Unobservable entries
The signal is usable only after its complete one-second YM bucket, followed by explicit routing latency.
Does ES actually follow YM?
A frozen, costed, latency-aware test of the advertised 9:30–10:00 ET relationship.
Loading minutes-scale evidence…
Separate short-horizon test
This result tests seconds, not the advertised minutes-scale relationship.
Loading recorded evidence…
What this test protects against
Free trading
Every ES event includes adverse slippage and estimated round-trip commission.
Fake sample size
Signals are separated by a 30-second cooldown and results are summarized by independent day.
Parameter hindsight
Earlier sessions select the rule. Later sessions are untouched holdout with a shuffled-direction null test.