Which type of backtesting performance would you choose?

[Addison]

I would like to ask everyone how you would choose backtesting performance metrics?

Among the following backtesting results conducted in the same time period with slightly adjusted strategies, which one would you choose and why?

[Yashar Seyyedin]

I prefer to look at balance/equity curve and choose the one that has more consistent output.

But in general knowing if a backtest is reliable is really difficult. Sometimes you have to check every single trade to asses the reliability of backtest.

[Mahdi Ebrahimzadeh]

jowysher:

I would like to ask everyone how you would choose backtesting performance metrics?

Among the following backtesting results conducted in the same time period with slightly adjusted strategies, which one would you choose and why?

Use OnTester() function even with a very simple optimization code to optimize your EA. this way you can compare all metrics somehow in single parameter. 

this is a simple one↓

double OnTester()
{
   if (TesterStatistics(STAT_TRADES) <50) return(0.0); // Better to increase 50 based on your opt. interval and time frame you use. I choose 50 trades as minimum for H1 and 6 months

   return(pow((double) (TesterStatistics(STAT_PROFIT_TRADES) * TesterStatistics(STAT_TRADES)),0.25) * TesterStatistics(STAT_PROFIT_FACTOR) * TesterStatistics(STAT_RECOVERY_FACTOR)/TesterStatistics(STAT_EQUITY_DD));
} 

[Addison]

Yashar Seyyedin

I prefer to look at balance/equity curve and choose the one that has more consistent output.

But in general knowing if a backtest is reliable is really difficult. Sometimes you have to check every single trade to asses the reliability of backtest.

Thank you for your response

[Addison]

Mahdi Ebrahimzadeh #:

Use OnTester() function even with a very simple optimization code to optimize your EA. this way you can compare all metrics somehow in single parameter. 

this is a simple one↓

Thank you, I will try to verify in this way

[Cuong Vu]

Yashar Seyyedin #:

I prefer to look at balance/equity curve and choose the one that has more consistent output.

But in general knowing if a backtest is reliable is really difficult. Sometimes you have to check every single trade to asses the reliability of backtest.

Balance/Equity curve is a great thing to look at to see if a trading performance is good for certain strategies. These curves only record balance and equity values after a trade is closed, it won't display the real drawdown as we see it in real trading.

That is why I like to see the drawdown max value. Of course  the return is important, so the ratio between the return and max drawdown is my favorite factor to evaluate a strategy/performance.

Regarding to your (jowysher) two test results, I would choose the one on the top since it has the Return/DDmax higher and also has many more trades.

[Addison]

Cuong Vu #:

Balance/Equity curve is a great thing to look at to see if a trading performance is good for certain strategies. These curves only record balance and equity values after a trade is closed, it won't display the real drawdown as we see it in real trading.

That is why I like to see the drawdown max value. Of course  the return is important, so the ratio between the return and max drawdown is my favorite factor to evaluate a strategy/performance.

Regarding to your two test results, I would choose the one on the top since it has the Return/DDmax higher and also has many more trades.

Currently I also chose the above strategy first. I used expected value for the calculations, haha, the backtesting results above were also better.