Best Way To Approach Algo Trading?
Only people experienced with YOUR exact algo can answer. Nobody else.
Anyways, try to see how it performs out of sample. Here is a tool you can experiment with https://www.mql5.com/ru/code/27755
Only people experienced with YOUR exact algo can answer. Nobody else.
Anyways, try to see how it performs out of sample. Here is a tool you can experiment with https://www.mql5.com/ru/code/27755
Hi thanks for replying. Man, google translate does such a bad job on that page, I am really having a hard time to even understand what this tool does and how this can help me.
I thought that my "problem" was universal. There is nothing special about my algo, it´s very simple system using heikin ashi and MA cross.
The tool is called Validate and it allows you to "validate" any EA from yourself or even the market. If it does not validate, it is s***.
Try translating to English instead to whatever you speak. For me it works just fine with the internal translator, i am not a native English speaker and only have 3 brain cells to boot. If i can figure it out, so can you.
The tool is called Validate and it allows you to "validate" any EA from yourself or even the market. If it does not validate, it is ****
Try translating to English instead to whatever you speak. For me it works just fine with the internal translator, i am not a native English speaker and only have 3 brain cells to boot. If i can figure it out, so can you.
Yeah I got that, but the problem is that isn´t a tutorial for it. It is not explained in a didactic manner how to use it, it has a bunch of files it needs and tbh I don´t have patience to sort this out. Now, if you know of any youtube video of someone using this thing, that would be great.
What is considered the best practice when algo trading?
I have developed a robot for myself and it uses a simple strategy that "works".
I mean, I can run optimization and find settings that make a million dollars in a year or so.
Now the thing is that that setting only works for that period of time, if you try using that setting further in the past, it won´t work.
So the question is: do we run optimization and find the best settings for the very recent market? Like, do I use only the current month, or last 3 months only when optimizing, then grab the best settings and hope it keeps working (until it doesn´t). Then rinse repeat?
As with manual trading, there is no one right way. It depends what you want for yourself. I've been testing strategies for about 10 years. Out of sample is about 3 years. This means I can have the strategy traded live for 3 years without any problems. Then I would optimize again. But there are also algo traders who, for example, re-optimize every week and use new parameters.
I don't think there is a right or wrong there.
With a shorter backtest period, the risk of over-optimizing is of course higher. However, your strategy is also adapted to the current market phase.
However, to achieve a higher sample size, I would recommend choosing longer periods for backtest and out of sample test. The statistical significance is then simply higher.
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
You agree to website policy and terms of use
I have developed a robot for myself and it uses a simple strategy that "works".
I mean, I can run optimization and find settings that make a million dollars in a year or so.
Now the thing is that that setting only works for that period of time, if you try using that setting further in the past, it won´t work.
So the question is: do we run optimization and find the best settings for the very recent market? Like, do I use only the current month, or last 3 months only when optimizing, then grab the best settings and hope it keeps working (until it doesn´t). Then rinse repeat?