Question
Can a long-only EMA 50/200 crossover system on BTC/USDT outperform a simple Bitcoin buy-and-hold benchmark over the 2020–2026 evaluation window?
Hypothesis
The initial hypothesis is that an EMA 50/200 trend filter may reduce some of the risk of permanent market exposure by exiting during bearish regimes. Accepting the hypothesis requires comparing return and risk metrics against the benchmark; profitability alone is not sufficient.
Dataset
BTCUSDT Spot 1D data was obtained from the Binance Public Data Archive. Archive files were verified using official Binance CHECKSUM files and the normalized dataset was frozen. The source range, including EMA warm-up, is 2017-08-17 to 2026-08-16. Source dataset rows: 3,287. Frozen dataset SHA256: b7a00cff14ae915c99f6d02a8fb2a2a14fa1abd36d287704b19cc67ba232697f.
Methodology
EMA 50 and EMA 200 are calculated on daily candles. Long entries occur after a bullish crossover confirmed at the daily close and execute at the next candle open. Exits follow a bearish crossover and execute at the next open. Position size is 100% of strategy equity, pyramiding is disabled, leverage is 1x, fee is 0.10% and slippage is 0.05%. The evaluation period is 2020-01-01 through 2026-08-16 and the benchmark is BTC Buy & Hold. The run was executed with runner NXQR-1.1.0.
Result
In the verified First Lab Run, the EMA 50/200 strategy produced a cumulative return of 550.26% and an annualized return of 32.67%. BTC Buy & Hold returned 771.57% over the same evaluation window, outperforming the strategy by about 221.31 percentage points. The strategy captured approximately 71.32% of the benchmark's cumulative return. Only 5 trades were generated, with a 60.00% win rate, 9.29 profit factor, -61.73% maximum drawdown, 0.83 Sharpe, 1.20 Sortino, 0.53 Calmar and 89.57 expectancy.
Conclusion
The baseline was profitable but did not outperform Buy & Hold on cumulative return. The high profit factor and 60% win rate are based on only five trades, which is not enough for strong statistical inference. A roughly 61.73% maximum drawdown also shows substantial residual risk. This experiment does not provide sufficient evidence that the EMA 50/200 baseline is superior to Buy & Hold; it is better treated as a reproducible baseline for subsequent strategy variants.
Limitations
This result covers only BTC/USDT Spot, the daily timeframe, one exchange/data source and one historical window. Only five trades were generated. Fees and slippage are fixed assumptions, while market impact, latency, liquidity constraints and other execution details are not fully simulated. No walk-forward, out-of-sample or multi-market validation is included. The current benchmark comparison is primarily return-focused and does not independently calculate every benchmark risk metric.