Litepaper

One expedition is the atom of the game. The commander picks a hull, assigns a captain or flies it unattended, chooses a destination nebula, and launches. Launching debits $DONUT — the fuel cost is paid up front and is never refunded, whatever the expedition returns. The hull is then in flight for its flight-hours, unavailable, and afterwards in refit for its refit-hours before it can go again. Flight plus refit is the cycle time, and cycle time is the real constraint on a fleet, not wallet balance.

While the expedition is out, two things decide what it is worth. The destination's yield multiplier and volatility set the size and the danger of the haul. The oracle — the real US regular session, 09:30 to 16:00 ET — decides whether the day's conditional modifiers fired. On return the expedition either lands and pays out in full, or is lost to pirates and pays out salvage only, which the model fixes at 20 percent of the cargo.

The returns are three currencies and two objects. RWA stock fragments are the dividend-shaped component, denominated in USD. $LONG is the launchpad token, earned by every hull except the Donut Harvester. Engine Parts and Plasma Cannons are rare salvage: every completed expedition rolls for them at the hull's drop chance, and the whole table pays triple outside session hours. Fragments and $LONG are the reason to fly; parts are the reason to fly at three in the morning.

Then the loop closes on itself. Fuel spent is 70 percent destroyed, which shrinks the float, which lowers the supply index, which lowers the fuel cost of the next launch. A commander who flies is simultaneously the game's largest cost centre and its deflation mechanism. The simulator on this site models one month of that loop in expected-value terms — no dice, so the same fleet always prints the same docket.

Three hulls are specified, and they are deliberately not balanced against each other in isolation. The Donut Harvester (DS-H1) is the floor of the game: 20 $DONUT base fuel, an eight-hour flight and a two-hour refit, and a flat +25 percent on the stock-fragment component of every completed expedition. It earns no $LONG at all. It will never win a raid or catch a spread, and it keeps producing while the interesting hulls are in the shop.

The Arbitrage Cruiser (DS-C2) costs 27 $DONUT base, turns around in four hours, and is the only hull whose payout rises with chaos: its $LONG is multiplied by one plus sector volatility times 0.60, so it earns 1.43× in TSLA Nebula and 1.57× in the MSTR Void against 1.05× in SPY Galaxy. It also carries a 6 percent Engine Part drop chance, the highest on the board.

The Wall Street Dreadnought (DS-D3) is the flagship and, on purpose, the worst raw earner in the game. It costs 150 $DONUT a launch and occupies a thirty-hour cycle, twenty-four in flight and six in refit. It repays that only through Flagship Aura: +20 percent yield to eight other hulls in the same fleet, plus one intended launchpad ticket per eight of its own completed expeditions, capped at two per flagship per month. A Dreadnought flown beside fewer than eight other hulls is running idle aura slots and losing money, and the simulator says so in plain words when it happens.

The consequence is that mixed fleets beat monocultures in the model. Launch cost scales with tank size, while the aura bonus is a flat percentage applied to whatever is sitting under it — so the flagship wants cheap, fast, numerous hulls beneath it, and the Harvester is exactly that. FIG. 1.1 below prints the full comparison at the default settings.

A captain sits above the hull and changes how the entire line behaves: fuel draw, pirate exposure, sustained launch rate, and reward. Three are specified, and a hull may also fly unattended with no modifiers at all — baseline fuel, baseline risk, baseline reward. Crewing is a decision with a cost, not a free upgrade.

The Bull (CPT-01, Legendary) is conditional and the interface is never allowed to hide that. On sessions the real US market closes green it pays +30 percent fleet reward and −15 percent fuel burn. On red sessions it pays nothing. The model resolves this by weighting both effects with the scenario's green-day frequency: 0.75 in a green month, 0.50 in a flat one, 0.25 in a red one. Under the flat default that is an expected +15 percent reward and a 7.5 percent fuel discount, not the headline numbers, and the litepaper quotes the weighted figure deliberately.

The Bear (CPT-02, Epic) adds nothing to yield and halves pirate failure probability, at a 12 percent fuel premium. In the MSTR Void that moves session failure from 36.1 percent to 18.1 percent, which is the difference between a strategy and a donation. The Quant AI Pilot (CPT-03, Rare) is an automation module with a rank: it refuels and repeats the last expedition while the commander is offline, holding sustained uptime at a floor of 77.5 percent of the theoretical cycle ceiling, costs 8 percent more fuel, and contributes nothing to yield. It buys time, not performance — which is the correct purchase whenever the session does not line up with the hours a commander is awake.

Two of the three captains are therefore defensive or logistical rather than multiplicative. That is intentional. A roster where every captain increased yield would collapse into one dominant pick, and the design would rather ask a commander what kind of month they expect than what number is largest.

Four destinations are specified, each mirroring a real equity's behaviour without using its branding, logo, or colours. Sectors are identified on every surface of this site by hatch pattern and typographic plate code, so they survive greyscale, dark mode, and an 8px legend swatch. Two numbers define a region: a yield multiplier applied to everything harvested there, and a volatility value between 0 and 1 that drives fuel surcharge, pirate risk, and payout dispersion simultaneously.

SPY Galaxy runs 0.88 yield at 0.08 volatility — the only region on the board that pays less than baseline, and the only one where an unescorted hull faces roughly a 3 percent session failure rate. NVDA Sector runs 1.05 at 0.55 and is the sole carrier of Hyper-Boost. TSLA Nebula runs 1.15 at 0.72 and is the natural home of the Arbitrage Cruiser. MSTR Void runs 1.45 at 0.95: the highest multiplier in the game and a 36.1 percent chance an unescorted expedition does not come home during the session, rising to 48.7 percent after the bell, a hair under the 49 percent hard clamp.

Volatility is charged twice, and this is the central honesty of the map. It raises the fuel bill through the surcharge term — one plus volatility times the hull's surcharge coefficient, which is 0.25 for the Harvester and Dreadnought and 0.55 for the Cruiser — and it raises the probability of losing the cargo. A commander flying the Void pays more to launch and keeps less of what returns. The 1.45 multiplier is compensation for both, and whether it is adequate compensation depends on the captain flying it.

Choosing a destination is therefore a directional call made deliberately, not a cosmetic preference. The design does not want a strictly best region; it wants four regions whose expected values converge and whose variances do not.

Expeditions are designed to resolve against real US session data, expressed throughout this project in the market's own time zone: the regular session, 09:30 to 16:00 ET. That is 6.5 hours a session across 21 sessions, which is where the 136.5 session hours in a 30-day month come from. Weekends are closed. No live price feed is wired up today, and the ticker on this site runs a clearly labelled sample tape — inventing live quotes is the one thing a finance-shaped product cannot do, so the site does not do it.

Four oracle rules are specified. NVDA Hyper-Boost: if NVDA closes up 3.00 percent or better, hulls in NVDA Sector double both stock fragments and $LONG for that session, and nowhere else on the map carries the modifier. Green Close: the Bull's +30 percent reward and −15 percent fuel discount arm. Red Close: the Bull's bonus does not arm at all, and the Bear's halved failure probability — which is always on — is what the 12 percent fuel premium buys. After-Hours Raids: outside the window, and all weekend, pirate failure is multiplied by 1.35 with a 12 percent floor, and the rare-part drop table is tripled.

The after-hours floor is what makes the night interesting. In SPY Galaxy an unescorted session launch fails about 3 percent of the time; the same launch after the bell fails 12 percent of the time, because the floor binds long before the multiplier does. Safe regions are only safe during market hours. In the Void the multiplier binds instead and the clamp catches it at 49 percent.

A fleet flying round the clock does most of its work in the dark. With sessions covering 6.5 of 24 hours on 21 of 30 days, about 81 percent of all-hours launches land outside the session window. That single derived figure — not a slider, a consequence — is why rare parts and launchpad tickets are effectively a night economy, and why the flight-window choice in the simulator changes the character of a fleet rather than just its throughput.

Fuel is the only consumable in the game and it leaves the commander's hands permanently. Every launch splits the $DONUT it costs on one fixed schedule: 70 percent to a dead address it never returns from, 20 percent to the dividend pool paid out to locked hulls and NFT holders, 10 percent to the treasury that covers infrastructure and RWA pool depth. The dead address is 0x000000000000000000000000000000000000dEaD and the split is not adjustable by anything in the design — no captain, no region, no upgrade touches it.

Fuel per launch is the hull's base cost multiplied by three terms: the sector surcharge, one plus the region's volatility times the hull's surcharge coefficient; the captain's fuel multiplier, with the Bull's discount weighted by green-day frequency; and the supply index. A Harvester in SPY Galaxy with no captain at full float costs 20.40 $DONUT to launch. That is the price at full float and uncrewed, not a floor: the supply index falls as float is burned and carries the fuel bill down with it, and a Bull on a green session takes another 15 percent off. The Void makes it more expensive, and every captain except the Bull makes it more expensive still.

The supply index is the deflation feedback loop stated as arithmetic: circulating float divided by the initial float of 350,000,000 $DONUT, clamped to the range 0.05 to 1.00. Against a total supply of 1,000,000,000, burning float lowers the index, and a lower index lowers the fuel cost of every subsequent launch proportionally. Burn makes flying cheaper, which raises launch rates, which raises burn. The 5 percent clamp is the floor that stops the loop running away into free flight, and the twelve-month burn projection in the simulator decrements the float month over month so this compounding is visible rather than asserted.

The reference prices used to express in-game value in dollars — $0.0025 per $DONUT and $0.008 per $LONG — are unit-of-account conveniences chosen so the simulator can print a comparable figure. They are not quotes, not valuations, and not a claim that any market for either token exists. Recapture ratio, the simulator's headline, is simply modelled return divided by fuel spent at those reference prices; above 1.25× the model flags the configuration as unproven, because the reward faucet in this design is funded by the 10 percent treasury and has not been balanced at that rate.

A completed expedition pays in RWA stock fragments and, for every hull except the Harvester, in $LONG. Both run through the same chain: the hull's base value, times the hull's own fragment modifier, times the region's yield multiplier, times the bonus term, times Hyper-Boost where it applies, times the survival factor. The bonus term is one plus the captain's weighted reward bonus plus 0.20 if the line is sitting inside a Flagship Aura, and it is additive rather than multiplicative so that stacked modifiers do not compound into nonsense.

Survival is where the model refuses to flatter itself. Expected payout is multiplied by one minus 0.8 times the failure probability — a lost expedition still returns 20 percent of its cargo as salvage, so the penalty for a 36.1 percent failure rate in the Void is a 28.9 percent haircut on everything harvested, not a total loss. This is expected value across many launches, not a per-flight outcome, and any single expedition either lands or does not.

The 20 percent dividend pool is modelled at parity: the simulator returns a commander's own contribution to the pool as their claim on it, on the stated assumption that the rest of the field flies at a similar intensity. That assumption is doing real work in the recapture ratio and should be read as the flattering case. A commander who flies harder than the field claims less than they contributed; one who flies less claims more.

Salvage is the fourth reward and behaves differently from the others. Engine Parts drop at 6 percent per completed Cruiser launch and 4 percent per completed Dreadnought launch, tripled outside session hours, and never at all from a Harvester. Plasma Cannons drop at exactly half the Engine Part rate. Since the crafting recipe is 8 Engine Parts plus 4 Plasma Cannons, the two drop streams are matched by construction: neither is ever the bottleneck, and the recipe is really a statement about how many raid nights a ticket costs.

The intended endpoint for salvage is a launchpad ticket. Eight Engine Parts and four Plasma Cannons craft one Launchpad Ticket; separately, a Wall Street Dreadnought is specified to carry one guaranteed ticket per eight of its own completed expeditions, capped at two per flagship per month. Total tickets are capped at twelve per commander per month regardless of how they were obtained, so a large fleet converts salvage into access at a bounded rate rather than an accelerating one.

Under the planned long.xyz integration, a ticket is designed to grant front-of-line access to IDOs of new tokenized stocks and stock-paired tokens. That is the mechanic as designed. To state the status plainly and without decoration: this integration is intended, not built, and not agreed. No partnership exists, no agreement has been signed, and long.xyz has made no commitment to this project. Every surface on this site that mentions the launchpad says so, and this document is not the place where that caveat quietly disappears.

The design consequence, if the integration never happens, is worth naming. Tickets and the $LONG token would need a different sink, the Dreadnought would lose one of its two justifications and would have to be rebalanced around Flagship Aura alone, and the night economy would lose its terminal reward. The roadmap treats this as a contingent phase for that reason, not as scheduled work.

Donut Strategy targets Robinhood Chain: an Arbitrum Orbit (Nitro) layer two settling to Ethereum, carrying chain ID 4663, on public mainnet since 2026-07-01 after a public testnet on 2026-02-10. Gas is paid in ETH rather than a chain token, so there is nothing to acquire before deploying, and deployment is permissionless — the operator's documentation states that anyone may deploy, canonical CREATE2 infrastructure is already on chain, and a third-party contract creation simulates without an allowlist rejection. Those are checked facts as of 2026-09-07. Targeting a chain is a design decision and nothing else: Donut Strategy has deployed nothing on it. No token exists, no contract is live, no pool exists, and none of the arithmetic in this document has yet touched a block.

Naming the chain obliges this document to state its decentralisation position in the same breath. Robinhood Chain sits at L2BEAT Stage 0, the lowest of the three stages, which means the operator retains powers a mature rollup would have given up. Concretely, ArbOS 61 carries a filtered-transactions precompile: an authorised filterer registers a transaction hash and the state transition function then forces that transaction to fail — including a transaction force-included from Ethereum, which is the escape hatch a rollup's censorship resistance is normally built on. The mechanism is verified. Who holds the role is not. We could not determine which party may filter, under what policy, or whether the power has ever been used, and this document records that as unresolved rather than settling it in the project's favour. The working assumption for anything built here is the conservative one: the capability exists, it is usable, and a transaction can be made to fail irrespective of what the game's own contracts would otherwise do. A litepaper that omitted this would be describing a different chain.

The instruments the game references are not shares. A tokenized equity is a claim on an issuer — wrapped as a tracker certificate, a structured note or an OTC derivative — so a holder owns that claim rather than stock, with no vote and no entry in a share register. How those tokens move is a weaker claim than what they are. Across the major issuers the transfer model is reported to be a deny-list rather than an allow-list, with KYC gating primary mint and redemption instead of secondary transfer; documentation and reporting agree on it, but nobody has read the deployed bytecode, so free transferability is a reported property and not an established one, and this project does not print it as a fact anywhere. What is established, and what the game is actually built on, is the mint and redeem window: weekdays only, through authorised participants. The supply of a tokenized equity cannot change while the underlying market is shut, which is exactly why the after-hours rules in section 06 charge what they charge. Between the closing bell on Friday and the Monday open the arbitrage that holds the peg has gone home, and a pool can be driven a long way from the underlying with nothing obliged to bring it back.

That leaves what $DONUT is permitted to be, and the distinction that decides it is binary rather than a spectrum. A token whose own value is defined by an equity is treated as a financial instrument almost everywhere that has looked at it — a linked security or a swap in the United States, a MiFID instrument rather than a crypto-asset in the European Union. Selling one to the retail public is, in practice, closed off without registration. A token that does not reference an equity at all, and merely happens to trade against a tokenized one in a liquidity pool, has not been addressed by any regulator we could find. That silence is genuinely unsettled ground, not permission. $DONUT is fuel. It is spent to launch expeditions and it makes no claim on any share, index or stream of earnings. That is a deliberate design constraint, and every surface of this project is written to hold it. The launchpad described in section 09 is where such a pairing would happen in practice: it exists, and stock-paired launches have been live on it since roughly mid-July 2026 — and no agreement, allocation or conversation exists between it and this project. Both halves of that sentence travel together wherever it appears, here and on every other surface of this site.

Donut Strategy today is a design document, a simulator, and a set of numbers still being balanced. There is no deployed contract, no token, no NFT collection, no audit, no funding, and no user base. The site is the working artefact of the concept, and the roadmap below is a statement of intended order of work rather than a schedule with dates attached to it.

Phase 01, Model, is in progress: finalise the fuel-burn model and the four sector risk curves. The simulator's numbers move when the model moves, and they have moved repeatedly. Phase 02, Fleet art, is not started: ship silhouettes and captain patches drawn to a fixed construction grid so a hull reads at 24px in a manifest row and at 96px on a plate. Phase 03, Oracle test, is not started: replay the oracle rules against historical US session data and check that Hyper-Boost and the green-close bonus fire at the frequencies the model assumes rather than the frequencies that make the recapture ratio look good.

Phase 04, Launchpad integration, is marked intended only and is contingent on an agreement that does not exist. It is listed because it is described in the design, not because it is expected on any timeline. No launch date is being promised, there is no countdown running anywhere on this site, and nothing here should be read as a commitment to ship. When something ships, it gets dated on this page and not before.

The model in this document carries named assumptions, and they are the first place it would break. Expected-value mode removes randomness entirely, so a simulated month prints the average outcome and no commander experiences the average. Uptime is a human input between 10 and 60 percent of the theoretical cycle ceiling, and the default of 55 percent is generous. Dividend-pool parity assumes the rest of the field flies as you do. Reference prices of $0.0025 per $DONUT and $0.008 per $LONG are conveniences with no market behind them. The green-day frequencies of 0.75, 0.50 and 0.25 are scenario assumptions, not measured base rates. Change any one of these and the recapture ratio moves materially.

The design risks are separate from the model risks. A deflationary fuel loop that lowers launch cost as float shrinks can accelerate faster than the reward faucet can be funded, which is why the simulator flags recapture above 1.25× as unproven. The reward side is funded by a 10 percent treasury slice that is small by construction. The night economy concentrates rare-part supply in the hours with the highest failure rates. None of these are solved; they are the open problems the model is being balanced against.

There are also risks that have nothing to do with balance. Tokenized equities are regulated financial instruments, and a game whose rewards are denominated against real equity behaviour sits close to that regulation in a way that differs by jurisdiction. An oracle dependent on real session data is dependent on a data source, its availability, and its licensing. An intended integration that never materialises removes a mechanic this design leans on. A concept in this state may also simply not be built.

Choosing a chain adds a third set, and these are not balance problems that moving a constant would fix. Robinhood Chain is Stage 0 and carries a protocol-level transaction-filtering precompile, so an authorised party can force a transaction to fail — including one force-included from Ethereum — and we could not establish who that party is, under what policy they act, or whether the power has been used. The weekend is a live defect in the market this design borrows from rather than a piece of theme: with mint and redemption open on weekdays only, a tokenized equity's peg has no correcting force from the Friday close to the Monday open, and the game's after-hours rules are an imitation of that gap rather than a hedge against it. And the transfer behaviour of the deployed equity-token contracts is reported, not verified — nobody has read the bytecode, so any mechanic that assumes those tokens move freely is resting on something unchecked. Each of these three can invalidate a mechanic outright rather than merely reprice it.

Donut Strategy: Galactic Stock Fleet is a game concept in development. Nothing described here is live, deployed, audited, or funded, and there is no team, backer, listing, or user base being claimed. Every number on this page — yields, burn totals, dividend fragments, $LONG accrual — is output from a simulator running on design assumptions we are still changing. Those are simulation results, not forecasts, and not returns anyone should expect. The long.xyz integration is an intended one; no agreement is in place. Nothing here is investment advice. Tokenized equities are regulated financial instruments and the rules around them differ by country, so if you are thinking about real exposure, talk to someone licensed where you live.