Experiment - page 200

New comment
 

Does this graph show the magnitude of the difference in movement? The straighter, the more correlated?

 
Aleksei Stepanenko:

Does this graph show the magnitude of the difference in movement? The straighter, the more correlated?

The closer to 100%, the more correlated. These are the ones you need to start truncating with.

I am currently being offered this by autosampling, but need to review everything manually, where the most optimal ones are and where there are no positions yet


 
khorosh:

There is information on correlationhere , including an interactive table.

Vitaly Muzichenko:

The closer to 100% - the more correlated. These are the ones you need to start truncating with.

I've got autosampling suggesting this now, but need to go over everything manually, where the most optimal ones are and where there are no positions yet


oh, thank you!

 
Aleksei Stepanenko:

Oh, thank you!

Here are more copies, you can find more to go with them, Yusufu can remove a few and leave one


 
Vitaly, cool! Thank you.
 
Aleksei Stepanenko:
Vitaly, cool! Thank you.

Alternatively, we can add a block for checking correlations instead of truncating pairs.

The principle is simple: a position is opened, a signal to open another one is received, we check the correlation, if more than 80%, we do not open this pair until it closes at that one.

When several positions are opened, we gather all the open ones into an array and, when a signal is received, substitute them one by one and check if the position already opened has the correlation of more than 80%.

This way is better and easier than truncating pairs.

 
Aleksei Stepanenko:

So I'm serious. I haven't done dependency studies or trading correlations. And I hang EAs on every pair, like Yusuf, where the eye falls ;).

Now I want to know how to do it right and get rid of doubles.

Don't listen to Vitya, he will give you lots of advice. Google Forex correlation calculator and you will find everything.
 
Олег avtomat:

A small error can manifest itself in a completely unexpected way.

I'm not talking about a property of the system (its inertia, etc.), but about the implementation of the algorithm in the code.

Yeah, that's true.
For example, I did not immediately recognize that I had made a contradiction and it has cost me a considerable drawdown. I spent three days on reversing the EA. Now it will start to work out, I hope.
If I'm using cents or demo, I don't care. But with quid, they don't forgive that.
 
Renat Akhtyamov:
Yeah, that's right.
For example, I did not immediately realise that I had made a contradiction and it cost me a significant drawdown. I spent three days on the EA. Now it will start to work out, I hope.
If I'm using cents or demo, I don't care. But that's not forgivable on quid.
Yeah, the whole of Wall Street is rallying to the hunt for 50 quid in an unprecedented way.
 
vladavd:
Yeah, all of Wall Street is rallying to the hunt for 50 quid in unprecedented ways.
What, did you just say that?
Come on.
Show me how you're gonna make some money on the sub, we'll have a laugh.
1...193194195196197198199200201202203204205206207...277
New comment