An interesting game - who can understand from transaction history how an advisor works - page 23

New comment
 
Folks have a suggestion in this thread to start/continue discussions on Safonov's ideas. Sanctus if off-topic stop.
 
I'm all for it.
 
Read the book, not much detail though. Does the first degree of derivation and forward/backward play really work.
 
Dezil >> :
Read the book, not much detail though. Does the first degree of derivation and forward/backward play really work.


Safonov's ideas resonate with the opinion of another member of our forum. I don't remember who, but the idea was as follows: Many EAs are profitable, but there are periods of the market for which the EA is not designed and during this period it fails. What we do, we just take another EA that is customized for the current market, and suspend this one on the market. That is all! This is not a strategy.
 
Yes the first advisor started to lose, we put another one for the changed market, it also started to lose because the market changed again, we put the first one back on.... we'll be one step behind.
 
There is a point here. Safonov proposes to focus on the law of random number inertia. And not to stop EAs on the demo so that it is possible to work in an additional dimension with each of them.
 
I have a question. There is a widespread opinion that one should enter only those trades in which TP/SL >= 3.00 (or in the vicinity). However Safonov showed that such proportions cause a large number of "lossless" deals. I'm closer to trend friends in trading and it is actually the most important criterion to filter my entries. But Safonov convincingly demonstrated the disadvantages of this filter in his calculations. Gentlemen (and especially Ryukhi TViMS) have an opinion on this idea.
 
To be honest, I have also glanced through the book so far, but I am thinking of using it in the following example. We have an EA with two MA with a certain period. As soon as the market changes, we stop the EA and look for other MA parameters for the current market and go back to work. The search is performed in accordance with the model proposed by Safonov..... Please correct me if I'm wrong. Maybe I haven't fully understood the book yet....
 
nikelodeon >> :
To be honest, I have also glanced through the book so far, but I am thinking to use it in the following example. We have an EA with two MA with a certain period. As soon as the market changes, we stop the EA and look for other MA parameters for the current market and go back to work. The search is performed in accordance with the model proposed by Safonov..... Please correct me if I'm wrong. Maybe I haven't fully understood the book yet....


Take the second position with something completely different from the mashka. I'd suggest something channeled. It's better when the strategies complement each other.
 
IlyaA >> :


Get something completely different from the mashka for the second position. I would suggest something from the channel. It's better when the strategies complement each other.

And it would be even better if you take 10-15-20 non-correlated systems, ideally even more. The only question is the price of the hardware that would be required.

1...161718192021222324252627282930
New comment