# Xenos reproducible baseline evidence

Experiment: `xenos-btcusdt-m15-ema20-50-2026h1-v1`

- Symbol: BTCUSDT
- Timeframe: M15
- Dataset: 2026-01-01T00:00:00Z through 2026-06-30T23:59:59.999000Z
- Source: Binance public Spot market-data REST API (`https://data-api.binance.vision/api/v3/klines`)
- Candles: 17,376
- Dataset SHA-256: `d95f21741eb23a58ce60d274c97b4cf94253bec7fa89bd716f7d628bfff83a96`
- Strategy: Long-only EMA 20/50 crossover baseline
- Initial simulated capital: $10,000.00
- Completed trades: 164
- Total return: -41.40%
- Win rate: 23.17%
- Profit factor: 0.57
- Maximum drawdown: 44.62%

## Reproduce

From the project root:

```powershell
python research\run_public_baseline.py
```

The script reuses `candles.csv` when it exists. Delete that file only if you intentionally
want to download the dataset again from the recorded public source.

## Important limitation

This is a simulated historical research baseline, not live trading performance, financial
advice, or a promise of future returns. It uses next-candle-open execution, a fixed 0.10%
fee per side and fixed 0.02% slippage per side. It does not model order-book liquidity.