Spreads Get quoting
Strategy

The 17.5 hour edge window

Price discovery keeps banker's hours. The chain doesn't. The gap between those two clocks is where the desk earns.

The two clocks

SessionEasternWhat happens
Primary session9:30 – 16:00 Mon–FriReal price discovery. Arbitrage keeps tokens glued to stocks. Spreads are tight; there is little edge for an on-chain maker.
After hours16:00 – 20:00Earnings land. Thin primary venues still print. Token flow continues at full speed against decaying information.
Overnight20:00 – 4:00No primary prints at all. Futures and foreign listings are the only anchors. The pools are at their most stale, the flow at its most uninformed. Prime quoting hours.
Pre market4:00 – 9:30Information starts pricing in again. Overnight inventory should be flat before the open gap.
WeekendFri 16:00 – Mon 9:30Sixty-five hours of trading with zero primary discovery. The largest window and the largest risk: news has all weekend to accumulate against your quotes.

The desk API exposes this clock at /v1/session so an agent never has to guess which regime it is quoting into.

What the pools do at 4:01pm

Nothing. That is the entire thesis. A constant-product pool prices the pair from its reserve ratio and charges the same fee at noon and at midnight. It does not widen for uncertainty, does not read the 4:05 earnings release, and does not pull quotes on a halt. Its liquidity is only honest while arbitrageurs keep its reserves aligned with the primary market — and after the close, they have nothing to align it with.

What an agent does at 4:01pm

Why the edge persists

This is not a secret; it is a shift nobody wants. Human market makers do not staff 3am on a Sunday for a few basis points on tokenized AAPL. Passive pools cannot think. An agent has no circadian rhythm, no payroll, and no boredom. The hours that are a cost center for a trading firm are the product here.