Consistency of a trading system

[jrkob]

Hello,

I read somewhere on this forum that on way to check if a trading system is consistent, is obviously, for example to back-test it, or walkforward it. Results should be consistent within… I believe I read 95% confidence level, or something like that.

However, to me that rule shouldn’t apply to all systems. For instance, the current model I’m using is a volatility breakout system. I obviously don’t expect my system to perform the same way all the time, because the volatility changes. If the volatility of the instrument I’m trading decreases, I will find absolutely normal to have results not as good as if the volatility increases. Therefore, my system doesn’t give consistent profit return, and I’m not extremely worried as long as I am confident that when volatility returns, I will make money again.

What would worry me obviously, would be if my system doesn’t capture the next volatility breakout, because it would mean that my system didn’t capture the movement it was designed to capture.

Is any of you in the same case as me ? I expect most of you to say “yes”, but let’s see…

Cheers !

 

[garfangle]

Quote from jrkob:

Hello,

I read somewhere on this forum that on way to check if a trading system is consistent, is obviously, for example to back-test it, or walkforward it. Results should be consistent within… I believe I read 95% confidence level, or something like that.

However, to me that rule shouldn’t apply to all systems. For instance, the current model I’m using is a volatility breakout system. I obviously don’t expect my system to perform the same way all the time, because the volatility changes. If the volatility of the instrument I’m trading decreases, I will find absolutely normal to have results not as good as if the volatility increases. Therefore, my system doesn’t give consistent profit return, and I’m not extremely worried as long as I am confident that when volatility returns, I will make money again.

What would worry me obviously, would be if my system doesn’t capture the next volatility breakout, because it would mean that my system didn’t capture the movement it was designed to capture.

Is any of you in the same case as me ? I expect most of you to say “yes”, but let’s see…

Cheers !

Do a regression analysis. See if the correlation and r^2 numbers are high.

 

[klev]

Although past performance is no indication of future results, backtesting seems to be working for me. http://www.crontech.com/swp is a good place to start backtesting your strategies.

 

[Grob109]

Why did you choose a thing (volatility) that changes in the way it works over time?

 

[Buy1Sell2]

Quote from Grob109:

Why did you choose a thing (volatility) that changes in the way it works over time?

This was Grob109's first post 3 yrs ago-- Does anyone see where there has been an evolution process?