Avalanche - page 114
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
17 reversals... A corridor, on the borders of which we put the pendants, is visible. Without verbiage, show in the picture, an example of closing positions, overlapping, locking, manipulating volumes in an avalanche, what you will do to get out of this hell of a flat.
ZS. The more "scary" intervals are 11.2008 and 05.2009
(по каждому инструменту)
объем позиций
(по каждому инструменту)
Это у А***. У других, возможно, иначе.
Yes in others, because I've seen it, by the way, in A*******, this idea worked, but not on demo-micro (not on demo-micro)
Once again, I would like to emphasise that this is the best result that can be squeezed out by optimising the channel width.
In essence, it's an adjustment.
And if the channel width (and lot size increment) is made a function of the number of iterations, is it considered a fitting?
The topicstartner is certainly not the author of an invention called avalanche, if one considers that the main distinguishing part of it from other systems is the use of locking orders with consecutive increase in lot size.
.
Conclusions when researching Yuri's (horosh) TS - not to be confused with Avalanche. (Avalanche has many times worse results and there is no point in discussing it).
1) The tester works wonders as always and allows us to find parameters and parts of history, where we MIGHT make good profit.
2. TC has extremely low stability. At a slight change of any of the parameters or the starting point the results deteriorate at times or even by orders of magnitude, or else it is dumped. The optimization map looks like a sieve. In trading, it will be a minefield. The drawdowns increase LAVINOUSLY, the FV drops to 1-2 and below in 2 years and that's in the best case. Take comfort in the fact that thousands of trades are statistically accurate in this case is wrong. You should be guided by the worst result in the optimization chart when you shift any parameter by 10% and there are failures.
3. The obtained results have low MO (expected payoff). In reality we will lose on opening/closing and it will decrease even more.
4. With the achieved glossiness of individual runs, I would never risk (and do not advise others) to put such a TS on the real.
.
Here is the best result that the otiimizer produced after a night of work:
I am stopping any further work with this TS due to its futility. This is at best a self-deception.
All conclusions IMHO.
.
By the way, the tester, or rather its graphic tricks, helps us to fool ourselves. The maximum drawdown of equity at 8K on the initial balance of 3K should look like a deep sagging green line. For some reason it doesn't. Alexey (Mathemat) has pointed it out recently, and Reshetov accused someone of photoshopping. But no - it's the tester that helps us feel like gourams.
1. as always the tester works wonders and allows us to find parameters and areas of history where we MIGHT have made a good profit.
2. ...If we change any of the parameters or starting points slightly, the results deteriorate by times or even by orders of magnitude, or we are completely screwed...
You don't have to answer my questions, the topic will die down and everyone will calm down. :)
SZY. "It" - the possibility, when setting a certain checkbox, to create a report as of "last moment of no positions", and not after the forced closing of all positions on the "date" of the end of the testing period, would be my wish to improve the tester, but most likely it (wish) will not be implemented, and to collect a real!!!! statistics can be in function deinit(), checking for open orders. Here is just the optimization of....
>> "come to a stop"!? :)
Yes, THEY did. Actually an MC, not a smooth flush - I didn't put it accurately.
ZS I'm afraid the topic won't die for a long time - someone is benefiting from it.
Yes, THEY are. In fact MK.
Wrong. It's the end date of testing and the forced closing of positions precisely because of that, not the MK, which, by the way, when locking....... silent, silent, :):):):):) especially as it was written about it a few pages earlier.
Я уверен, что его советник сделан непрофессионально, Я уже указывал на его ошибки в программировании. Почему у goldtrader'a также использовавшего неттинговый подход результаты совсем другие?
Perhaps it is unprofessional... it was never intended to be a complete product... Maybe it's sloppy and sloppily written... features are not written... The design style suffers...But this does not mean that it is worse (in terms of functionality). Algorithm of avalanche is accurate to a decimal point. Both in the Expert Advisor with Avalanche itself and in its netting variant.
The algorithm is simple, elementary, very easy to formalize. And even a person who has never seen mql before can implement it in an EA... Don't try to distort things. The results are obvious.
No one has seen your version. But I'm absolutely convinced that there is something else besides the avalanche... It is this "something" that plays the principal and predominant role, and the avalanche is only one of the possible ways out. That's why the results are so different.
I wrote Avalanche without any additives and modifications, and I didn't write a "product", I wrote a "handicraft"... Exactly to make sure of its futility, and to lay out real evidence, not verbiage, which for hundreds of pages the topicstarter feeds us...
Wrong. This is the end date of testing and the forced closing of positions is because of this, not the MC, which, by the way, when locking....... silent, silent, :):):) :):) especially as it was written about it a few pages earlier.
1. I've already written I don't use locking. It does not give an advantage in anything. This is my strong opinion, I do not think it is necessary to argue and prove it.
2. I do runs on parmeters, where I get some holes in the optimization map and get the MC. In some cases the deposit may be enough and indeed the last position after the limited reversal is closed with a big loss, it depends on your luck. I.e. it depends on the ratio of start deposit/stop lot. But I chose it so as not to torment the tester for a long time.