What Is Glass-Box AI Trading?
Most "AI trading signal" services are black boxes: you get a BUY or SELL with no reasoning. Glass-box AI is the opposite — every verdict comes with the models that produced it, ranked by influence.
The problem with black-box signals
When a signal is wrong, you have no way to learn from it. When it's right, you can't tell whether it was skill or luck. Black-box services make you dependent on their output without building your edge.
How Tidava's glass-box approach works
Tidava runs 30+ quantitative engines per asset simultaneously:
- Momentum engines — RSI, MACD, rate-of-change across 1m to 1D timeframes
- Regime detection — classifies market state (trend, range, breakout, reversal)
- Volatility flow — ATR, Bollinger, implied-vol surface
- Options positioning — put/call ratio, gamma exposure, max pain
- Macro overlay — live CPI, NFP, Fed policy, yield curve pressure
- Sentiment — funding rates, long/short ratios, social momentum
Each engine votes. Tidava surfaces the top 3 reasons that drove the verdict — ranked by conviction weight. You see which engines agreed, which dissented, and how strongly.
Why this matters for traders
When you understand why a signal fired, you can:
- Decide whether the reasoning matches your own thesis
- Size the position appropriately based on conviction
- Know exactly when the signal is invalidated
- Learn which engine combinations work in which regimes
Glass-box vs black-box: a direct comparison
| Feature | Black-box signal | Tidava glass-box |
|---|---|---|
| Shows reasoning | ✗ | ✓ Top 3 engines |
| Auditable when wrong | ✗ | ✓ |
| Builds trader edge | ✗ | ✓ |
| Regime-aware | Rarely | ✓ Always |
| Macro integrated | ✗ | ✓ Live data |