then tuned the refund to deliver it.
🛠️ Prototype the core early. A rough, playable loop taught us the flywheel feel in a day, more than another page of notes could.
🃏 Borrow a familiar door. Building the controls on a card hand let new players reach the deep part fast.

What's next for Trapmancer

Next, TRAPMANCER becomes a fully remixable Meta Horizon world. The goal is to ship it not only as a game to play, but as an open, documented template that other creators can clone, reskin, and build on directly inside Meta Horizon Studio:

🧩 Remixable by design. The trap deck, the elemental reaction matrix, the style scoring, and the wave system are exposed as clean, reusable building blocks. A creator can drop in their own theme, a haunted manor, a sci-fi reactor, a candy fortress, and keep the chain-and-style core intact.
🤝 Deeper co-op chains. Cross-player reactions with a bonus refund when two mages link a combo, and role decks so a vault is a team, not just more hands.
👹 A bigger bestiary and bosses. Flying enemies that only wall and lightning traps reach, shielded foes that need a physical hit first, trap-breakers, and a vault-guardian boss with a weak point only a long chain can crack.
🃏 A trap collection. A fair, horizontal gacha where rarity unlocks new chains, not bigger numbers, plus seasonal trap schools and a deeper deck-builder.
⚔️ Trap Duel (PvP). Two mages defend and send waves at each other, with style score breaking ties.
🏆 Daily mutators and an endless gauntlet, ranked by style, not just by waves cleared.

Single-player stays the heart. Everything new, including the remixable toolkit for Meta Horizon creators, exists to let more people feel the chain land, and to let them make their own version of it.