StockIQ
paper
--:-- PT —
…

Watchlist 10 names · live

Evaluating all ten…

Agent

ask · instruct

Decision log

every screen, entry and exit

Backtest

M2 · lower bound
Daily bars cannot say whether the high or the low came first, so any day touching both target and stop is scored a loss. Results are a pessimistic floor, never an estimate.
No backtest yet. 120 sessions takes a couple of minutes.

Standing instructions

—
LockedRisk limits and the flatten. Changed in the rulebook, applied on deploy.
ConfigThresholds and gates. Live in universe.yaml, applied at next screen.
StandingPlain-English rules. Editable here, applied at next screen.

Probability model

—
Weights

Values are log-odds, not percentages: +0.5 lifts a score of 26 to about 37, +1.0 to about 49. Rows marked fitted came from data and are locked — overriding a measured number with an opinion is how a model stops describing the market and starts describing you. Everything else is a documented zero you can set, and reset.