Looking for an MTS. giving a steady 20% or more per month - page 33
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 day X has arrived.
The market has been running for more than 7 hours and the account and investment password are still missing.
>> see my personal message.
1. quotation history.
2. System mathematics.
Optimization of the Expert Advisor for currency pair.
- The system uses StopLoss, TakeProfit, TrallingStop.
The simple mathematics of optimizing such an Expert Advisor should be at least as follows
StopLoss >= TakeProfit * K
TakeProfit <= TrallingStop * N
K, N - coefficients
4. Money management
5. Using news.
6. Technical analysis.
7. Not to interfere with expert trading.
8. Development of the Expert Advisor.
If you do not want to have a separate topic, let's continue in this one.
By the way, the topicstarter just threw a single post and no longer appears on the forum at all :)
On points.
1. If my using the quotes from different brokerage companies, I may be surprised by the difference in quotes but I have not found any fundamental difference between them except for non-market deals (spikes). If you want to have a base of history of all working pairs with one particular brokerage company you will mostly work with, then it will be enough as a reasonable variant. Such history is self-synchronized and takes into account peculiarities of the brokerage company.
2. Mathematics, of course, is the salt of any strategy.
3) Optimization for each currency pair - it's absolutely obvious.
Really, I think this postulate:
StopLoss >= TakeProfit * K
is of limited use. I would say this variant is good for a strategy based on the flat. For a trend-following strategy, the opposite is true - Take Profit should be higher than Stop Loss.
4. absolutely obvious.
5. How does it take into account the news in an Expert Advisor?
6. Sabo somoi :)
7. It's also absolutely obvious. You have made your Expert Advisor, do not disturb it. Let me add the following - a week is the absolute minimum for strategies working on short timeframes. For longer strategies, a month is needed. And these figures are really absolute minimums. Some well-known short-term strategies can give quite a long drawdown. I would also add that besides testing on a demo account, small additional testing on a micro-real is needed.
8. Plus I would add that in addition to the development on paper, you must thoroughly test on a long history, it does not matter, in Metatrader or Matlab, or in any other tools. After all, there are often people who haven't done any deals, have a vague idea about metatrader, don't even speak Russian, and for a "brilliant idea" on paper are willing to charge the others about $1k.
For my part, I would add the 9th point.
Before you start working on real EAs, especially multicurrency ones, you need a good working framework, that allows you to create and test strategies at all three levels with minimal costs and maximum reliability - tester, demo and microreal-time. The development of this framework takes much more effort than the coding of the strategy. But the costs are fully recouped, because the further use of such a framework saves you from unnecessary labor costs and stupid mistakes in future developments.
By the way, I have sufficiently thought out and developed such a framework in MQL4, but its final completion is still a long way off. At least 2-3 months. I think when I finish it I'll post it in codebase.
5. How do you consider the news in your EA?
Personally, I can remember when I signed up to work in the news back in 2006. Of course it was my own bullshit, but nevertheless I still have the impression.
I either prohibit to trade on the news or N hours beforehand if there are no open positions, or try to cover open positions. Stops are tightened, positions are locked. This variant will disappear now because of the smart guys ...
In general, the news is a fast movement, fast money, maximum adrenaline, as my friend says, it's not good for a normal guy.
>> you want to get some sleep. >> wait till 2:00 a.m. :)
Firemast, it's like a Russian proverb with you.
Why would you open all the pairs in one day? You had a week ahead, the entries on some pairs were not optimal and you had to wait for the right moment. But you rushed into the market with your orders.
I either forbid to trade on news or N hours in advance if positions are not open, or try to cover open positions. Stops are tightened, positions are locked, now this option will fall away, there are smart guys...
No, I got it. I mean, is it technically feasible in MQL4? I checked all the functions again - there seems to be nothing for working with news. How to determine the news release, its type? Or just by time?
When I was doing a statistical analysis of quotes, I noticed three distinct volatility spikes. The smallest was half an hour. But even for short-term strategies, half an hour is too short a time to be accounted for in any way. I have found that the techno-analysis contribution is much larger than the half-hour spike. The other is twice as much - the start of each hour.
But the most significant contribution to powerful market movements are obtained at the openings of major exchanges. But this is pretty much an obvious point and not too difficult to account for all year except for the winter/summer time changeover moments. In another thread I touched on the issue of different countries having different time changes. In general, the dissinchra can stretch over a month.
No, I got that. I mean, is it technically feasible within MQL4? I looked through the whole set of functions again - there seems to be nothing for working with news. How to determine the news release, its type? Or just by time?
When I was doing a statistical analysis of quotes, I noticed three distinct volatility spikes. The smallest was half an hour. But even for short-term strategies, half an hour is too short a time to be accounted for in any way. I have found that the techno-analysis contribution is much larger than the half-hour spike. The other is twice as much - the start of each hour.
But the most significant contribution to powerful market movements are obtained at the openings of major exchanges. But this is pretty much an obvious point and not too difficult to account for all year except for the winter/summer time changeover moments. In another thread I touched on the issue of different countries having different time changes. In general, the dissinchra can stretch over a month.
This is where I take the news for the whole month at once.
http://www.forexfactory.com/calendar.php?month=5&year=2009&fullmonth=1
Then I prepare a list of news. There is a file with important news, to which the market reacts. Expert Advisor reads information in arrays and compares time of news and time of terminal and adjusts time of brokerage company and calendar saved in GTM. If I find important news the flag of blocking new orders opened 2 hours before the news releases has been triggered, then we pull stops and close positions. 15 minutes before the news release either cover the trade or lock the position to avoid sharp ups and downs. What to do when you know about an event is everyone's business.
...
As many have noticed I use many currency pairs in my strategy, all these pairs are synchronous with each other, minute by minute. This gives a huge advantage when testing such multicurrency Expert Advisors. Selection of optimal parameters for each pair. If necessary, we merge all the reports into one report with a number of conditions and checks. Extraction of favourable moments for closing of some or all pairs.
....
For a tester, yes you can. How do you do synchronization on the real. I would be interested to see an example. Let it be an indicator of the sum of all currencies, the main thing that would be synchronized quotes for currency pairs.
Thanks in advance.
I do not have any link to your website in my profile.
Yeah, I see.
A set of countermeasures is generally obvious. The only thing I think is worth doing is back-testing the countermeasures over a long period of time. Sometimes you think that the contribution of a measure should be big, but it is small, and on the contrary, you underestimate something, and the back-tests show that it should not be neglected.
Thus, I thought that pulling a stop lossless position at the first opportunity must seriously increase the efficiency. In flat strategies pulling did not give the expected result, while pulling at the first opportunity worked better in trend strategies than losing at the first opportunity.
A closer look reveals that a close (breakeven) stop often licks off small swings, even when price has already moved in the right direction. A sort of "lost profit".
Have you tried to test the effectiveness of each measure individually?
For a tester, yes you can. How do you achieve synchronisation in real life? It would be interesting to see an example. Let it be an indicator of the sum of all currencies, the main thing that would be synchronized quotes for currency pairs.
Thanks in advance.
Z.U. And in the profile is no link to your site
You can not achieve synchronization in real time and will not achieve synchronization, only for the tester. but correctly tested expert, in real time and sees itself more confident.
I don't have a website yet. Still under construction ....