All fields editable — applies to the next signal cycle, resets on restart
Risk gate
Fusion weights — index
Fusion weights — single stock
Strategy matrix
Every cycle, the best-fit selector scores each enabled strategy applicable to the direction read by risk-return (expected profit ÷ max loss, where expected profit already accounts for how likely that category is to pay off given current conviction) and picks the highest score. Uncheck a strategy to remove it from consideration.
Strategy
Category
Direction
Stop
Profit target
07
Session
Data sources
Connections
Emergency control
Cancels every pending order, flattens every open position at market, and suspends auto-trigger system-wide until re-armed.
08
Run Log
Every scheduled run, success or failure
Started
Type
Status
Detail
09
Signal Scan
Every instrument, every cycle — not just the ones that surfaced a card
Instrument
Segment
Direction
Conviction
Expected move
TA votes
Outcome
Macro drivers
News drivers
10
Learning
Calibration by event type — last 7 days
Direction bias / magnitude scale are the corrections learned from resolved outcomes, shrunk toward 0 / 1 until enough samples accumulate. See manthan/learning/calibration.py.
Event type
Samples
Hit rate
Predicted dir. (avg)
Actual dir. (avg)
Direction bias
Magnitude scale
Reaction trend — average move since detection
Every tracked event, checked at fixed horizons after detection (not just the final resolution) — how the average reaction builds or fades over the following minutes.
Horizon
Samples
Mean move
Recently resolved events
Resolved
Instrument
Event type
Headline
Predicted dir.
Actual move
⏻ Halt everything?
This cancels all pending orders, flattens all open positions at market, and suspends auto-trigger until you re-arm the system.