QuorumWiki
Jugar
Diseño

13. Exit and Legacy

The prestige layer: Exits, Legacy, Ventures, and Mutators.

Esta página aún no está traducida al español.

A company ends in an Exit, and the Exit pays Legacy. Legacy buys permanent perks. Every Venture and Mutator is open when the player founds a company, so the rules of a company are a choice rather than a purchase. This is the idle game prestige pattern (Influences). Ventures and Mutators keep each company different, which answers the Upload Labs complaint about repetition.

Exits#

ExitAvailable fromPayout basisMultiplierNotes
Acqui-hireSeries AEquity × min(valuation, $50M)0.5Ends early. Good for a fast restart with a new Venture.
AcquisitionSeries BEquity × valuation1.0The buyer pays the current valuation.
IPOPre-IPO checklistEquity × valuation at listing1.5Continues into the Public Era.
Step downPublicEquity × current market value1.5, + quarter streak bonusEnds the company from the Public Era.
payout=equity×valuation×exit multiplierlegacy=floor⁡(payout/$1,000,000×founding multiplier)\begin{aligned} \text{payout} &= \text{equity} \times \text{valuation} \times \text{exit multiplier} \\ \text{legacy} &= \operatorname{floor}\left(\sqrt{\text{payout} / \$1{,}000{,}000} \times \text{founding multiplier}\right) \end{aligned}
PayoutLegacy
$1 million1
$100 million10
$1 billion31
$10 billion100
$100 billion316

The table is for a founding multiplier of 1. The multiplier is the Venture's Legacy multiplier times one plus the sum of the Mutators' shares (Ventures, Mutators).

The square root is the common prestige curve. Each Exit is worth more than the last, but not exponentially more, so the player keeps a reason to improve rather than repeat (Balancing).

What resets. Cash, Insight, Dev Hours, users, research, the board, the team, the product levers, and Equity.

What stays. Legacy, Legacy shop purchases, Playbooks, achievements, Challenge results, and the engineering glossary entries the player has discovered.

The Legacy shop#

PerkLevelsCost (★) per levelEffect per level
Angel network52, 5, 9, 18, 36Starting Cash ×2
Alumni network53, 7, 14, 27, 54+1 starting engineer (from the Pre-seed Era)
Institutional knowledge106 eachKeep one researched item of your choice through the next Exit
Pattern recognition55, 9, 18, 36, 72Insight +20%
Founder's reputation55, 9, 18, 36, 72User growth +10%
Better terms57, 14, 27, 54, 108Dilution per round −2 points
Night shift36, 17, 51Dev Hours +10% per level: the team runs a second shift
Time dilation223, 90Unlocks 20× and then 40× speed
Playbooks46, 17, 51, 152Playbook slots: one free, then 2, 5, 10, 25 (Playbooks)
Autopilot311, 34, 101Policies, each a switch in the Team panel: queue the cheapest Study; hire while payroll is under a quarter of revenue; buy the next size for App Servers above 70% load
War stories39, 27, 81Open action items raise incident odds by half as much
Head start328, 84, 253Start the next company in the Pre-seed Era, then in Seed, then in Series A, with that Era's Cash, Equity, Users and research, and a Design that has a Cache (and a Load Balancer behind a second App Server from Series A)

The twelve perks cost 1,766 ★ in all. Costs grow so that a full shop takes several Exits, which is the target for 80 to 150 hours.

Playbooks#

A Playbook is a saved sub-design: Nodes, links, Stacks, and settings. It plays the role of a Factorio blueprint (Influences). A Playbook does not carry Product levers, because a lever is product state rather than hardware.

  • The player saves a selection on the board as a Playbook.

  • Placing a Playbook places all its Nodes in one action, if the player has the Cash and the research. Missing research shows as ghost Nodes.

  • Playbooks survive Exits. The second company's first hour is faster because the player places a known-good Garage design in one action.

  • Playbooks export as a text string, so players can share them.

Playbooks remove the post-prestige tedium that Upload Labs players report (Influences).

Ventures#

A Venture is the kind of company the player founds. Each Venture changes the request mix, demand shape, constraints, and incidents. B2B SaaS is the default, and every Venture is open at founding. A Venture sets the mix, and it can cap a Product lever at its first step when the product change would erase the Venture's constraint (Progression).

The mix is the Venture's own traffic, not a multiplier on the Era's. A profile gives every request kind a share and sums to one, and each Era masks it to the kinds that Era serves: the Garage opens on pages and reads for every Venture, and a Series A company carries video only once the Video Feature ships. The profile is in catalog/ventures.ts and its shares, kind by kind, are on Ventures.

VentureDemand shapeTrafficConstraintSignature incident
B2B SaaS (default)Evening peak (20:00), weekend 1×The Era's mixSLA credits from Series BBig tenant hot shard
Consumer socialEvening peak, weekend 1.15×, viral spikesFeeds and video; reads far outnumber writesVery low value per request; huge volumeCelebrity post
FintechBusiness hours, month-end peaksCheckouts and writes; a webhook per settled chargeAny stale on money kinds is a compliance fineDouble charge
Streaming mediaEvening peak, premieresVideo is most requests; egress decides the marginEgress dominates the billPremiere spike
Game backendEvening peak by regionRealtime messages; a state write for each sessionrt Apdex T 50 ms; regions matter from Series ALaunch day
AI startupFlat, with launch spikesPrompts and uploads; batch jobs and reportsGPUs are scarce; inference cost decides marginGPU shortage
E-commerceDaily, with a yearly eventCatalogue pages and site search; checkouts are a small shareOne scheduled sale event per calendar month at ×20 demandFlash sale
Developer platformFlat, bursty clients (peak 20:00, ±10%)API reads and webhooks; the dashboard is a small shareNo naive retries on any Link: customers write their own retry loopsRetry storm from a customer
HealthcareBusiness hours (14:00), weekend 0.4×Reads and writes; patient lookups and chartsData Components stay in the home region from Series AAudit finding
Ad techFlat, enormous (peak 20:00, ±5%)Bid requests; crawlers and verification followA request whose mean passes 100 ms earns nothing (real-time bidding deadlines)Latency regression
Bootstrapped indieAnyThe Era's mixNo funding rounds. Equity stays 100%. Salaries from revenue only.Runway
Government contractBusiness hours (14:00), weekend 0.2×, deadline spikesPages and records; forms that arrive as uploadsMust use Java or C# for the core service; long procurement for new Components (placement takes 1 calendar day)Deadline day

Every profile gives every kind a share, because a kind with none would dry up its Insight pool and stall a default recipe (Progression). The shares follow the shape each kind of company runs in production: an OLTP system is usually half reads and under a third writes, a social feed reads tens of times for every post, a checkout is a fraction of the requests that lead to it, and a bid is answered inside 100 ms or not at all. The value multiplier keeps each profile earning what the slice tuned for that Venture, so the shapes change the Design the Venture needs rather than the money it makes.

Four Ventures cap a lever. Fintech caps buy at the first step, because a regulated checkout cannot become an upsell. AI startup caps ai, because GPUs are scarce. Healthcare caps upload, because patient data stays minimal. Government contract caps push, because a public service does not send marketing. A capped lever still has its first step, so the pool is not stuck; it just cannot become the product's main surface.

Every Venture has its demand shape and its rule. Developer platform, Healthcare, Ad tech, and Government contract also have their signature incident. So do the other Ventures, with one exception: Bootstrapped indie's row names Runway, which is a figure rather than an incident, so it has none. Four of those eight enforce the rule their row names: Bootstrapped indie's no-funding rule, and the three below. Consumer social's celebrity post is the Hot key incident, and its own row names what the player meets.

What each Venture starts with:

VentureCashEquityStack and researchLegacy ×Signature incident
B2B SaaS$300100%The Garage's1.0×Big tenant hot shard
Consumer social$300100%The Garage's1.1×Hot key
Fintech$300100%The Garage's1.3×Double charge
Streaming media$300100%The Garage's1.2×Premiere spike
Game backend$300100%The Garage's1.3×Launch day
AI startup$300100%The Garage's1.3×GPU shortage
E-commerce$300100%The Garage's1.2×Flash sale
Developer platform$300100%The Garage's1.35×Retry storm
Healthcare$300100%The Garage's1.4×Audit finding
Ad tech$400100%The Garage's1.3×Latency regression
Bootstrapped indie$300100%The Garage's1.4×None
Government contract$900100%Tomcat, and lang-java researched1.5×Deadline day

All twelve Ventures start in the Garage, and the Head start perk still opens Pre-seed.

The game owns Cash, Equity, the demand curves and amplitudes, the traffic profiles, the market and value multipliers, and the Legacy multipliers. This document sets the rest: the 100 ms bid deadline, the Java-or-C# rule, calendar-day procurement, Series A residency, the no-naive-retry rule, the Series B SLA, realtime regions from Series A, scarce GPUs, and the signature incidents. The SLA's one SLO level and the AI startup's two GPU Nodes are the game's own.

Each signature incident hits for a few minutes:

  • A hot shard costs three times the CPU per query.

  • A premiere lifts video four times.

  • A launch day lifts realtime six times.

  • A flash sale lifts checkouts ×20.

The rule of each Venture is enforced where the player meets it. Developer platform refuses naive retries on a Link (retryRefusal). Healthcare refuses to place a data Component outside the home region from Series A (venturePlaceRefusal). Ad tech's requests earn nothing once the kind's mean latency passes its deadline (earnsInTime). Government contract refuses a runtime that is not Java or C# (runtimeRefusal), and a bought Component runs a calendar day after it is paid for (placedLife).

Three of the older Ventures have their row's rule too. B2B SaaS signs its SLA from Series B: the availability target it is held to moves one level above the Era's (sloOf), so its error budget is the contract's and a missed week is a breach. With Enterprise sales researched, that breach pays credits back at 10% of subscription revenue while the week is short (Economy). Game backend keeps one Realtime Hub per region from Series A (venturePlaceRefusal): realtime is served where the players are. AI startup may hold two GPU Nodes, whatever their size, and resizes them rather than buying more (venturePlaceRefusal): GPUs are scarce.

The other four rows name a rule that is not a refusal or a target. Each runs in the Simulation:

  • Consumer social's constraint is its market shape: a little over half the value per request and twice the market (value 0.55 and market 2 in catalog/ventures.ts, read by features.ts), and viral spikes come twice as often (spikeWeight).

  • Streaming media's egress dominates the bill: video is more than a third of its requests, and every byte leaving the Design bills three times the price (egress). The Egress Shock achievement asks for the bill.

  • Fintech's compliance fine charges 20 times the value of a request for each stale answer to a money kind (read, write, buy), as fines on the Ledger, taken from Cash beside Burn.

  • E-commerce's sale is on the calendar: every 30 calendar days, at noon, for two calendar hours, buy demand is ×20 (saleOf in calendar.ts). The Flash sale stays the Venture's signature incident, at the same ×20.

A signature incident weighs four times as much for its own Venture (SIGNATURE_WEIGHT) and keeps a weakness of 2 (SIGNATURE_WEAKNESS), so it can fire on a well-built Design. Six incidents exist only as signatures: Latency regression, Deadline day, Big tenant hot shard, Premiere spike, Launch day and Flash sale. No other Venture's weakness can start them. Retry storm, Audit finding, Hot key, Double charge and GPU shortage stay general incidents, so Developer platform, Healthcare, Consumer social, Fintech and AI startup meet theirs four times as often.

Ad tech serves the bid request kind the design names (value $0.001, Apdex T 100 ms). The kind arrives with the Series A mix, and three of every ten requests Ad tech serves are bids, so its Earn per request follows from the kind rather than from a page/read stand-in. Its 100 ms deadline is still checked against the mean latency of every kind it serves, and the bid kind's own Apdex target is that same 100 ms.

Each Venture has its own Legacy multiplier (1.0× to 1.5×), shown before founding.

Mutators#

A Mutator is an optional rule that makes a company harder in exchange for more Legacy. Every Mutator is open at founding, so a player can pick the rule that teaches the lesson they want. Each one adds a Legacy multiplier.

MutatorRuleLegacy
Chaos MonkeyInstance failures are four times as likely, from the Seed Era+25%
Hacker News front pageTraffic spikes are three times as frequent+15%
Legacy monolithStart with a large PHP codebase: migrations cost 3× and tech debt starts at 40%+20%
One region foreverMulti-region is disabled+20%
No cachesThe Cache and CDN are disabled+40%
Shared tenancyNoisy neighbours steal up to 30% of an instance's CPU, redrawn every 30 sim seconds+10%
Tight budgetFunding grants are halved+20%
Single languageEvery App Server must run PHP+15%
Postgres for everythingThe Cache, the Wide-column Store, Distributed SQL and the Data Warehouse cannot be placed+30%
Strict SLOThe SLO is one level higher in every Era+20%

"Postgres for everything" is a real engineering argument, and this Mutator lets the player make it. Postgres can serve q-search with full-text search, ai lookups with pgvector, and queues with SKIP LOCKED, each with a lower ceiling than the dedicated Component.

Company history#

Each Exit adds a row to the company history: the Company's name, Venture, Mutators, time played, peak users, Exit type, payout, Legacy earned, final Stack mix, and a snapshot of the board. The history shows how the player's engineering changed over time.