Confidence is a trust score that tells you how reliable the AI believes its forecast range is at the moment. It is not a promise and it is not a “probability of going up.” It simply answers : " How much can I trust the projected price corridor for this horizon, given current market conditions ? " The numbers Q10, Target (Mid), and Q90 describe the forecast itself. Q10 is the lower bound of the expected range, the Target is the central estimate, and Q90 is the upper bound. Together they form a probabilistic corridor where price is expected to evolve over the selected time window.

Confidence does not move Q10 or Q90 up or down directly; it tells you how dependable that corridor is right now.You will see a confidence per horizon (1h, 2h, 4h, 6h) because predictability changes with time. Short horizons often benefit from clearer microstructure signals, while longer horizons face more uncertainty.

The global confidence is a single headline score for the whole snapshot: it starts from the model’s base reliability and then applies environmental adjustments when the market context is harder to model, such as unusual or unfamiliar structure, noisier signals, or atypical volatility/activity. In practice, a higher confidence means the range is more reliable for planning scenarios around supports/resistances, while a lower confidence means you should treat the range as a rough guide and rely more on confirmation (breakout/breakdown) before acting.

Below, detailed information on how this confidence score is computed.