The Gas Floor

HEX trades in dozens of pools at once.  What holds them to one price is arbitrage, and arbitrage runs on gas — so the cost of a swap decides whether the HEX price you see in any pool is a price or a leftover.  An on-chain measurement, read nightly since 2026-08-23; the figures below are from the last read.

The Question

Dozens of pools, different partners, different DEXs, different depths — how can there be one price?  The textbook answer is arbitrage: any pool that drifts from the others is a free trade, someone takes it, and the drift closes.  The answer has a condition, and the condition is measurable.

The Measurement

Every night the sweep reads every HEX and eHEX pool it can find on both chains — 52 pools at the last read, 2026-09-18 — and compares prices within partner families, every eHEX/WPLS pool against its siblings, so no dollar conversion touches the measure.  Each pool’s drift is read against the deepest pool of its family, the yardstick, never counted as evidence of itself.  Pools are tiered by the dollars on their token side: deep from ^$10,000, small from ^$1,000, dust below.

The deep pools quote one price.  On PulseChain, 3 of the 4 deep pools sit within 1% of their family’s deepest pool, and 2 of the 4 small ones; every exception is on switchx.  On Ethereum, 1 of the 2 deep pools sits within 1%, and none of the 8 dust pools do.  The count changes where the depth sits, not what it quotes.

The Tail — and the Law

The pools that disagree come in two kinds.  On Ethereum the dust pools sit as far as 7,328% from their family, quotes nobody has touched in months.  That is the law this page is named for: arbitrage corrects a pool only while the correction pays more than the gas and fees it costs.  That threshold is the gas floor.  On PulseChain a swap costs a fraction of a cent, so the floor is the swap fee, about 0.3% a leg, and pools of a few hundred dollars stay inside it.  On Ethereum a swap costs cents to a dollar, and a pool whose drift is worth less than that stays where it is — the smaller the pool, the wider the band it wanders in before anyone bothers.  A chain’s price web is only as tight as its gas is low.

On PulseChain no pool of any size sits more than 5% off tonight.  The three switchx pools the sweep read off the market from 2026-08-23 to 2026-10-19 are concentrated-liquidity pools; a switchx pool’s balances are not its price, and read from the pool itself each sits within 1% of PulseX.

Where eHEX’s Liquidity Lives

About half of all pooled eHEX sits on PulseChain as the bridged twin, half on Ethereum, the chain that issued it (the 2026-10-19 read).  Where the price is made would take a lead–lag measurement across the bridge, which this sweep does not make.  Only about 1–1.5% of each token’s liquid supply works in pools at all (see % of Liquid in LPs); the rest is held or staked — staking alone holds over 90% of everything allocated, which is the design working: HEX is built to be held.

The Straggler Board

Every pool the nightly sweep reads that sits more than 1% from the deepest pool of its family.  Coverage is not total: the PulseChain discovery source lists up to 30 pools per token, so a straggler ranked past that cap waits until churn surfaces it; the Ethereum side is read from the Uniswap factories.  The estimate column is the constant-product arbitrage profit, the pool’s token side in dollars × (√1+drift − 1)², before gas and fees — deliberately, since gas and fees are why most of these pools are here at all.  Nothing here is advice; the board states what the chain says, and people will do what they do.

*Exact for v2 pools, before gas and fees; v3 rows can overstate, and no row reads more than the pool holds of the token the trade takes out.  Method and full per-family data: pool-dispersion.json.

Why It Matters Here

For HTT markets the lesson is narrower than it looks.  HTT pools live on PulseChain, where the gas floor barely exists, so where a maturity has more than one pool, cheap gas keeps them agreeing.  Most maturities have one pool, and a lone pool’s quote is its last trader’s opinion — its discount to HEX is a judgment about the maturity, not an arbitrage anyone can close.  The distinction the Manual’s liquidity chapter draws between a pool’s depth and whether its quote is the market is the distinction this measurement makes visible.  Definitions: Gas Floor.

Epilogue

We assumed that arbitrage is the only force holding a token’s pools to one price, that gas is the only cost it has to clear, that every pool’s balances are its price, that a family’s average was a fair yardstick, and that a number typed on 2026-08-23 would still be true in September.

What held.  On PulseChain every deep pool outside switchx has agreed on every read since 2026-09-18, and the Ethereum dust follows the gas-and-fee floor.  On Ethereum the two deep eHEX/USDC pools did not: the Uniswap v3 0.3% pool sat 1.1% to 8.4% from the v2 pool on most reads from 2026-09-18 to 2026-10-19.

What was wrong, and what we did about it:

• Gas is not the only cost.  Swap fees set the floor wherever gas is nearly free, so the law now reads gas and fees.
• Not every pool prices by its balances.  Reading the switchx pools’ balances as a price put three sound pools on the board; the sweep now reads their price from the pool itself, and the projection below that they would stay on the board rested on that misread.
• An average is not a fair yardstick.  One that includes a broken pool bends the reading of the sound pool beside it, so every drift is now read against the deepest pool of its family.
• Typed numbers age.  The pool counts, the gas cost and one drift figure all went stale, so every figure above now comes from the nightly file.

The next month, to 2026-10-18, we project:

• The deep pools on both chains stay within 1% of their family’s deepest pool on every read.
• The Ethereum dust pools do not move.
• The switchx pools remain on the board unless the DEX itself changes.
• The board holds between ten and twenty rows.

If any of those fails, the law as written here is wrong somewhere, and this section will say where.