Signal Grading Model (SGM)
A machine-learning score, measured against reality before it's trusted.
The Signal Grading Model (SGM) is a machine-learning layer that scores each signal on top of the rule-based engine. Crucially, it runs in measurement mode: its scores are recorded and evaluated against real outcomes before they are ever allowed to influence a live decision.
This is the disciplined way to introduce ML into trading — prove it on out-of-sample results first, then promote it only if it demonstrably adds edge. No black-box score is trusted on faith.
ML scoring layer
A learned score that sits on top of the transparent rule-based signals.
Measurement mode
Scores are logged and benchmarked against real outcomes before going live.
Outcome-anchored
Every prediction is checked against what actually happened, not a backtest fantasy.
Promotion by proof
A model earns influence only after it demonstrates edge on unseen data.
SGM is how the terminal improves over time without becoming a mystery box — it complements the rules rather than replacing the logic you can inspect.
What is measurement mode?
The model scores signals but those scores are only recorded and evaluated — they don't drive live decisions until proven.
Is this a black-box AI score?
No. It sits on top of transparent rule-based signals and must prove edge on real outcomes before it's trusted.
When does SGM go live?
Only after out-of-sample results justify it. Until then it runs purely as a measured, benchmarked layer.
Get early founder access
The TradrQuant terminal is opening early to founders. Join the waitlist and get in as modules like Signal Grading Model (SGM) roll out.
Coming soon · no card · one email at launch
TradrQuant is an educational and analytical platform. Nothing here is financial advice. Trading involves substantial risk and past performance does not guarantee future results.