EA N7S_AO_772012 - page 6
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
Well, the Expert Advisor worked on the demo from 8.00 MSK Monday until 11.00 MSK this Friday. I optimized it by 4+1 preliminarily. I have worked with 10000$ account with lot 1. During this period of time there were 24 trades, 13 of which were losing (stop loss). The amount of profitable deals without trailing stop was $70-160. Several times I set a trailing stop of 35 pips. On these trades I received from $730 to $1,800 (maximum) in profit. Of the profitable deals that showed a modest result of $70-160 at least 3 reached a profit of $ 1000-1500, but because of the lack of trailing stop rolled to a minimum break-even level.
Result - at 11:00 pm Moscow time on Friday there was 6192 left in my account.
My conclusions.
1. The optimization period is insufficient.
2. The Expert Advisor has an inertia that affects the results when it changes the movement. The price has moved upwards, it tries to sell, and vice versa. And this happens several times in a row :-((.
3) I need a trailing stop.
Next week I will try Casper's L5_trail with optimization according to the 4+4 scheme.
And still some questions.
On what pairs and with what sets the EA has traded.
My account balance is now $3600 vs. $2400 on Monday, i.e. +1200 of which $800 and $400 of float have been fixed.
That is why I have to look into it and see if there is something wrong.
At the weekend I will give the final results for all nine pairs and give recommendations
L5_trail by Casper sometimes gives error 1 (unchanged parameters) when modifying. This is not my EA anymore, I cannot say anything, ask Casper to fix it.
I do not welcome this kind of trailing because I believe that profits should be allowed to grow.
However in the next version I will separate initial SL and step of pulling up to the market and switch to trailing either by SAR or by own trailing after some profit level.
I also agree that losses should be cut, but now when I look at the chart, I see that the market has changed, and the EA did not know about it and has acted in accordance with zeros and ones. We can try to prohibit opening of a second order following the losing one in the same direction on the same bar. But as applied to analogous systems, it has led to the twofold decrease of the drawdown and threefold decrease of the profit.
SHOOTER777, if you don't mind drawing a 4+4 diagram.
Because I only understand your periods with the diagram. :)
I will explain it once again without the schematic.
Your Expert Advisor has six groups of parameters which are optimized one by one.
Six groups are divided into two optimization periods, i.e. three groups for each period.
In each period, three groups are two layers NN.
For the first period, there are two groups of x and y in the first layer and one group of z in the second layer
For the second period, the first layer has two groups X and Y and the second layer has one group Z
So, although the Expert Advisor is universal, these periods can be different for every pair, don't get hung up on 4+4.
Try and search for each one depending on its variability.
Also, pairs are characterized by different volatility, so the SL optimization range can also be varied.
Regarding 4+4 in this case it means that both the first period and the second period are the same and equal to 4 weeks.
I would ask to use a different form of notation.
For example : 4&4 or 2&2 means the above, and 4+4 4+2 means that the second period is shifted relative to the first one by a week ahead.
...Continued.
We choose the best option and run it on test, admiring the picture.
Then we proceed to the second step. Once again open the Expert Advisor properties and go to the Input Parameters tab and load the file _step_2=y _l3.set
Run the optimization and wait for the results.
We select the best option and run it through and admire the picture.
Then proceed to the third stage. Once again open the Expert Advisor properties and go to the input parameters tab and load the file _step_3=z _l3.set
Run the optimization and wait for the results.
We select the best variant and run it through and look at the picture.
At this point the first optimization level is over. Besides, we have an opportunity to look into the future and test the Expert Advisor with the obtained results in the next fifth week preceding the sixth week.
And regret that we have missed a good profit.
To be continued...
I have no results in the second stage of optimization, just nothing.
What could this be related to and what to do?
I have a strange thing during the first stage too.
I have optimized the Expert Advisor several times during the first phase, resetting all parameters to initial values and loading file _step_1=x_l3.set, and each time I get different results during optimization.
What can it be related to? After all, I didn't change the price chart and optimization period.
Vovanych, this is actually quite simple.
The optimizer uses genetic algorithms to select optimization parameters, so it turns out that out of 10 million possible combinations, only 10 thousand chosen by this algorithm are enumerated. And coincidence of the whole set of numbers in several runs is unlikely. :)
Vovanych, this is actually quite simple.
The optimizer uses genetic algorithms to select optimization parameters, so it turns out that out of 10 million possible combinations, only 10 thousands chosen by this algorithm are enumerated. And coincidence of the whole set of numbers in several runs is unlikely. :)
So I have to disable "genetic algorithms" checkbox?
Then the result will always be the same?
And I always get no more than 5,500 results, not 10,000 as you said.
P.S. I have read it at your link, it is interesting))))) I will try to assimilate these chromosomes)))
At the weekend I will give the final results for all nine pairs and recommendations.
If possible, do the sampling from the history (since Jan. 5) and make a chart with the drawdown and other things please - I would like to see what turns out with the simultaneous use of the advisor on 9 pairs
Thank you
Results.
Last week's demo result is in the zip archive. It is difficult to filter the earlier data, because I am using one account to test different versions of the EA and it was not standing on all nine pairs all the time. The Excel file has a weekly summary with tests for the past three weeks. The results on the demo from the tester differ by no more than 10%.
Weekly totals (Cent. real. lot 0.1 deposit 5000)
eurusd = -185
gbpusd = 37.00
usdjpy = -16.40
eurgbp = 30.84
eurjpy = 495.36
gbpjpy = -304.29
usdcad = 203.57
usdchf = -126.07
Taking into account the statistics given by Nikolay for 3 weeks, we can consider that the pairs using this algorithm work normally
gbpusd
eurgbp
eurjpy
usdcad
Guaranteed to be profitable (Requiring another optimization than 1-4 + 4-5).
usdchf
audusd
The rest are either lucky to have profits around 0 or +- depending on the week.
Has anyone tried to optimise all 6 groups of parameters on a week and put the EA on 1 day (next)?