Looking for patterns - page 35

 
Aleksei Stepanenko:
...

EURUSD. H1 timeframe. Period 2000-2020.

...

Alexei, you've been using 20 years' worth of data since the beginning of the thread. Now even hours. This is already non-trivial. Could you please indicate the source of this data?

 
Aleksei Stepanenko:

Maznev Expert Rails Expert Advisor 1.0

I would like to congratulate everyone, this is the first Expert Advisor we have created together. It is far from perfect, but it is a good start. Hurrah!

So, the Expert Advisor has the only Rails indicator so far. Deals are opened by the signals of this indicator and closed by Stop or Take. It is a wooden exit yet, but we will change it soon.

An interesting feature of the profitability displaying of the Expert Advisor is OnTester value. It consists of 4 digits: the first two represent the percentage of profit gained and the second two represent the percentage of profitable trades. Thus, during optimization we can immediately see the quality of trades opened by the Expert Advisor with certain input parameters.

All testing should always be performed in the "Open prices" mode to save our life time. The algorithm is implemented without unnecessary loops, so everything should fly.

Points for Stop, Take, and in general, here and always indicate a 4-digit quote. The Expert Advisor will translate it inside if it is needed.

I have run the optimization with the genetic algorithm as a hastily as possible. EURUSD. H1 timeframe. Period of 2000-2020. Parameters in the set-file below. The following results have been obtained:

What can I say? Weak. The first thing we should do is to get rid of the iron stop and teach our Expert Advisor to exit on a pullback that will minimize losses. The stop will be reserved for those rare cases when the price is flying with the arrow.

Alexey, maybe try binding a stop and take to the ATR?

However, at different volatility different values of take and stop are desirable.

And you won't have to opt for their values for each instrument.

 
Vladimir:

Could you please indicate the source of this data?

Vladimir, hello there!

I did not understand the question a bit. Regarding the timeframe, I am used to considering the daily movements, on H1 everything is visible, and you can test it quickly. The period of 20 years - I take the maximum time interval, excluding the daily movements on the hour timeframe till 1999. The source of quotes is a regular brokerage company.

Spread the word, we can change something.

 
Aleksei Stepanenko:

Morning, Alexey. What period, on what TF and with what parameters were you testing?

 
Boris Gulikov:

tie the stop and take to the ATR ?

Thanks for the idea, it's worth a try. I don't really like averaging, of course, but I can't do without it in some places. What are your impressions and nuances concerning the improvement of the system by using ATR?

My current plan is to go over results of deals on the chart and see which ones are winning or losing and why. I think the loss is influenced by trend. At first I thought to improve stops, but the right thing to do would be to look into the cause of successes and failures.

Boris, if you have any ideas, get in touch. I would appreciate it.
 
Алексей Тарабанов:

Shorting apples is impossible, a short already means a derivative, whatever that short might be.

Once again, a fakap on the flat. you're giving up, Colonel )

 
Vitaliy Maznev:

Vitaly, the period is 2000-2020, on H1.

Parameters are in the second file, they can be loaded into the Expert Advisor in "Expert properties"

 
Aleksei Stepanenko:

Vitaly, the period is 2000-2020, on H1.

Parameters are in the second file, they can be loaded into Expert Advisor in the Expert Advisor Properties

The reason I asked... The system was originally designed for a day, not for an hour. I meant the indicator parameters.

 

I transferred the indicator to the EA and it has the same parameters: 30-60-50-40-300 and a stop with a 180-160 take.

As for the daily, you can try it that way. It will not have many trades, any result will be unreliable, I think. You may try it, though.

 
Aleksei Stepanenko:

Thanks for the idea, it's worth a try. I don't really like averaging, of course, but I can't do without it in some places. Now I plan to go through the results of trades on the chart and see which trades are in the plus, which are in the red and why. I think the loss is influenced by trend.

At first I thought to improve stops, but it would be more correct to understand the reason of successes and failures.


Boris, if you have any ideas, get in touch. I would be grateful.

Thanks for the invitation! )

What does averaging have to do with it?

No averaging. Only dynamic stop loss and take profit.

I think it's worth doing - the test results should improve considerably. That's a good place to start.

If you have volatility of, say, 20 bars - 50 pips, then take and stop of 20 pips is ok. But if volatility is 200 points, then take and stop of 20 points is not enough.

Attach trend filtering wizard. It is the simplest trend filter. It will do for starters.