What's your opinion on optimization? I need some advice. - page 2
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
My opinion:
Backtesting as well as optimization is a complete waste of time.
It does not generate any profits.Well said .
At least we walk away from backtests (paired with coding) with a very clear cut example of what overfitting is.
Actually an encouraging graph, better than most. Usually a good formula has no need for optimising. Optimised parameters may has a chance that it will not work sometime into the future and there is no way to know when it will fail. Hence it is like you have made something wonderful but turned it into a time bomb? Having sufficient fund will solve the problem there.
Isn't a good formula some form of optimization ?
A way to make the approach relevant to the price action?
Actually an encouraging graph, better than most. Usually a good formula has no need for optimising. Optimised parameters may has a chance that it will not work sometime into the future and there is no way to know when it will fail. Hence it is like you have made something wonderful but turned it into a time bomb? Having sufficient fund will solve the problem there.
Hi, thank you for your words.
That graph was from my EA without any optimization, the parameters I was using were just "guessed". After posting it, what I did was run an optimization on the worst year (2012), where I had the big drawdown (43%).
From the optimization results (EURUSD 2012), I chose the set of parameters that could reliably run on any pair. The graphs below are out-of-sample backtests using the parameters I chose.
EURUSD 2012-2017
GBPUSD 2012-2017
EURJPY 2012-2017
USDSJPY 2015-2017
With a set of parameters I know I can trust (by "trust" I mean: they where chosen for a reason, not just becaused I guessed them) I am improving my EA. What I do is run visual backtests to understand how and when drawdowns occur. Then I change or create new sets of rules to counter it. If these new rules makes the EA run better (on all pairs and all years above and some more), I keep the changes. If not, I try something different.
After I am satisfied with the EA, I'll run walk-forward analysis (probably on more data, like 2007-2017, on all pairs I intend to run the EA) and do several monte carlo analysis. If it passes all tests, I'll run it live.
Before I do the walk-forward analysis, I'll return to this topic to ask for advice on the best pratices on how to run it.
Hi, thank you for your words.
That graph was from my EA without any optimization, the parameters I was using were just "guessed". After posting it, what I did was run an optimization on the worst year (2012), where I had the big drawdown (43%).
From the optimization results (EURUSD 2012), I chose the set of parameters that could reliably run on any pair. The graphs below are out-of-sample backtests using the parameters I chose.
EURUSD 2012-2017
GBPUSD 2012-2017
EURJPY 2012-2017
USDSJPY 2015-2017
With a set of parameters I know I can trust (by "trust" I mean: they where chosen for a reason, not just becaused I guessed them) I am improving my EA. What I do is run visual backtests to understand how and when drawdowns occur. Then I change or create new sets of rules to counter it. If these new rules makes the EA run better (on all pairs and all years above and some more), I keep the changes. If not, I try something different.
After I am satisfied with the EA, I'll run walk-forward analysis (probably on more data, like 2007-2017, on all pairs I intend to run the EA) and do several monte carlo analysis. If it passes all tests, I'll run it live.
Before I do the walk-forward analysis, I'll return to this topic to ask advice on the best pratices on how to run it.
It is very stable everywhere, much more relevant than a long speech, just go on !
Great work you did there Rafael !
It is very stable everywhere, much more relevant than a long speech, just go on !
Great work you did there Rafael !
Thank you for your kind words!
Encouragement is always welcome :)
Thank you for your kind words!
Encouragement is always welcome :)
Nor encouragement, nor kind ... descriptive, with nor love, nor hate : it is perfect.
Hi guys.
I've been learning lots of new coding tricks and making great improvements on my EA.
I'm currently downloading data from 2007 to 2018 to start walk forward analysis and I have some questions...
1. How many parameters should I optimize?
2. How do you chose which parameters to optimize? (My EA currently have 13 parameters that could be optimized)
3. What about parameters optimization steps... how many do you aim for? (For example if a parameter can be between 0 and 10, do you set step as 1 and have 10 results or set step as 5 and have 3 results (0,5,10)?
4. What about time period?
Following this example, I would optimize 8 months and run the next 4. Repeat process each 4 months. How do you chose how long should be optimzed, how long it should backtest and how often it should be done?
I really appreciate any help I can get.
PS- I'll post some new backtest and even videos of my EA running if you are interested in seeing how it's performing.
Hi guys.
I've been learning lots of new coding tricks and making great improvements on my EA.
I'm currently downloading data from 2007 to 2018 to start walk forward analysis and I have some questions...
1. How many parameters should I optimize?
2. How do you chose which parameters to optimize? (My EA currently have 13 parameters that could be optimized)
3. What about parameters optimization steps... how many do you aim for? (For example if a parameter can be between 0 and 10, do you set step as 1 and have 10 results or set step as 5 and have 3 results (0,5,10)?
4. What about time period?
Following this example, I would optimize 8 months and run the next 4. Repeat process each 4 months. How do you chose how long should be optimzed, how long it should backtest and how often it should be done?
I really appreciate any help I can get.
PS- I'll post some new backtest and even videos of my EA running if you are interested in seeing how it's performing.
Hi Rafael !
If you wishes to show its stability over the last x years to sell it as a ready to use for example, then you could do 3/4 + forward 1/4.
But in truth you won't need more than 6 months to match last market conditions. To repeat quaterly or as soon as you notice it slow down.
Maybe someone will answer you better than me.