Momentum Foresight
Bitcoin Elite Trend Alpha
Predict trend outperformance before it happens. This model forecasts the future slope of Bitcoin's price trajectory, identifying potential regime shifts and hidden momentum.
Trend Reliability
Up to 81% Precision
Distinguishing trend from noise
Multi-horizon forecasts (4H, 12H, 24H, 60H)
Up to 81% precision in trend filtering
Institutional taker/maker imbalance detection
No-repaint momentum signatures
Trend Performance Matrix
Validated performance across multiple predictive horizons.
| Horizon | Avg. Accuracy | Precision (Alpha) | Trader Utility |
|---|---|---|---|
| 60-Hour (Swing) | 67.2% | 0.63 | Swing Continuation |
| 12-Hour (Intra) | 65.2% | 0.76 | High Reliability |
| 4-Hour (Scalp) | 66.5% | 0.74 | Scalp Filter |
How it works
Every hour the model ingests fresh market structure, re-computes the Slope-derived trend, and delivers a probability that Bitcoin holds above or below the line over the next 60-hour horizon.
Four horizons, one decision
Each hour, four independent models score trend durability at 4h, 12h, 24h, and 60h lookouts. When they converge, conviction is high.
Probability, not prediction
The models don't guess price β they estimate the odds the current trend persists. You trade the probability, not the noise.
Confluence-based signals
Entry triggers only when short and long horizons align above threshold. Exits fire when any horizon weakens β protecting capital early.
Signal decoding
Probability β₯ 0.95 signals strong bullish continuation with high conviction.
Probability β€ 0.05 highlights exhaustion or an imminent reversal risk.
Intermediate readings flag indecision zones where the edge is limited.
Interpreting the target
Conceptually, the probability ring highlights how the model chases the 0.95 buy trigger, while the trend trajectory sketch mirrors what you saw in the BTC example: green while bases hold, red where the slope usually breaks, and the highlighted zone where continuation has been most reliable historically.
Backtest performance snapshot
Recent validation on 8,853 hourly candles delivered a 75.33% accuracy score with tightly clustered precision/recall near 0.75 for both continuation (class 0) and exhaustion (class 1) scenarios. The indicator preserves balance across bullish and bearish calls, which you can see in the confusion matrix and class-by-class breakdown. All metrics are derived from a walk-forward validation loop so the model only ever sees historical data when scoring each forward window.
Accuracy
Validated on 8,853 hourly candles.
Macro F1
Balanced harmonic mean across both classes.
Class 0 precision
Continuation class (trend holds).
Class 1 precision
Reversal class (trend breaks).
Bottom line: The strategy delivers +15.79% alpha over buy-and-hold with a Sharpe Ratio of 2.14 and maximum drawdown of just 7.57%. With 185 trades achieving a 68.1% win rate, this represents a statistically validated edge confirmed by Monte Carlo, Bootstrap, Walk-Forward, Permutation, and T-test analyses.
Statistical Tests
- p < 0.001βMonte Carlo Simulation
- 95% CI > 0βBootstrap Analysis
- 5/5 profitableβWalk-Forward Validation
- p < 0.001βPermutation Test
- p = 0.0045βT-test vs Buy & Hold
π’ Entry Conditions
- P(4h) > 0.8 β short-term momentum confirmed
- P(12h) > 0.8 β medium-term trend aligned
- P(24h) > 0.7 β daily bias supportive
- P(60h) > 0.6 β multi-day trend intact
π΄ Exit Conditions
- P(4h) < 0.3 β short-term momentum fading
- OR P(12h) < 0.4 β medium-term trend weakening
Run your own backtest
Want to validate the strategy yourself or experiment with different parameters? Use this Python script with <a href='https://vectorbt.dev/' target='_blank' class='text-marbell-orange hover:underline'>vectorbt</a> to fetch live data from our API and run your own backtests. The script includes parameter optimization, detailed trade analysis, and interactive Plotly charts.
Requirements
- Python 3.9+ with pip
- Install dependencies:
pip install vectorbt pandas requests python-dotenv plotly - Create a
.envfile with your API key:API_KEY=your_key_here
What the script does
- Fetches probability data for 4h, 12h, 24h, and 60h horizons
- Runs vectorbt backtest with configurable entry/exit thresholds
- Generates interactive HTML charts with equity curve and drawdown
- Includes parameter optimization to find the best settings
What you see on the chart
- Gold line β AI-calculated hourly trend fit (trend_1h_fit).
- Lime markers β high-confidence buy triggers when probability β₯ 95%.
- Red markers β high-confidence sell alerts when probability β€ 5%.
- Gray markers β neutral segments where the model stays cautious.
- Optional aqua line β raw probability stream for quantitative overlays.
Why it matters
Traditional indicators react to priceβthis one scores the staying power of the move. With a 75.3% historical accuracy window, you know when the market is behaving like past stretches that kept trending versus segments that usually rolled over.
- Quantifies how likely the current trend is to persist over the next 60 hours instead of guessing exact prices.
- Detect continuation before follow-through is obvious on the chart, with a 75.3% validation accuracy backdrop.
- Spot false breakouts and trend fades before they trap capital.
- Commit capital only when model odds align with your setup.
Built for serious traders
- AI-driven probability engine trained on institutional flow data.
- High-conviction trend identification for swing trades.
- Secure API integration for automated execution or charts.
- Removes emotional bias with objective statistical scoring.
How to use it
- 1Monitor the 60H probability for major regime alignment.
- 2Use 4H and 12H horizons for tactical entry/exit precision.
- 3Scale positions based on multi-horizon signal consensus.
Deploy inside TrendSpider
Paste the script into a TrendSpider custom JavaScript indicator, replace the placeholder API key with your credential, and save. The overlay streams the AI trend fit plus probability-driven markers in real time.
- Open TrendSpider, create a new JavaScript indicator, and clear the default template.
- Paste the snippet, swap
YOUR_API_KEY_HEREfor your key, then click Save. - Add the indicator to your chart in overlay mode to see the trend fit and markers.
Turn complex market structure into clarity
BTC 1H Trend (AI) keeps you aligned with dominant momentum and out of low-conviction chop. Activate it today and pair it with your existing playbook for higher conviction entries.
π Get Started in 3 Minutes
Follow these simple steps to start receiving AI-powered trend signals for your Bitcoin trading.
Create Free Account
Sign up with Google or TikTok in one click. No credit card required to explore.
- βInstant access to dashboard
- βView sample signals
- βExplore all indicators
Subscribe to Indicator
Choose your plan and get instant access. Secure payment via Stripe.
- βCHF 49/month
- βCancel anytime β’ No hidden fees
Start Trading Smarter
Access live signals via dashboard, API, or TrendSpider integration.
- βReal-time dashboard
- βREST API for automation
- βTrendSpider charts
π Secure checkout powered by Stripe β’ π³ All major cards accepted
185
Backtested Trades
68.1%
Win Rate
2.14
Sharpe Ratio
5/5
Statistical Tests Passed
Disclaimer
The Signals provided by Marbell AG are for informational and educational purposes only and are not investment advice or a recommendation to buy or sell any financial instrument. The Signals are non-personalised and do not consider your objectives, financial situation, or risk tolerance. Past or simulated performance is not indicative of future results. Marbell AG does not execute trades or provide portfolio management or copy-trading services.