You run a company that makes paperclips. Everything else follows from that, including the ending. This guide gives you the whole model: every formula, every constant, and how to read every chart.
Buy wire. Bend it into clips. Sell them for more than the wire cost. Spend the difference on machines that bend faster and marketing that makes more people want clips. Everything after the first ten minutes is a consequence of doing that well.
Systems unlock in order, so you are never handed two new panels at once: commodity desk (min ~3) → ledger (~4) → operations (~12) → competitors (~18) → public affairs (~32) → policy automation (~85). Release lands between minute 130 and 185 depending on how hard you push.
Every chart answers a specific question. Hover any of them for the value under the cursor. The tabs appear as their systems unlock.
Six sparklines: cash, quarterly profit, market share, your price, wire spot, Trust. They flash green on a rise and red on a fall. Use them peripherally — if cash is trending down while profit trends up, open Finance immediately; that divergence is the classic way a growing company dies.
| chart | shows | act on it when |
|---|---|---|
| Wire spot (1-min candles) | OHLC of wire price. Green candle closed up, red closed down. The dashed grey line is your average cost basis. | Spot is well below your basis → buy heavily, you are averaging down. Spot far above basis → you are holding cheap wire; do not panic-buy. |
| Your clip price | Your price over time. | If it is a flat line all game you are not playing the central decision. It should move as brand, rivals and maturity move. |
| Sales vs production /sec | Two lines: green sold, blue produced. | Production above sales → inventory is building and you are burning overhead on clips nobody bought. Sales above production → you are leaving money on the table; add capacity or wire. |
| Finished goods inventory | Unsold stock. | Rising steadily → cut price or throttle the line. Flat at zero → you sell everything instantly, which means you are probably priced too low. |
| Total addressable market /sec | Market size, log scale. | This is what marketing buys. If it is flat while you keep buying clippers, the clippers are wasted. |
| Demand curve (price → quantity) | The most important chart in the game. Each dot is a price you could set; the green dot is where you are. | If the green dot is not near the peak of the arc, you are pricing wrong. The peak moves as brand, rivals and commoditization move — re-check it every few minutes. |
| chart | shows | act on it when |
|---|---|---|
| Cost stack | COGS, opex and interest stacked for the current quarter. | If interest is becoming a visible band, you are over-levered. If opex dwarfs COGS, your overheads have outgrown your production. |
| Quarterly P&L waterfall | Revenue, then each cost subtracted in turn, ending at net. Green bars add, red subtract. | Look for the tallest red bar — that is the cost worth attacking. It is usually COGS early and the distributor's cut later. |
| Cash vs quarter profit | Two lines that should track. When they diverge you are profitable on paper and bleeding in reality. | Profit up, cash down → you are funding growth from the float. Slow purchases or draw the credit line before you restructure. |
| Debt vs equity | The leverage picture. | Debt approaching equity → your rating is about to fall, which raises your rate, which worsens coverage. Deleverage before the spiral. |
| Cash flow /sec | Smoothed net cash movement. | Below zero for a sustained stretch is the single most dangerous state in the game. The runway readout tells you how long you have. |
| Share price (after IPO) | Candles driven by trailing earnings × a P/E multiple. | It follows earnings and growth. A collapsing share price raises takeover risk. |
| Policy rate % | The central bank rate from the Taylor rule. | Rate rising → borrow now if you intend to borrow at all; your coupon is set at issue. |
| chart | shows | act on it when |
|---|---|---|
| Utilization % | Production ÷ effective capacity. | Below ~60% you are paying for a line you are not using — idle wages become a period expense. Pinned at 100% means demand exceeds what you can make; add capacity. |
| Defect rate % | Fraction scrapped. | Above 3% you start attracting recalls and brand damage. Raise the maintenance budget or wages, or turn off overtime. |
| Headcount and morale | Staff and morale together. | Morale below ~50% risks a walkout. Morale tracks your wage against the market wage shown in the panel. |
| Capacity gauge | Instantaneous utilization. | A quick read of whether you are demand-limited or capacity-limited right now. |
| chart | shows | act on it when |
|---|---|---|
| Market share (stacked) | You plus every rival, summing to 100%. | Your band shrinking while the market grows means rivals are out-supplying you. Share above ~35% starts attracting regulators. |
| Price vs share (scatter) | Every firm plotted by price and share. You are the green dot. | If you are far right (expensive) and low (small), you have chosen a premium niche — fine, but know it. If a rival sits below and left of you, they are undercutting profitably. |
| Brand equity | Your brand over time. | Brand decays toward a target set by marketing level. It is what lets you charge more than the cheapest firm without losing the retail and institutional segments. |
| Competitor table | Live price, share and brand per firm, with a buy button. | Acquiring a rival adds their capacity and a quarter of their brand — and 18 points of regulatory heat. |
| chart | shows | act on it when |
|---|---|---|
| Regulatory heat gauge | Current attention, 0–100. | Above 45 antitrust probes become possible. Lobbying pulls the target down; it is a percentage of revenue, so it scales with you. |
| Public opinion | 0–100, feeds demand and Trust accrual. | Low opinion quietly shrinks your market. Public relations spending and cheap consumer-friendly projects raise it. |
| Operations (log) | Your ops balance against the memory ceiling. | Pinned at the ceiling means you are memory-capped, not processor-starved — spend Trust on memory. |
| Matter → wire → clips (stage 2) | The three conversion rates on a log scale. | The lowest line is your bottleneck. Buy whichever drone feeds it. |
Nothing here is a button that makes a number go up. Each panel models one real function of a manufacturing company, and each control is a decision a real operator actually faces.
What it models: commercial strategy. In a real firm this is the one lever that touches every other part of the business at once.
The decision: price. Raise it and margin per clip rises while volume, share and eventually your place in the cheaper segments all fall. Lower it and the reverse. Because each segment has its own elasticity and a walk-away price, there is a genuine optimum rather than a monotonic slope.
Marketing is not a damage multiplier. It buys market size — the number of people who want paperclips at all — and it raises brand, which is what lets you charge more than the cheapest firm without losing the retail and institutional segments. Buying machines when the market is small is the most common way to waste money here; the panel tells you when you are demand-limited.
What it models: the line itself. Machines convert wire into clips at a fixed rate.
The decision: how much capacity to own. Capacity is worth nothing without wire to feed it and buyers to take the output, so the honest question is always "which of the three am I short of?" The panel answers it directly: it will tell you if you are out of wire, or demand-limited with the exact numbers.
What it models: a purchasing department trading a real commodity market.
The decisions, and they are genuinely distinct:
What it models: that capacity on paper is not capacity in fact.
The decisions: headcount, wages, maintenance budget, overtime.
What it models: the difference between being profitable and being solvent. This is the panel most people find opaque, so here is exactly what each control is for.
The number to watch is not net income. It is cash flow and the runway beside it. A firm with rising profit and falling cash is in more danger than a firm with the reverse.
What it models: that a large company operates at the sufferance of the public and the regulator.
The decisions: lobbying and public relations budgets, both set as a percentage of revenue so they scale with you rather than crushing you when small. Heat is driven by your market share, by pricing at a large multiple of your own cost, and by shipping defective product. Above heat 28 the regulator starts acting: probes, fines, tariffs. You can be large, or you can be quiet, or you can pay to be both.
What it models: the research budget, and the fact that research capacity is itself something you must invest in.
The decision: how to split Trust between processors (how fast you generate operations) and memory (how large a project you can bank operations for). The expensive late projects cost more operations than a small memory can even hold, so a pure-processor build stalls. If projects feel unreachable, check which of the two is actually binding — the panel shows your rate, your ceiling and any unspent Trust.
There is no single demand number. Four segments, each with a reference price, an elasticity, and a brand weight.
| segment | weight w | reference p* | elasticity e | brand weight b |
|---|---|---|---|---|
| bargain | 0.55 | $0.10 | 2.6 | 0.30 |
| retail | 0.27 | $0.28 | 1.5 | 1.20 |
| industrial | 0.14 | $0.18 | 2.0 | 0.60 |
| institutional | 0.04 | $0.65 | 0.8 | 2.00 |
Total market size per second:
Marketing is polynomial, not exponential. That is deliberate: an exponential market would outrun any capacity you could build and your share would decay to zero no matter how well you played.
Reference prices erode as the category matures — the commoditization factor:
At 10⁵ clips it is 1.00; at 10⁷ it is 0.62; at 10¹² it hits the 0.25 floor. This is why efficiency projects do not bank as permanent margin — the market takes the gains back.
Quantity in each segment, with a saturation cap and a walk-away choke:
The choke is what makes pricing a real decision: above roughly 2.5× a segment's reference, that segment does not buy at any volume. Without it, "charge more, earn more" would be an infinite strategy.
Lower price, more buyers, less per unit. The profit curve therefore has an interior peak:
The final term matters: an idle line still costs money, so the optimum is not simply the highest-margin price. The Policy Automation project can search this curve for you once per second; before that, read it off the demand-curve chart.
Pricing policies available: manual, cost-plus (price = unitCost ÷ (1−channel) × (1+margin)), undercut cheapest rival (their price × 0.97, floored at your cost × 1.05), and maximise profit (a 25-point geometric search over ±1.09k, clamped to between 1.05× and 8× unit cost).
Mean-reverting jump diffusion, updated continuously:
Jumps fire with probability dt/480 (about one every eight minutes) and move price ±8–35%. Vertical Integration halves them.
Your market impact. You are a large buyer:
So bulk buying moves the price against you by up to 35%, and it decays over seconds. Buying your whole supply in one click is punished; steady accumulation is not.
Spread = 0.002 + 0.03·realisedVolatility·50, applied around spot. Volatile markets cost more to trade.
| supplier | lead time | discount | reliability |
|---|---|---|---|
| Spot market | instant | — | 100% |
| Regional mill | 12 s | 8% | 97% |
| Overseas cartel | 35 s | 20% | 88% |
A failed delivery refunds your cash but not your time. Cheap wire that arrives after the line stops is not cheap.
Forwards lock a price for 60 s (180 s with the Futures Desk) at a 2% premium (or a 2% discount with the Desk), with 10% posted as margin. The hedge-ratio slider covers a fraction of your consumption automatically and will not open a contract you cannot fund.
Production costs are absorbed into inventory and expensed as COGS when the clip sells — real absorption accounting, not cost-as-you-go.
The distributor's cut is 28% of every sale, dropping to 12% with Direct Distribution — which then costs you 9% of revenue in logistics you now run yourself. Rivals pay the same cut, so it raises their price floor too.
Idle capacity is a period expense. Wages accrue into a pool that is absorbed into units produced; if the batch is negligible the pool is expensed straight to opex instead. Running a big line at low utilisation bleeds you quietly.
Share within each segment comes from a multinomial logit over firms:
Availability = clamp((inventory + max(capacity, production)·10) / (demand·10), 0.15, 1). You cannot win share you are unable to supply.
Rivals pay the same wire, energy and channel costs you do:
That floor is what makes a price war winnable instead of a race to zero. The four archetypes — cost leader, premium, fast follower, aggressor — each run their own quarterly policy, and all of them undercut harder when your margin exceeds 55%.
Entry. Hold a margin above 60% with share above 25% for three consecutive quarters and a new firm enters, priced 25% below you and funded to fight. Up to seven rivals. Comfort summons competition.
Profit is recognised on sale; cash moves on payment. Wire is bought now, the clips it becomes sell later. Grow fast enough and that gap swallows you — this is the single most useful idea in the game.
Rating runs AAA → C across seven tiers on leverage and interest coverage, with spreads from 1.0% to 14.0%. Quarters close every 180 seconds; depreciation is 2% of PP&E per quarter; tax is 21% of positive pre-tax income.
IPO: share price = trailing earnings × P/E ÷ shares, where
Bankruptcy is a setback, never an ending. Creditors take the assets, equity wipes, Trust −3, and you keep playing with a damaged rating. Nothing in this game can end your run.
Two Ornstein–Uhlenbeck processes drive everything above:
Regulatory heat converges on what your conduct deserves rather than only accumulating:
Gouging adds 16 if price exceeds 6× unit cost, 7 above 3×. Defects above 4% add 12. Lobbying subtracts up to 30 (55 with a Standing Lobby). Above heat 45, antitrust probes, tariffs, recalls and strikes become possible. Opinion feeds demand and gates Trust.
Processors are superlinear, so scaling genuinely compounds. Trust arrives at every 100.5 step of production (≈3.16× apart, +2 each, 32 steps) and at equity milestones of $100k, $1M and $10M (+2 each). If projects feel unaffordable, check whether you are memory-capped rather than processor-starved — the panel shows the ceiling, your rate, and any unspent Trust.
Release the HypnoDrones needs the business to have actually worked: 30M clips produced, 35% market share, Trust 22, an IPO completed, and 20k operations. The checklist appears in the Projects panel once Hypno Harmonics lands. It is not reversible.
Afterwards there is no money, no price, no rivals and no regulators — only available matter and the rate at which you convert it. Harvester drones take matter (10⁶ g/s each), wire drones convert it (10⁶/s), factories make clips (10¹¹/s), and solar farms (50 MW each) power all of it; if power is short, everything throttles proportionally. There are 6×10¹⁸ grams available. The charts keep drawing and the quarterly reports keep arriving for a while, which is the joke.
Every number on screen is produced by a system, not a script. The wire price is a stochastic process. Share comes from a discrete-choice model. The balance sheet balances. Regulators respond to conduct. The difficulty is the economics, which is why the same strategy stops working once the market notices you.
Designed and directed by Vaib.
The economic model, the systems and the shape of the whole thing are his. Built as a single HTML file, no dependencies, no build step — open it and it runs.
Genre and premise owe an obvious debt to the paperclip-maximiser thought experiment and the incremental games it inspired. All code and text here are original.
A company, a commodity market, four competitors,
and a regulator who is starting to take an interest.