PROTOCOL

There is no mint page. A wallet becomes eligible for one Genesis ship when its qualifying buys exceed its qualifying sells by 0.08 WETH on the canonical market, and the ship is minted in that same transaction.

Netting matters. A gross counter would be farmed by cycling the same small trade, so the ledger is signed: a sell before qualification pushes you back down, and a wallet that sold before it ever bought starts underwater and has to climb through zero first.

After the mint the ledger stops mattering entirely. Selling every token you own does not burn, disable or claw back the ship. The threshold is a qualification, not a lock.

The specification denominates qualification in USD, which needs an ETH/USD feed. Robinhood Chain has none. The ways out were an owner-updated price, which the specification’s own security section lists as a priority risk; the pool’s spot price, which it forbids outright; or denominating in the asset the protocol already holds.

This takes the third. The threshold is a fixed WETH amount set once at deployment and immutable after. The honest cost is that a ship’s dollar price drifts with ETH. The benefit is that there is no oracle to go stale, no keeper to trust, and no path where a broken feed changes who qualifies.

This is the part that decides the architecture. Uniswap v4 lets a hook adjust the specified currency before a swap and the unspecified currency after it. Which of those is WETH flips with the direction of the trade:

• Buy — WETH is the specified currency, so the 2% is skimmed in beforeSwap, before it reaches the curve.
• Sell — WETH is the unspecified currency on the way out, so the 3% comes off the gross output in afterSwap.

A hook that only implemented afterSwap would collect the buy levy in SWARS instead of WETH, and the treasury would stop holding an asset the protocol cannot print. That is the whole economic argument, so the hook carries four permissions and its address has to end in the matching bits.

Exact-output swaps are refused rather than mispriced. On those the levy basis stops being the amount the trader actually brought, and a stated limitation beats quietly wrong accounting.

Revenue accumulates instead of firing a buyback on every trade, because a predictable order on the tape is a gift to anyone watching. An epoch then sorts it into four pockets that can only be spent for what they are for.

Operations can never reach the staking pocket, and nothing in the treasury can reach staker principal at all — that lives in a different contract, in a different token. The bought SWARS goes straight to the contract that pays users rather than sitting in a treasury wallet.

The buyback pays the levy like everybody else. Exempting it would have meant an address the hook treats specially, and that permission is worth more than the 2% it saves.

The canonical pool was seeded single-side: the whole launch float placed in one range entirely below the opening tick, so the position started as 100% SWARS and engaged no WETH whatsoever. Every ETH in the pool arrived from a buyer.

Staking distributes tokens that are already in the contract. Seasons pay out of a budget that was funded before claims opened, checked against that season’s own balance so one season overspending can never reach another’s tokens.

Gameplay happens off chain because putting every raid on chain would make a worse game. What lands on chain is the settlement: a Merkle root over the final scores, replaceable during review and frozen the moment the first wallet claims against it.

A 28-day season pays a share of a bounded pool. ORE never converts to SWARS at a fixed rate, because that would make emissions grow with gameplay production.

• The levy is not universal. It applies where the hook applies. Anyone may open a second pool and trade around it. SpaceWars answers that by concentrating liquidity, qualification and routing on the canonical market, not by taxing the token itself.
• The threshold is not Sybil resistance. It guarantees a defined level of net acquisition before a mint. It does not guarantee one human per ship.
• Buybacks are not a floor. They can support recurring demand while there is revenue. They cannot guarantee a price, and the system is built to pay less rather than borrow against a promise.
• There is no fixed APY, and the only figure worth quoting is the rate currently funded.

No administrator can set a 50% sell tax, block exits, or mint a 5,001st ship. These are constants in the bytecode, not parameters.