Please tell us your opinion. - page 2
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
The result is great, and there is no skew in the number of trades. Could you be more specific: what does adaptive mean, what is the input or at least what is the dimensionality of the input vector, what is the size of the second layer, output is one or two (four), what was the network built on (it is clear from the avatar), how does the second test (network) differ from the first, what are the decision thresholds (or you have a flip), have you tried with other TF and instruments? I am interested, as I myself am torturing PNN, but have not achieved such results. I have not got any effect, thank you.
P.S. Answering my question: I have no doubts, that it will work in the future too.
Well, I will try to answer in brief. Adaptive means that it is retrained (changed) on every new bar. The dimension of the input vector is what? I did not write the network, but one good programmer did it for me, that's why I do not know its nuances. It is clear from the avatar that it is rewritten in MQL. There is only one way out. Tests differ in different network parameters and different indices that are on the input and output. The decision thresholds are selected by the genoptimizer. I tried it on 15 min, not bad either. I cannot use minutes, like the bettor does. I cannot understand why. But on the other hand, do I really need those minutes?
I have only one question to solve - how to evaluate the functionality of the TS in the future?
There's more about testing. For example, I take a period of history, for example 2 years, in which the price does not go higher than the initial one. After optimization or training I include almost the whole period (except for the last year) and smooth out the trend. After that I include only the last year and draw conclusions.
2 years on which TF?
to LeoV
The result is excellent. Let me explain. The market was changing at the beginning of February.
Everything non-adaptive that was working well in autumn started to fall down in February.
But your Expert Advisor held on in January and variant 2 managed to adapt to two different markets. That's great!
About testing also. For example, I take a segment of history of 2 years, in which the price does not go higher than the initial one. After optimization or training, I have included almost the entire period (except for the last year) and smoothed the trend. Then I include only the last year and draw conclusions.
2 years on which TF?
EURUSD 2004.01.01 - 2006.10.01
About testing also. I, for example, take a strip in the history of e.g. 2 years, in which the price does not rise above the initial one for the period. After optimization or trainings I enable almost the whole period (except for the last year) and smooth the trend. After that I include only the last year and make conclusions.
2 years on which TF?
EURUSD 2004.01.01 - 2006.10.01
what is the timeframe? a minute, 5 minutes, 15 minutes, 30 minutes, 1 hour, 4 hours, day, week? the number of bars depends on it.....
About the testing still. For example, I take a period of history, for example 2 years, in which the price of the period does not go higher than the initial one. After optimization or training I analyze almost the entire period (except for the last year) and smooth out the trend. After that I include only the last year and make conclusions.
2 years on which TF?
EURUSD 2004.01.01 - 2006.10.01
Here is an example of lean and mean time 2007.02.02 - 2008.04.02, SL90/TP30, LOT 0.1
The last month obviously points to overtraining
About testing also. For example, I take a period of history, for example 2 years, in which the price does not go higher than the initial one. After optimization or trainings I enable most of them (except for the last year) and smooth the tendency. After that I include only the last year and make conclusions.
2 years on which TF?
EURUSD 2004.01.01 - 2006.10.01
Here's an example of a skinny underperformance SL90/TP30, LOT 0.1
It's a good one. But on what timeframe?
The last month clearly indicates overtraining
What kind of indication? Do you have a neural net or something?
The last month clearly indicates overtraining
What kind of indication? Do you have a neural network or something?
No, it's a simple combination of two trend indicators with constant share on the trend. i'm just a follower of this strategy