Learning progress

Trust Score tracks whether the model is simply collecting data, calibrating weights, validating consistency, or ready for more serious allocation review.

Model confidence

MACAS
Score
0
Foundation stage
0 evaluated runs

Phase tracker

1
Foundation
Foundation stage

Early phase: collect predictions and normalize the signal stack.

2
Calibration
Calibration stage

Weight learning begins once enough evaluated outcomes exist.

3
Validation
Validation stage

Walk-forward validation and stability checks are now meaningful.

4
Ready
Ready for allocation

The system has enough track record to support capital allocation review.

Trust score is based on 0 evaluated predictions. Minimum 200 evaluations are required for the ready state.