Concepts
Signal Interpretation
How to interpret daily model weights safely and consistently in the Stacking Sats workflow.
Signal interpretation in Stacking Sats is about relative context, not point prediction. A higher signal means the model views current conditions as more favorable than the baseline, within its historical framework.
Getting signals in the app: Complete the Quick Start path (profile → plan → API token); your AI agent then reads the signal via the Agent API.
What A Daily Signal Means
- Signals are model outputs, not investment advice.
- Signal values are relative to the model's training and backtest assumptions.
- A single day should be interpreted as one data point in a sequence, not as a standalone forecast.
Practical Interpretation Bands
| Signal Weight | Practical Interpretation |
|---|---|
| < 0.90 | More defensive vs uniform baseline |
| 0.90-1.10 | Near baseline allocation behavior |
| > 1.10 | More aggressive vs uniform baseline |
Use these bands as operational shorthand. For full assumptions and constraints, see Assumptions and Limitations.
Worked Example
Assume your uniform daily allocation is $10.00 and today's dynamic signal weight is 1.35.
- Dynamic allocation =
10.00 * 1.35 = $13.50 - Relative change vs uniform =
+35%
If tomorrow's signal is 0.85, the same uniform baseline becomes $8.50. Interpretation should focus on consistency over time, not any single-day value.
Risk Controls For Interpretation
- Anchor every interpretation to Assumptions and Limitations.
- Validate execution logic against Weight Computation and Signal Composition.
- Treat backtest outputs as historical validation only (Backtest Overview).