TradingLab MACD + 200 EMA
publicmade in Claude Code
@djgambhirBTCUSDT · Binance USD-M perpetual · 15m · 2023-01-01 → 2026-09-01 · 128,545 bars in 123.0ms
Underperformed the market. Holding BTCUSDT returned +376.55% over the same period, against this strategy’s -81.28%.
The rules
A Pro feature. Pro isn’t on sale yet; beta testers get it free — ask on the pricing page.
BTCUSDT with every trade
A Pro feature. Pro isn’t on sale yet; beta testers get it free — ask on the pricing page.
Month by month
6 of 45 months up · best +9.1% Oct 2025 · worst −15.1% May 2023 · 10 losing months in a row at worst
| Year | Jan | Feb | Mar | Apr | May | Jun | Jul | Aug | Sep | Oct | Nov | Dec | Year |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2023 | −2.4% | −5.6% | +2.2% | −1.8% | −15.1% | −1.2% | −4.9% | −4.9% | −4.2% | +0.6% | −7.3% | −2.6% | −38.9% |
| 2024 | −10.3% | −5.3% | −3.3% | −5.7% | −9.8% | −5.2% | −3.5% | +4.5% | −6.3% | +5.5% | −3.4% | −3.3% | −38.2% |
| 2025 | −3.3% | −1.2% | −3.3% | −2.6% | −4.1% | +0.4% | −7.5% | −0.2% | −0.4% | +9.1% | −4.7% | −5.3% | −21.6% |
| 2026 | −2.9% | −4.3% | −6.2% | −9.7% | −5.2% | −1.3% | −10.1% | −4.4% | 0.0% | · | · | · | −36.7% |
Training · validation · out-of-sample
A Pro feature. Pro isn’t on sale yet; beta testers get it free — ask on the pricing page.
Risk-adjusted
- Sharpe
- -3.07
- Sortino
- -4.25
- Calmar
- -0.45
- Return ÷ max drawdown
- -0.99
- CAGR
- -36.69%
- Exposure
- 34.0%
A Pro feature. Pro isn’t on sale yet; beta testers get it free — ask on the pricing page.
A Pro feature. Pro isn’t on sale yet; beta testers get it free — ask on the pricing page.
Costs and mechanics
- Fees paid
- −$5,030.43
- Funding
- −$49.21
- Liquidations
- 0
- Warmup bars(loaded before the start date)
- 203
- Benchmark(buy and hold)
- +376.55%
Monte Carlo
A Pro feature. Pro isn’t on sale yet; beta testers get it free — ask on the pricing page.
All 1163 trades
A Pro feature. Pro isn’t on sale yet; beta testers get it free — ask on the pricing page.
How this was run
Orders fill at the next candle’s open, never at the close that produced the signal. When one candle’s range covered both the stop and the target, the stop was taken, because the data cannot say which came first. Fees, slippage and funding are all charged.
Warmup history was loaded from before 2023-01-01 so the requested window is fully tradable.