Climate, then trigger, then risk.
The Futures engine is a pipeline, not a crystal ball. Regime is a filter. FLAT is microstructure context. Styles are independent experiments. If that sounds slower than a Telegram long — good.
FLAT microstructure
Crowding tax. Persistent extremes change expected value; they do not print a market order by themselves.
Positioning shocks. Build vs. cover, squeeze vs. unwind. Location of liquidity matters more than a pretty wick.
Tape / impulse quality. Who is lifting offers vs. fading. Used as confirmation, not as a lone trigger.
Session and time-of-day. London vs. New York vs. dead hours. Some styles are simply closed overnight.
Decision pipeline
Classify 15m / 1h / 4h / 1d as trend, range, or chop. A scalp is illegal in the wrong 15m climate even if the 1d is trending.
Where is price relative to value, range extremes, and prior day. Middle of nowhere is not a setup.
Style-specific confirms (break, reclaim, VWAP, impulse, …). One style, one ticket. No stacking.
Invalidation as a price. Size in R. Macro blackout can veto even a perfect trigger.
Independent styles
Short-horizon continuation and fade. Tight invalidation. No stacking with slower books.
Session structure, VWAP interaction, and funding-aware continuation. Closed before overnight risk when the profile says so.
Multi-session trends with 4h regime confirmation. Designed to ignore 15m noise.
Daily regime and higher-timeframe location. Fewer trades, larger R targets, strict macro gates.
Prints around scheduled macro. Either stands down into the print or trades the defined post-print reclaim — never both blindly.
Run the live engine
Regime, setup queue, and FLAT histogram update inside the app. This page is the textbook; the desk is the lab.