Getting more out of a robust strategy
6 replies
Oliver
3 years ago #259567
Hi
I have a strategy
this strategy has proven itself on demo and has passed all robustness tests. I have many others which are producing good results but im curious how i can get more out of them?
This particular strategy i have experimented with parameters – turning on duplicate trades. this has increased the ret/dd ratio. I have then run it with said settings in all robustness tests and its still passing
I have other strategies which do not have such positive results when allowing duplicate trades. but i wish to experiment by allowing duplicate trades, by setting a limit to the number of trades its allowed to take at any given time to see if i can still get positive results. id like to begin by allowing a max of 2 then potentially change that max to 3 etc to see if there is an optimal amount before it changes my opinion to not trade with live funds
I therefore ask if anyone can help me understand how to alter the strategy to give me the control i mention
In addition i have also tried to turn the strategy into a template to see if i can use such strategy to find on other markets but every time i generate a template to use in sqx, it yields no strategies so i feel im doing something wrong. therefore some help here would be most appreciated.
thanks in advance
Oliver
Oliver
3 years ago #259673
Anyone?
kasinath
3 years ago #260523
Did you ever get an answer to this? perhaps elsewhere?
mabi
3 years ago #260526
I a couple running that performs good in different versions that i altered in simple optimizer surprisingly they are pretty uncorrelated on daily basis but still profitable since 1 year back. On Demo thought.
hankeys
3 years ago #260558
playing too much with a strategy will lead to overoptimising it and worse results in the future
its better to have more strategies with different logics
You want to be a profitable algotrader? We started using StrateQuant software in early 2014. For now we have a very big knowhow for building EAs for every possible types of markets. We share this knowhow, apps, tools and also all final strategies with real traders. If you want to join us, fill in the FORM.
Insanity82007
3 years ago #260606
I agree with Hankeys on this one. Generate other strategies instead of optimising existing ones too much and have more strategies in your portfolio with low risk for each one than looking for a silver bullet
kasinath
3 years ago #260627
Agreed, for the most part.
95% of the time if we try to squeeze more performance out of a ‘good’ strategy, it will end up being overfit.
There is, however the 5% of cases (perhaps much less) where there can be additional gains.
It does depend on the strategy. It’s why I put ‘good’ in quotes, ha. For example, with a strategy designed to ride strong ‘fat tail’ (outlier) trends, you might be successful amending the strategy to add orders to winning positions. This wouldnt be applicable with a scalping strategy.
@oliver: I’ve attached a slide deck that may be helpful. It outlines one man’s approach to designing and refining an algo.
You might find some tips on refining a strategy, starting from the slide titled “Let’s Build a Trading System”
Viewing 6 replies - 1 through 6 (of 6 total)