Bootstrapped LP

Why this mechanism

A venue where traders bet against a pool needs the pool to pay winners. Usually someone must fund it first and be paid to take that risk; bets are then capped by what the pool can cover, and a bad run, or a bad actor, can empty it. Pools of earlier on-chain games of chance were seeded this way, with central capital or with token rewards that complicated everything else, and users of such venues still worry whether the house can pay.

The exchange removes both constraints, the capital up front and the cap on bet size, by replacing instant payouts with a payout queue.

The payout queue

Instead of an LP that pays every winner on demand, wins the LP cannot cover wait in a first-in, first-out queue. The protocol may be insolvent in aggregate for a while, but winners are paid, in order, as soon as the money is there:

1. A position closes in profit.
2. If claims are already queued, the new win joins the back of the queue, and a harvest pays from the front whenever there is cash.
3. If nothing is queued, the LP pays the win from its cash; whatever it cannot pay starts the queue.

The queue absorbs swings that an empty LP could not. One harvest can pay several claims at the front when enough cash has come in. In our accounting, margin lost on positions funded from a free balance becomes LP cash (minus the loss fee, and minus any builder fee) while the queue is empty; while claims are queued it is set aside in the side bucket and harvested into queued payouts instead (Solvency & queue).

Why token emissions help

A game with an edge for the house can still be attractive to players if their losses earn a claim on the protocol’s future value. The traders who carry the most risk, the earliest ones, should receive the most.

Here that claim is TOKEN, minted on the mint curve: at launch a flat 100 TOKEN per dollar of mint basis while the tracked LP is below $2,000,000, including while it is underwater, then a smooth decay. The flat rate is what makes trading worthwhile before the LP is solvent. After the switch to the supply-share rule, each dollar mints a fixed share of the supply, and the part of a loss that refills an underwater LP counts up to 5×: the same incentive, aimed at the moments the LP needs it most.

The paper

The scheme follows a payout-queue paper published in 2024 by a pseudonymous author under the name “Martingaler”, on bootstrapping asynchronous games of chance without an initial bankroll. We cite it by name only; this page describes our implementation, which differs from the paper in four ways:

• Emissions follow the LP, not the number of plays. Losers receive tokens at a flat rate while the tracked LP is small and at a decaying rate after that, rather than on a schedule that falls with every play.
• No escape valve. Queued winners cannot give up their winnings to take back their stake.
• No yield on the LP. The LP holds plain USDC.
• Earmarking. While claims are queued, every loss is set aside in a side bucket for the queue.