Accuracy Rate
accuracy_rate measures the percentage of an analyst's predictions that were directionally correct.
Definition
A prediction is directionally correct if the price moved in the direction of the call by the end of the measurement window:
- Bullish call → stock price is higher than at time of call → ✓ correct
- Bearish call → stock price is lower than at time of call → ✓ correct
The measurement window is 3 months by default.
Minimum prediction threshold
Accuracy is only shown when an analyst has 10 or more matured predictions. Below this threshold, accuracy_rate is null.
This prevents a single lucky call from showing a misleading 100% accuracy rate. An analyst with 3/3 correct is statistically indistinguishable from random; an analyst with 70/100 correct is not.
What accuracy does not measure
- Magnitude — an analyst can be 80% accurate with tiny returns, or 55% accurate with large returns. Always look at
overall_return_pctalongside accuracy. - Timing — a bullish call is correct if the price is higher at window maturity, even if it dropped 20% before recovering.
- Conviction size — all predictions are weighted equally regardless of the analyst's stated confidence.
Accuracy vs return — which matters more?
Neither alone is sufficient. The most useful signal is high accuracy + high return:
| Accuracy | Return | Interpretation |
|---|---|---|
| High | High | Strong analyst with good magnitude calls |
| High | Low | Correct direction but small moves |
| Low | High | High-risk, high-reward style; volatile |
| Low | Low | Poor signal overall |
Use --min-accuracy and --min-return together to find analysts in the top-left quadrant:
traderbro analyst list --min-accuracy 65 --min-return 10 --min-predictions 20 --json