14. Sandbox, Challenges, and achievements

Sandbox#

A Sandbox is a Board outside any Company. It runs the same Simulation as a Company. Nothing in a Sandbox changes a Company, and nothing in a Company changes a Sandbox. There are two kinds: guided and free. ADR 0003 records why.

The Sandbox fixes one problem. A Company teaches each mechanic when the player first needs it, so a player who wants to learn zones or GC pauses must first play the Company to Series A. The Garage Era is the tutorial for a new player (Progression).

Guided Sandboxes#

A guided Sandbox teaches one mechanic through two to four ordered objectives, like a Factorio tutorial level.

• Start. A fixed Design with the problem already on the Board, or one that appears within the first minute: a red card, a long tail, or Stranded requests.

• Objectives. Each objective is one sentence that states the goal, for example "Keep p99 under 300 ms for 2 minutes." It passes when its rule on the Ledger holds for 90 to 180 sim seconds without a break. A break starts the hold again.

• Escalation. Later objectives raise demand, add a Burn budget, or script an Incident, so the first fix stops being enough. An Incident the Design must survive fires again after each failure. An Incident the player must answer by hand, such as a slow query, fires once.

• Palette. The research the Sandbox grants. It holds the fix, plus at least one wrong option that a reasonable player would try. The wrong option plays out, so the player sees why it fails.

• Demand and seed. Fixed and flat, with no daily cycle. Users do not grow or leave.

• Visibility. The Sandbox grants the visibility step that the mechanic needs. A player cannot learn tail latency without seeing the p99.

• Money. None. Placing and resizing are free; Burn shows, because some objectives set a budget.

• Length. Varies with the mechanic. At 1× speed the fastest pass takes 4 to 11 sim minutes, and a first attempt takes longer. Later tracks are longer, as later Factorio tutorials are.

A guided Sandbox has no text overlays, arrows, or step-by-step prompts. The Board, the Node cards, and the Tradeoff cards work as they do in a Company. The game teaches through consequences (Vision).

Results are pass or fail. There is no score, no histogram, and no server. Nothing pays Cash, Insight, or Legacy, because a reward would make the Sandbox a required grind (pillar 5).

Sandbox tracks#

A Sandbox track is the ordered set of guided Sandboxes for one Era. Every track is available from the start. Inside a track, each Sandbox opens when the player passes the one before it.

The list is a balance target. Each track arrives with the milestone that adds its mechanics (Scope and roadmap). All six tracks hold 18 Sandboxes in all, and a test plays each one with its fix and with a wrong option.

Free Sandbox#

The free Sandbox is a test bench.

• Every Component, Stack, research item, and visibility step is unlocked.

• Cash is unlimited. Burn still shows, so the player sees what a Design costs.

• The player sets the users and an Era's Request kind mix, sets the speed, and triggers any Incident in the catalog.

• There are no Users to grow, no Funding rounds, no Team, no Dev Hours, and no Insight. A free Sandbox is Design and traffic only.

• The player can copy the current Company's Design into a free Sandbox, like a staging environment, to test a change before making it in the Company.

• The free Sandbox keeps its Board in local storage between visits, apart from the Company save.

• Achievements do not count in any Sandbox.

Challenges#

A Challenge is a fixed puzzle, separate from the company. It gives a brief, a fixed demand, a fixed palette, a fixed budget, and a duration. The player builds a design and runs it. A Challenge is complete when the design meets its targets.

This mode follows Opus Magnum: a crude solution is easy, and the depth is in improving it (Influences).

Scoring#

Every completed Challenge scores three axes. Lower is better on all three.

Each axis has a histogram of every player's best result. The player sees where they fall, for example "Cost: better than 72% of players". The game never shows a rank or a top-10 list. The question is whether the player can reach the 70th percentile, not whether they can be first.

The three axes conflict. A cheap design has a long tail. A short tail needs more Nodes. So the player picks a direction, and each player's best results usually come from different designs.

Rules#

• The first Challenge is open from the start. Completing a Challenge opens the next in the global list, including across Era boundaries. Completed Challenges remain open for replay; the daily drill is always open.

• A Challenge's palette is fixed. Legacy perks do not apply. Every player has the same tools.

• The seed is fixed. Incidents in a Challenge are scripted and the same for everyone.

• Results upload as a design plus a seed, and the server re-runs them to verify. The server score is the authoritative one, and a score on the player's device is provisional until the server confirms it. The simulation is deterministic (Simulation).

• Rewards: a badge, and Legacy on first completion (1 to 10 ★ by difficulty). No Cash, no power.

Challenge list#

Each Challenge is inspired by a real event or a real engineering problem. Briefs name the event as an inspiration; the numbers are the game's.

The daily drill#

Every calendar day (real day, UTC), all players get the same short Challenge: a seeded design with a seeded incident. Results go on a daily histogram. The drill takes 5 to 15 minutes. It is optional and never rewards streaks, so missing a day costs nothing.

Achievements#

Achievements reward knowledge, unusual builds, and real engineering moments. They do not reward hours played. Some unlock Mutators or Ventures (Exit and Legacy).