Canonical Pricing & Pool Dependence

Every price the engine consumes is derived from the canonical AMM pool and nowhere else.

The spot price is the pool's reserve ratio, read directly from the pair. For the current ETH-paired implementation:

P = ethReserve / tokenReserve

The formula generalises to any base-asset pair; nothing in the engine assumes ETH specifically beyond the deployment's configuration.

The same reserves feed the engine's structural anchor — the dynamic floor:

P_floor = ethReserve / (totalSupply − treasuryTokenBalance)

the price that would prevail if every token outside the Treasury sat in the pool simultaneously. The full anchor system built on P_floor is defined in State Variables & Price Anchors.

The design constraints are absolute: no external oracles, no time-weighted averages, no cross-pool aggregation, no off-chain data feeds. This trades sophistication for robustness — one explicit, auditable source of truth, no oracle-manipulation surface, no ambiguity about what the engine is responding to.

Two implementation details matter for anyone reading the contract's numbers:

The dead-pool guard. A pool whose base-asset reserve has been fully drained would make the floor zero and poison downstream math with division-by-zero and underflow. The floor computation guards this single chokepoint explicitly: at a zero reserve it reverts (DeadPool) rather than propagating a degenerate value.

The band is exposed on-chain. The Treasury publishes its own price-anchor band through public view functions — getFloor(), getPLiqStart(), and getPLiqMax() — computed on demand from live pool state and the immutable multipliers. Anything monitoring a launch (an interface, a keeper, an analyst) reads the band from the contract's own arithmetic rather than re-implementing the formulas, so a displayed anchor can never silently diverge from what the engine will actually do.