Esta estrategia explota un mito bien conocido, que dice que si compras al cierre de la sesión del mercado de valores y luego simplemente vendes al dÃa siguiente en la apertura, entonces te irá bien. Puedes ver por ti mismo que los resultados son... diversos, por decir lo menos. ¿Puede comprar al cierre y vender en la apertura hacerte ganar dinero? ¿Es sostenible? Compruébalo tú mismo, en los resultados del backtest a continuación. Recuerda que tal vez hay activos donde eso podrÃa tener sentido. De todas formas, no pudimos cubrir todas las acciones y ETFs de EE.UU.
El backtest cubre 12.6 months de datos NVDA • 10 Minutes (NVIDIA Corporation), desde June 28, 2024 hasta July 11, 2025.
La curva de equidad es el rendimiento de la estrategia a lo largo del tiempo. Debes compararla con el rendimiento de Compra y Mantén del activo. En general, quieres que el área azul esté bien por encima del área gris.
El drawdown es cuántas pérdidas (realizadas o no realizadas) ha tenido la estrategia si se compara con el pico más 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 área roja debe estar bien dentro del área gris.
Entonces, hemos hecho backtest de Buy at close/Sell at open en 12.6 months de velas NVDA • 10 Minutes. Este backtest resultó en 255 posiciones, con una tasa de ganancia promedio de 50% y una relación riesgo-recompensa de 1.06. Si asumes que la relación riesgo-recompensa de 1.06 se mantiene, necesitas una tasa de ganancia mÃnima de 48.5 para ser rentable. Asà que vas bien hasta ahora. Las métricas clave son las siguientes:
Con esa exposición en mente, puedes ver que para 5% tiempo-en-mercado, obtienes 14.59% del potencial alcista del activo, y 48.83% del potencial bajista del activo.
All of the following: # Yankee 10min Date&Time, Custom TZ (UTC+3), Time = 1550
All of the following: # Bravo 10min Date&Time, Custom TZ (UTC+3), Time = 930
The strategy looks quite mediocre from mathematical perspective. The win rate of exactly 50% with nearly identical win and loss sizes (0.86% vs -0.81%) suggests this is essentially a coin flip with tiny edge. While the Risk/Reward ratio of 1.06 is technically positive, it's far too small to be meaningful.
What concerns me most is the negative Sharpe ratio of -0.11. This indicates poor risk-adjusted returns - you're taking more risk than the returns justify. The market exposure of only 5.1% is quite low, which explains why you're significantly underperforming buy & hold (4.8% vs 32.9%). The strategy essentially misses most of the market moves while still exposing you to significant drawdown risk of -20.9%.
The high trade frequency (40.5 trades per month) combined with such small average gains suggests that transaction costs, which aren't included in this backtest, would likely eat up any theoretical profits. The strategy shows no statistical evidence of having predictive power - it appears to be just capturing random market noise rather than any real pattern. I would not recommend trading this without significant modifications to improve the edge.
Madre mÃa, this strategy is complete garbage! Let me tell you why, amigo.
First of all, you're making only 4.8% while the market made 32.9%? That's pathetic! You're basically losing money when you factor in inflation and trading costs. And look at that market exposure - only 5.1%! You're sitting out of the market 95% of the time during one of the strongest bull runs in NVDA's history. ¡Qué desastre!
The win rate is exactly 50-50, like flipping a coin, and your average wins barely exceed your average losses. Yes, you have a tiny positive R/R of 1.06, but with that horrible Sharpe ratio of -0.11, you're taking on risk without proper compensation. And that 20.9% drawdown? ¡Dios mÃo! That's massive for such little market exposure.
The only slightly positive thing I can say is that your win rate has some leeway above the minimal required win rate. But honestly, with these metrics, you'd be better off buying and holding or putting your money under your mattress. This strategy is a complete waste of time and computing power.
Total de Operaciones | 255 | Beneficio Neto | 4.8% | Beneficio Compra y Mantén | 32.9% |
Tasa de Ganancia | 50% | Ratio Riesgo/Recompensa | 1.06 | Máximo Drawdown | -20.9% |
Máximo Drawdown del Activo | -42.8% | Exposición | 5.1% | Promedio de Velas en Posición | 1.0 |
Ratio de Sharpe | -0.11 | Ratio de Sortino | 0.53 | Volatilidad Realizada | 17.27% |
Racha Máxima de Ganancia | 7 | Racha Promedio de Ganancia | 1.9 | Racha Máxima de Pérdida | 5 |
Racha Promedio de Pérdida | 1.8 | Promedio de Operaciones por Mes | 40.5 | Promedio de Operaciones por DÃa | 1.3 |
Desv. Est. del Retorno | 1.1 | Desv. Est. de la Pérdida | 0.7 | Desv. Est. de la Ganancia | 0.8 |
Expectativa | 0.0 | Beta | 0.1 |
common.strategy | exposición | rendimiento vs activo | drawdown vs activo | tasa de ganancia | recompensa/ riesgo |
---|---|---|---|---|---|
GLD • 10 Minutes | 5% | (-0.5%/43.7%) -0.01x | (-6.5%/-8.3%) 0.78x | 50 | 1.0 |
NVDA • 10 Minutes | 5% | (4.8%/32.9%) 0.15x | (-20.9%/-42.8%) 0.49x | 50 | 1.1 |
SPY • 10 Minutes | 5% | (1.0%/14.3%) 0.07x | (-4.8%/-20.7%) 0.23x | 51 | 1.0 |
TSLA • 10 Minutes | 5% | (24.5%/59.0%) 0.42x | (-12.6%/-55.3%) 0.23x | 50 | 1.2 |
WMT • 10 Minutes | 5% | (17.8%/40.1%) 0.44x | (-3.8%/-23.8%) 0.16x | 52 | 1.3 |