StratEvo
Quantitative trading: quotes, backtesting, screening, and genetic algorithm strategy evolution.
Installation
Ajoutez ceci à la configuration MCP de votre client (Claude Desktop : claude_desktop_config.json ; Cursor : .cursor/mcp.json ; VS Code et les autres utilisent la même forme « mcpServers »).
stratevoclaude mcp add stratevo -- uvx stratevo==6.0.3{
"mcpServers": {
"stratevo": {
"command": "uvx",
"args": [
"stratevo==6.0.3"
]
}
}
}Documentation
Tirée du README du projet sur GitHub, actualisée chaque semaine. Les liens ouvrent GitHub.
Sommaire
🦀 StratEvo
Stop writing trading strategies. Evolve them.
A genetic algorithm engine that breeds and walk-forward validates trading strategies across 484+ market factors.
Live Signals · Paper Trading · How It Works · Results · Robustness · Get Access
📡 Live Signals
Real-time buy/sell signals from evolved strategies. Updated daily. All signals are committed to git history — you can verify every one.
Latest Signals
| Date | Market | Action | Asset | Entry Price | DNA | Status |
|---|---|---|---|---|---|---|
| Signals will be posted here as Paper Trading goes live |
📁 Full signal history: signals/
📊 Paper Trading Performance
Forward-testing evolved strategies on real market data with simulated execution. No hindsight, no cherry-picking.
Paper Trading active — Crypto V13 live since 2026-04-18.
Current Paper Portfolio
| Strategy | Market | Start Date | Days | Return | Sharpe | MaxDD | Trades | Status |
|---|---|---|---|---|---|---|---|---|
| Crypto V13 | Crypto | 2026-04-18 | 0 | — | — | — | — | 🟢 Live |
📁 Daily P&L reports: paper-trading/
📈 Equity curves: paper-trading/charts/
Equity Curve (demo — real data accumulating)
Drawdown
How It Works
Most quant tools make you write the strategy. StratEvo evolves them instead.
You write the rules → StratEvo discovers the rules
You tune parameters → GA tunes parameters
You test on one period → Walk-forward tests on multiple windows
You hope it generalizes → Monte Carlo measures if it does Random DNA population (484 factor weights + risk parameters)
│
▼
┌──────────────────────┐
│ Walk-Forward Test │ Multi-window out-of-sample validation
│ each DNA candidate │ Real fees, slippage, position caps
└──────────┬───────────┘
│
▼
Keep the survivors (fitness = Sharpe × Return / MaxDD)
│
▼
Mutate + Crossover → next generation
│
▼
Repeat for N generationsEach DNA is a weight vector across 484+ factors plus risk/position parameters — all evolvable:
| Parameter | Range | What it controls |
|---|---|---|
| Factor weights (×484) | 0.0–1.0 | Which factors matter and how much |
hold_days | 2–60 | Day trades through swing trades |
trailing_stop | % | Trail below peak to lock in profits |
market_regime | sensitivity | Reduce exposure automatically in bear markets |
kelly_fraction | 0–1 | Position sizing from recent win rate |
Evolution Results
Numbers from our running evolution engines. Updated as generations progress.
🇺🇸 US Stocks V8 (100 S&P 500 stocks — Gen 136)
| Metric | Best DNA |
|---|---|
| Annual Return | 33.5% |
| Sharpe Ratio | 1.47 |
| Max Drawdown | 17.0% |
| Win Rate | 55.5% |
| Profit Factor | 1.75 |
| Total Trades | 179 |
₿ Crypto V13 (17 assets — Gen 53)
| Metric | Best DNA |
|---|---|
| Annual Return | 69.0% |
| Sharpe Ratio | 2.27 |
| Max Drawdown | 13.0% |
| Win Rate | 50.0% |
| Profit Factor | 1.58 |
| Total Trades | 174 |
These are backtests with walk-forward validation, not live trades. That's the whole point of paper trading — proving it works forward, not just backward.
Anti-Overfitting
We learned this the hard way. An early version showed 25,000% returns. Turned out to be a bug — look-ahead bias.
| Defense | What it does |
|---|---|
| Walk-Forward | Multi-window OOS validation. Must profit on data it never trained on. |
| Monte Carlo | 1,000 shuffled iterations. p-value < 0.05 or it's luck. |
| CPCV | Combinatorial Purged Cross-Validation. Industry standard for a reason. |
| Arena Mode | Multiple strategies compete head-to-head. Crowded signals get penalized. |
| Bias Detection | Look-ahead, snooping, survivorship — flagged automatically. |
| Turnover Penalty | Excessive trading is punished. Real transaction costs baked in. |
An honest 33% beats a fake 25,000%.
484+ Factors
| Category | Count | Examples |
|---|---|---|
| Crypto-Native | 200 | Funding rate, whale detection, liquidation cascade |
| Momentum | 14 | ROC, acceleration, trend strength |
| Volume & Flow | 13 | OBV, smart money, Wyckoff VSA |
| Volatility | 13 | ATR, Bollinger squeeze, vol-of-vol |
| Mean Reversion | 12 | Z-score, Keltner channel position |
| Trend Following | 14 | ADX, EMA golden cross, MA fan |
| Qlib Alpha158 | 11 | Microsoft Qlib compatible factors |
| + 5 more categories | 37 | Risk, quality, price structure, sentiment, DRL |
All factor weights are discovered by evolution. Zero manual tuning.
Strategy Styles
The algorithm converges on recognizable trading styles on its own:
| Style | What the DNA learned |
|---|---|
| Value Seeker | Buys cheap, holds patient |
| Momentum Rider | Chases runners, dumps laggards |
| Mean Reverter | Bets on bounce-backs |
| Flow Reader | Follows the money — volume leads price |
| Volatility Hunter | Profits from vol expansion |
| Crypto Native | 200 factors built for 24/7 markets |
Get Access
StratEvo Pro includes the evolution engine, paper trading, signal generation, and live exchange connectors.
📧 Contact: neuzhou@outlook.com
💬 Discord: discord.gg/kAQD7Cj8
Technical Papers
Check back daily for updated signals and paper trading results.
Questions fréquentes
Qu'est-ce que StratEvo ?
Quantitative trading: quotes, backtesting, screening, and genetic algorithm strategy evolution.
Comment installer StratEvo ?
Il s'exécute sur votre machine. Dans Claude Code : claude mcp add stratevo -- uvx stratevo==6.0.3. Pour les autres clients, collez la configuration JSON de cette page dans leurs réglages MCP.
Faut-il une clé API pour StratEvo ?
Son entrée au registre n'indique aucune clé ni secret obligatoire. Certaines fonctions peuvent tout de même demander un compte auprès du service auquel il se connecte.
StratEvo est-il open source ?
Son code est public sur GitHub, mais le dépôt ne déclare pas de licence standard.
Qui maintient StratEvo ?
Il est publié par NeuZhou. AI.tn le référence sans le maintenir ni le recommander.
Issu du registre MCP officiel. Être listé n'est pas une recommandation : vérifiez le code et les permissions d'un serveur avant de lui donner accès à vos données.