🖥️ PC1998Blizzard EntertainmentBRUTALStarCraft rebuild deck
BRUTAL
Three factions that share no units and still have to be equal is a balance problem measured in years, and underneath it sits the part nobody budgets for: two hundred units pathing through each other at sixty frames a second, where the naive approach dies at forty units and the correct one is flow fields plus local avoidance plus formation logic.
The hard part is not game design at all.
Brutal, and the brutality is front-loaded into two decisions you make on day one: flow fields instead of per-unit A*, and deterministic integer simulation instead of floats. Both are boring, both are invisible in a demo, and both are a rewrite if you skip them. Everything else — build trees, fog, combat — is ordinary work. Build one faction and one map, get 200 units moving properly, and treat balance as a thing you will never finish, because Blizzard did not finish it either until the professional scene had played it for a decade.
- Difficulty
- BRUTAL · 5 of 5
- First playable
- A project
- The original
- 60 team-years
- Released
- 1998 · PC
- Genre
- RTS
Build a real-time strategy game with one faction and working large-scale pathfinding. Any stack, but plan for 200 units on screen from the first line. Requirements: - Resource gathering, a build tree of at least eight structures, six unit types with rock-paper-scissors counters, fog of war with explored-but-not-visible shading, and a minimap. - Pathfinding is the architectural decision, so make it first. Use flow fields: compute one integration field per destination across the map grid, and every unit ordered to that destination reads its direction from the shared field. Layer local collision avoidance on top for unit-to-unit separation. Do NOT run A* per unit per frame — it is the standard approach in tutorials and it is why RTS clones stop at fifty units. - Selection and control that respects the genre: drag box select, control groups on number keys, shift-queued orders, attack-move, patrol, and a formation that keeps ranged units behind melee. - Combat with damage types against armour types, so counters are real numbers rather than a special case per matchup. - An AI opponent that runs a scripted build order, scouts at a fixed timing, and switches army composition in response to what it saw. Give it difficulty by reaction time and scouting frequency, never by resource cheats. - Fixed-point or integer maths for all simulation state, and a deterministic lockstep loop, even in single player. Retrofitting determinism after the fact is a rewrite. Definition of done: 200 units cross the map through a chokepoint without stalling the frame rate or clumping into a plug, and the AI switches its unit mix after scouting mine.
- Asymmetric balance, which took Blizzard years and a professional scene to reach and cannot be prompted
- Pathfinding at scale — a per-unit A* recomputed on collision will melt at a fraction of the unit count, and the fix is a different algorithm, not an optimisation
- An AI opponent worth playing, since RTS bots need build orders, scouting, army composition responses and macro management simultaneously
- The campaign, the briefings and the voice work, which is where most of the memory of this game actually lives
- Multiplayer with lockstep determinism, which is how RTS games sync hundreds of units over a modem and is unforgiving of a single floating-point discrepancy
One shared flow field per destination instead of one path per unit
It inverts the cost model. Per-unit pathfinding costs you the unit count multiplied by the search cost, which is why every naive RTS falls over as the army grows. A flow field costs one map-wide pass per destination no matter how many units use it, so a hundred units moving to the same place is the same work as one. They also arrive as a mass rather than a single file, which looks better and is closer to what the player meant when they dragged a box round their army.
ScopeA weekend— build just this, not the game
Build a flow field pathfinding demo as a single self-contained HTML file. Canvas 2D, no dependencies. - A 100x100 grid with walls you can paint in with the mouse, and a destination you set by right-click. - Compute a cost field from the destination outward with a breadth-first or Dijkstra pass, then derive a direction vector per cell pointing downhill. - Spawn up to 1,000 agents that steer by sampling the field at their position, with local separation so they do not overlap. - Render the field as arrows or as a heatmap on a toggle, and show the recompute time in milliseconds when walls change. - A comparison mode that switches the same agents to individual A* and displays both frame times side by side. Definition of done: 1,000 agents run at full frame rate on the flow field while the A* comparison mode visibly collapses, and moving a wall recomputes the field in under 10ms.
Not a clone — StarCraft re-themed and cut down to something you can finish. The angle is the important reel: it is what turns a project into a weekend. 5,880 ways to land.
Shogun Turn Order
StarCraft, but turn-based instead of real-time, set in feudal Japan. Rendered in CRT arcade with scanlines.
Build a browser game called "Shogun Turn Order". The pitch in one line: a rts in the spirit of StarCraft, but turn-based instead of real-time, set in feudal Japan. WHAT THIS IS Take the SHAPE of StarCraft — its core loop, and whatever makes that loop good — and rebuild it as your own game in a different world at a smaller scope. Use your own names, art and audio, which you want anyway: a reskin of somebody else's game is less interesting than a new one that works the same way. THE ANGLE — this is the important constraint, honour it above everything else Turn-based instead of real-time. Time only advances when the player acts. Removes every timing, netcode and animation-blending problem and turns reflexes into decisions. This is what makes the project finishable. If a decision would push the scope back towards the original's, take the smaller option every time. THE THEME Feudal Japan. Duels decided before the first strike, weather and season as mood, and restraint in everything including the interface. The theme is not a coat of paint. Let it change what the mechanics mean — a reload, a health pack and a locked door should all be things that make sense in this world and nowhere else. LOOK CRT arcade with scanlines. Post-process pass with scanlines, barrel distortion, phosphor bloom and slight chromatic aberration at the edges. Keep it subtle enough to read text through. REQUIREMENTS - Runs in a browser with no build step. Canvas 2D unless the angle genuinely requires 3D, in which case three.js. - Fixed-timestep update loop with interpolated rendering, so it behaves the same at 60Hz and 144Hz. - Every tuning constant in one CONFIG object at the top, exposed as live sliders in a debug panel. You will find the feel by dragging those, not by prompting. - All content — levels, entities, balance numbers — in JSON or plain data files, never inline in the logic. - Audio generated with the Web Audio API rather than asset files. - Respect prefers-reduced-motion: keep the fades, drop the shake and the parallax. BUILD ORDER 1. The core loop with placeholder rectangles, at the reduced scope the angle demands. No theme, no art, no audio. Make it fun as rectangles first. 2. The angle's consequences. Cutting to bots, or to turn-based, or to one level changes the design rather than just shrinking it — find out how and rewrite step 1. 3. The theme, as content and data. 4. The look, applied last as a rendering layer over a game that already works. Start by writing a one-page plan: the core loop, what the angle removes and what that frees you to do properly, and what you are deliberately leaving out. Wait for me to approve it before writing code.
Half the time one of these already does what you wanted, and the other half it saves you a week of solving a problem someone documented in 1997.
One count per game. No account — it is tied to a salted hash of your IP, which is not stored in a form anyone can reverse. Every build adds 60 team-years to the counter on the homepage.
Then try these
Questions
How hard is StarCraft to rebuild with an AI agent?
Yes — every game is buildable at some scope. The question is how hard, and StarCraft is BRUTAL (5 of 5). The hard part is not game design at all. Three factions that share no units and still have to be equal is a balance problem measured in years, and underneath it sits the part nobody budgets for: two hundred units pathing through each other at sixty frames a second, where the naive approach dies at forty units and the correct one is flow fields plus local avoidance plus formation logic. Reckon on a project to something you can actually play — not to something that matches the 60 team-years Blizzard Entertainment spent.
What makes StarCraft hard to rebuild?
Three factions that share no units and still have to be equal is a balance problem measured in years, and underneath it sits the part nobody budgets for: two hundred units pathing through each other at sixty frames a second, where the naive approach dies at forty units and the correct one is flow fields plus local avoidance plus formation logic. That is why it sits at BRUTAL rather than a tier either side of it. If you want the idea without the project, the transferable part is one shared flow field per destination instead of one path per unit — a weekend of work.
How long does it take to build a StarCraft clone?
Wrong question, and it is worth saying why. Nobody rebuilds StarCraft — the original took Blizzard Entertainment on the order of 60 team-years, released in 1998 on PC. What you can do is reach something playable and genuinely yours in about a project, and then keep going for as long as it stays interesting. Any figure that claims otherwise is measuring a prototype and calling it a game.
What do I lose by building my own StarCraft?
Asymmetric balance, which took Blizzard years and a professional scene to reach and cannot be prompted. Pathfinding at scale — a per-unit A* recomputed on collision will melt at a fraction of the unit count, and the fix is a different algorithm, not an optimisation. An AI opponent worth playing, since RTS bots need build orders, scouting, army composition responses and macro management simultaneously. Those are the parts a prompt will not hand you.
Which AI coding agent should I use to build it?
Any of Claude Code, Codex or Cursor will handle this prompt. Paste it as your first message, let the agent scaffold the project, then iterate on feel — the second, third and tenth prompts are where a game actually gets good. Work in an empty folder or a fresh git branch so you can throw it away cheaply.
One email when new games go on the list. No other reason to email you, ever.