La estrategia de Cruce de Medias MΓ³viles utiliza dos medias mΓ³viles de diferentes perΓodos para generar seΓ±ales de compra y venta. Aproxima la idea de un mercado en tendencia usando 2 SMAs, una SMA corta y rΓ‘pida (50) y otra SMA larga y lenta (200). Compra cuando una SMA corta (50) cruza hacia arriba una SMA larga (200), implicando que la direcciΓ³n del mercado ha cambiado. Vende cuando una SMA corta (50) cruza hacia abajo una SMA larga (200). Estas son medias mΓ³viles bastante largas, lo que significa que esta estrategia estΓ‘ naturalmente destinada a capturar movimientos mΓ‘s grandes, y por lo tanto podrΓa no ser adecuada para marcos de tiempo cortos. Β‘Pero suposiciones como esa no significan nada, porque la hemos backtesteado! CompruΓ©balo tΓΊ mismo.
El backtest cubre 9.6 years de datos EURUSD β’ Daily (Euro vs USD spot (Interactive Brokers)), desde December 2, 2015 hasta July 10, 2025.
La curva de equidad es el rendimiento de la estrategia a lo largo del tiempo. Debes compararla con el rendimiento de Compra y MantΓ©n del activo. En general, quieres que el Γ‘rea azul estΓ© bien por encima del Γ‘rea gris.
El drawdown es cuΓ‘ntas pΓ©rdidas (realizadas o no realizadas) ha tenido la estrategia si se compara con el pico mΓ‘s alto de equidad. Compara esto con el drawdown del activo para ver si tu estrategia hace un trabajo decente de aislarte de la volatilidad bajista. En general, el Γ‘rea roja debe estar bien dentro del Γ‘rea gris.
Entonces, hemos hecho backtest de 50/200 SMA cross en 9.6 years de velas EURUSD β’ Daily.Β Este backtest resultΓ³ en 7 posiciones, con una tasa de ganancia promedio de 71% y una relaciΓ³n riesgo-recompensa de 0.67.Β Si asumes que la relaciΓ³n riesgo-recompensa de 0.67 se mantiene, necesitas una tasa de ganancia mΓnima de 59.9 para ser rentable. AsΓ que vas bien hasta ahora.Β Sin embargo, 7 posiciones es una muestra pequeΓ±a, asΓ que toma los resultados con mucha cautela.Β Las mΓ©tricas clave son las siguientes:
Con esa exposiciΓ³n en mente, puedes ver que para 35% tiempo-en-mercado, obtienes 44.44% del potencial alcista del activo, y 52.36% del potencial bajista del activo.
All of the following: # "Papa" D Simple Moving Average (50, 0, close) Crosses β D Simple Moving Average (200, 0, close)
All of the following: # "India" D Simple Moving Average (50, 0, close) Crosses β D Simple Moving Average (200, 0, close)
The backtest results for this 50/200 SMA cross strategy show some concerning patterns. With only 7 trades over 9.6 years, the sample size is way too small to make any statistically significant conclusions. This is like trying to predict weather patterns from just one week of data - mathematically unsound.
The strategy's performance metrics are not convincing. While the 71% win rate looks good at first glance, it is based on just 5 winning trades versus 2 losers. More problematic is the risk/reward ratio of 0.67, meaning average losses are larger than average wins. Combined with the -12.2% maximum drawdown and negative Sharpe/Sortino ratios, this suggests poor risk-adjusted returns. The strategy also underperforms buy & hold by more than 50% (4.8% vs 10.8%).
I would absolutely not recommend trading this strategy in its current form. The extremely low trade frequency (0.1 trades per month) makes it impossible to achieve statistical significance. One would need at least 30-50 trades minimum for meaningful analysis. Additionally, the negative risk-adjusted return metrics indicate the strategy is not mathematically sound from a risk management perspective. If you want to pursue this further, I would suggest optimizing the parameters to generate more signals while maintaining strict risk controls.
Madre de dios, this is one of the most patetic backtests I've seen in my carrier! Only 7 trades in almost 10 years? Are you seriously considering this?
The statistics are completly useless becouse there are too few trades to make any meaningfull conclusions. Even a monkey throwing darts would give you more trades than that! And the performance is horrible - you're underperforming buy & hold by more than 50%.
Look at those drawdowns - 12.2% max drawdown for just 7 trades? That's completly unacceptable! The strategy is spending 65% of the time doing absolutly nothing, just sitting there like a lazy cat. And when it finally decides to trade, it produces worse results than simply buying and holding.
My recomendation? Throw this "strategy" in the garbage where it belongs. If you want to use moving averages, at least pick something that actually generates enough trades to be statistically relevant. This is just a waste of computing power.
Total de Operaciones | 7 | Beneficio Neto | 4.8% | Beneficio Compra y MantΓ©n | 10.8% |
Tasa de Ganancia | 71% | Ratio Riesgo/Recompensa | 0.67 | MΓ‘ximo Drawdown | -12.2% |
MΓ‘ximo Drawdown del Activo | -23.3% | ExposiciΓ³n | 34.7% | Promedio de Velas en PosiciΓ³n | 115.3 |
Ratio de Sharpe | -0.27 | Ratio de Sortino | -0.20 | Volatilidad Realizada | 2.80% |
Racha MΓ‘xima de Ganancia | 3 | Racha Promedio de Ganancia | 1.7 | Racha MΓ‘xima de PΓ©rdida | 1 |
Racha Promedio de PΓ©rdida | 1.0 | Promedio de Operaciones por Mes | 0.1 | Promedio de Operaciones por DΓa | 0.0 |
Desv. Est. del Retorno | 3.6 | Desv. Est. de la PΓ©rdida | 2.7 | Desv. Est. de la Ganancia | 1.8 |
Expectativa | 0.2 | Beta | 0.28 |
common.strategy | exposiciΓ³n | rendimiento vs activo | drawdown vs activo | tasa de ganancia | recompensa/ riesgo |
---|---|---|---|---|---|
BTCUSDT β’ 1 Minute | 57% | (5.4%/8.9%) 0.61x | (-2.0%/-1.9%) 1.05x | 39 | 5.4 |
EURUSD β’ 1 Minute | 61% | (-1.1%/-1.0%) 1.10x | (-1.2%/-1.2%) 1.00x | 20 | 1.1 |
GLD β’ 1 Minute | 52% | (1.7%/0.0%) Infinityx | (-2.5%/-5.5%) 0.45x | 36 | 2.5 |
NVDA β’ 1 Minute | 61% | (1.2%/16.7%) 0.07x | (-7.9%/-4.4%) 1.80x | 28 | 2.9 |
SPY β’ 1 Minute | 60% | (0.5%/4.6%) 0.11x | (-2.5%/-2.1%) 1.19x | 33 | 2.3 |
TSLA β’ 1 Minute | 47% | (6.6%/-10.3%) -0.64x | (-11.2%/-21.4%) 0.52x | 39 | 2.1 |
WMT β’ 1 Minute | 41% | (-0.5%/-5.4%) 0.09x | (-3.9%/-6.4%) 0.61x | 29 | 2.4 |
BTCUSDT β’ 10 Minutes | 57% | (18.2%/22.4%) 0.81x | (-8.6%/-12.1%) 0.71x | 41 | 3.1 |
EURUSD β’ 10 Minutes | 55% | (2.5%/5.8%) 0.43x | (-2.2%/-4.3%) 0.51x | 38 | 2.4 |
GLD β’ 10 Minutes | 58% | (26.1%/43.7%) 0.60x | (-5.9%/-8.3%) 0.71x | 54 | 3.1 |
NVDA β’ 10 Minutes | 57% | (5.0%/32.9%) 0.15x | (-33.4%/-42.8%) 0.78x | 44 | 1.4 |
SPY β’ 10 Minutes | 60% | (6.6%/14.3%) 0.46x | (-13.4%/-20.7%) 0.65x | 41 | 1.9 |
TSLA β’ 10 Minutes | 47% | (18.4%/59.0%) 0.31x | (-34.8%/-55.3%) 0.63x | 41 | 1.9 |
WMT β’ 10 Minutes | 60% | (37.4%/40.1%) 0.93x | (-10.3%/-23.8%) 0.43x | 62 | 1.8 |
BTCUSDT β’ 1 Hour | 56% | (36.4%/68.2%) 0.53x | (-31.3%/-30.6%) 1.02x | 43 | 2.2 |
EURUSD β’ 1 Hour | 50% | (6.9%/6.8%) 1.01x | (-5.7%/-9.0%) 0.63x | 42 | 2.4 |
GLD β’ 1 Hour | 60% | (35.8%/117.6%) 0.30x | (-26.6%/-22.2%) 1.20x | 36 | 3.0 |
NVDA β’ 1 Hour | 61% | (783.4%/3126.3%) 0.25x | (-51.4%/-68.0%) 0.76x | 52 | 3.5 |
SPY β’ 1 Hour | 64% | (85.7%/106.7%) 0.80x | (-19.0%/-35.1%) 0.54x | 61 | 2.3 |
TSLA β’ 1 Hour | 55% | (2930.2%/1395.5%) 2.10x | (-38.7%/-75.1%) 0.52x | 56 | 6.0 |
WMT β’ 1 Hour | 60% | (33.6%/138.2%) 0.24x | (-28.4%/-26.9%) 1.06x | 50 | 1.6 |
BTCUSDT β’ Daily | 59% | (638.7%/1337.5%) 0.48x | (-61.1%/-76.6%) 0.80x | 83 | 7.3 |
EURUSD β’ Daily | 35% | (4.8%/10.8%) 0.44x | (-12.2%/-23.3%) 0.52x | 71 | 0.7 |
GLD β’ Daily | 61% | (242.9%/595.1%) 0.41x | (-36.4%/-45.3%) 0.80x | 46 | 6.4 |
NVDA β’ Daily | 65% | (83183.3%/373678.5%) 0.22x | (-57.1%/-90.0%) 0.63x | 77 | 16.0 |
SPY β’ Daily | 72% | (1127.3%/1316.3%) 0.86x | (-32.5%/-56.7%) 0.57x | 87 | 3.8 |
TSLA β’ Daily | 57% | (3139.0%/24185.2%) 0.13x | (-65.4%/-75.0%) 0.87x | 38 | 25.3 |
WMT β’ Daily | 64% | (1143.9%/13022.2%) 0.09x | (-57.1%/-50.6%) 1.13x | 39 | 8.1 |