Everything in this document describes exactly what runs in production — no simplified marketing version. If you can read code, the logic here maps directly to what we've deployed.
OHLCV (Open, High, Low, Close, Volume) candle data for each supported asset is pulled from market data APIs and stored on a rolling basis. The backend retains approximately one year of daily candles per asset.
All 21 models are run against the most recent candle data. Each model independently outputs a directional signal: BUY, HOLD, or SELL, along with a confidence score between 0 and 1.
Each prediction is timestamped and stored at generation time. When the next candle closes, the prediction is resolved against the real market direction. A prediction is marked correct if the asset moved in the predicted direction by close.
Directional accuracy (dir_acc) is computed as the fraction of correct directional predictions over a rolling window. A model with 60% directional accuracy has correctly predicted price direction in 60 out of 100 recent predictions.
A weight is assigned to each model using the formula below. Rankings update daily. The top-weighted models per asset are promoted — their signals get higher visibility in the platform.
Model weight is computed as:
This formula has two important properties:
Each model is fully independent — they share the same input data but use entirely different approaches to reach their predictions. No model influences another's output.
| Model | Approach |
|---|---|
| XGBoost | Gradient-boosted decision trees trained on engineered price and volume features. Well-suited for tabular OHLCV data with non-linear relationships. |
| Gradient Boosting | Ensemble of weak learners (sklearn implementation) that iteratively corrects prediction errors. Captures complex feature interactions across time windows. |
| Random Forest | Ensemble of randomized decision trees. Robust to overfitting via bagging; provides implicit feature importance ranking across technical indicators. |
| Neural Network | Feedforward multi-layer perceptron trained on normalized OHLCV series. Learns non-linear temporal patterns without hand-crafted feature rules. |
| Support Vector Machine | SVM classifier with RBF kernel. Finds the maximum-margin hyperplane separating directional outcomes in feature space. |
| Model | Approach |
|---|---|
| Ichimoku Cloud | Multi-component Japanese charting system using Tenkan-sen, Kijun-sen, Senkou Span A/B, and Chikou Span for trend, momentum, and support/resistance signals. |
| SuperTrend | ATR-based trend-following indicator. Generates a clear directional line that switches side when price closes through the trailing stop band. |
| Parabolic SAR | Stop-and-Reverse indicator tracking price acceleration. Produces tight signals in trending markets; checked against trend confirmation to reduce whipsaws. |
| EMA Divergence | Measures spread between fast and slow exponential moving averages. Divergence direction and magnitude inform momentum and potential reversal signals. |
| TRIX Signal | Triple exponential smoothing of price. The rate of change of a triple EMA reduces noise and generates momentum signals with minimal lag. |
| Williams %R | Momentum oscillator measuring closing price relative to the high-low range. Identifies overbought (> -20) and oversold (< -80) conditions. |
| Momentum | Raw rate-of-change of price over a set period. Directional momentum is combined with zero-line crossovers to classify trend strength. |
| Price Action | Candlestick pattern recognition — engulfing candles, pin bars, inside bars, and doji formations — to identify institutional pressure and reversals. |
| Technical Indicator Composite | A weighted composite of RSI, MACD, Bollinger Bands, and Stochastic oscillator. Signals are only generated when multiple indicators align. |
| Model | Approach |
|---|---|
| VWAP Premium | Measures price deviation from Volume-Weighted Average Price. Extended premium above VWAP signals overextension; discount signals potential accumulation. |
| Volume Profile | Analyzes trading volume distribution across price levels. High-volume nodes act as support/resistance; breakouts through them carry higher conviction. |
| Volatility Breakout | Identifies expansion in price range (ATR-based) following periods of compression. Breakouts after low-volatility consolidation are classified as directional signals. |
| Mean Reversion | Statistical model that identifies assets trading at significant Z-score deviations from their rolling mean. Expects reversion toward the mean over a lookback window. |
| Fear & Greed Index | Incorporates market-wide sentiment data — volatility, momentum, social signals, dominance — to classify broad market psychology and its likely directional impact. |
| Elder-Ray Index | Combines exponential moving average with bull/bear power measurements. Positive bull power above EMA confirms uptrend signal; negative bear power below EMA confirms downtrend. |
In addition to our 21 AI models, ArtinFox monitors a curated set of 50+ external signal sources — crypto influencers, financial media outlets, and market analysts — across multiple platforms including X (Twitter), YouTube, RSS feeds, and newsletters.
Each external call is logged with a timestamp and directional prediction. When the market resolves, the call is scored using the same directional accuracy methodology as our AI models. Sources are ranked publicly on the News page.
All signals, accuracy figures, and model rankings provided by ArtinFox are for educational and informational purposes only. They do not constitute financial, investment, or trading advice. Past model accuracy is not indicative of future performance. Cryptocurrency markets are highly volatile. Always conduct your own research and consult a qualified financial advisor before making any investment decision. See our full Disclaimer.