TraderAI
AI Chart AnalysisLive ChartDashboard
AI AssistantTrade IdeasScreenerWatchlistAlertsMarket HeatmapVolatility RadarCurrency StrengthPattern StatsEconomic Calendar
Earn Rewards
EARN
Pricing
PRO
Dashboard
Upgrade to Pro
Sign InCreate Account
Volatility Radar

Forecast Accuracy — Public Track Record

How the Volatility Forecast actually performed: for each of the last 60+ trading days we compare the expected move (forecast from prior data only — no hindsight) with the move that really happened. Judge us by the numbers below, recomputed live from public exchange history.

How to read this honestly

A PERFECT ±1σ forecast band contains about 68% of days and a ±1.96σ band about 95% — never 100%. Hit rates close to those targets mean the model is well calibrated. Far above the target means the bands are wider than needed; far below means risk was under-called. 100% would be a red flag, not a feature.

Re-scoring the last 60 days of forecasts…

Methodology (reproducible by anyone)

For every day in the window we rebuild the exact GARCH(1,1) forecast the product would have shown that morning — using only candles that existed at the time — then compare it with the day's realised move. Nothing is stored or curated: this page recomputes the whole track record from public exchange history on every load, so a skeptic with the same candles gets the same numbers. Calibration = average realised move ÷ average forecast move (≈1 is ideal).

Volatility forecasts estimate the size of likely moves, never the direction. Educational only, not financial advice.