Performance without cherry-picking.
The goal is not to publish a glossy equity curve. It is to record what the engine believed before the move, whether the market confirmed it, which target levels were available and what happened next.
Freeze the logic
A research rule should be defined before its forward results are counted. Changing the rule after every losing day destroys the test.
Log the state
Record route, strength, grade, confirmation, target map and timestamps so later review cannot quietly rewrite the story.
Report both sides
Show wins, failures, conflicts, no-trade sessions and missing-data sessions. A research platform should make weak spots visible.
Forward collection is live.
The live feed is now collecting 1-minute NQ data into the route database and the dashboard is designed to refuse stale or incomplete context instead of fabricating a clean session.
Because the 2026 historical gap is not fully backfilled and the current live sample is still young, the community should expect transparent forward tracking rather than big performance claims.
What we plan to publish.
Route outcomes
How often the selected route matched the subsequent session behavior, segmented by grade and context.
Target reach
Which actual market levels were reached after confirmation, rather than only fixed-point extensions.
Failure modes
Conflict, late switches, incomplete sessions, stale context and cases where the engine should have stayed neutral.