How it works
Methodology
Deterministic research pipeline using public market data. No private trading APIs. No order execution.
Data & model flow
- Public OKX Data
- Technical Indicators
- Strategy / Signal Engine
- Prediction
- Historical Evaluation
- Statistical Analysis
Market data
- • Source: public OKX candle endpoints only
- • Assets: BTC-USDT, ETH-USDT, SOL-USDT, XRP-USDT
- • Timeframes: 15m, 1H, 4H
- • Predictions use the latest closed candle when possible
Indicators
- • EMA 20 / 50 / 200
- • RSI 14
- • MACD (line, signal, histogram)
- • ATR 14
- • Volume & volume SMA
Strategies researched
- • Baseline trend + momentum
- • Breakout
- • Trend following
- • Momentum
- • Mean reversion
- • Live prediction engine leans on breakout + indicator agreement
ML & statistics
- • Phase 10: ML signal filter (no durable edge)
- • Monte Carlo trade resampling
- • Bootstrap intervals
- • Random benchmark comparison
- • Cost / slippage sensitivity
- • Walk-forward / robust validation in earlier phases
Prediction outputs
- • Direction: LONG / SHORT / NEUTRAL
- • Signal strength: 0–100 analytical score (not probability)
- • Model levels: entry zone, stop, take-profit, R:R
- • Deterministic explanation factors
Historical evaluation
- • WIN — take-profit before stop-loss
- • LOSS — stop-loss first (including same-candle SL-first)
- • TIMEOUT — neither level within horizon
- • SKIPPED — NEUTRAL / missing levels
- • Causal: predict on candles ≤ N; evaluate on N+1…
What QuantLab is not
- • Not a trading bot or exchange
- • Not an investment advisory service
- • Not a claim of profitable foresight
NO ROBUST EDGE DETECTED
See research timeline →