[D1] "Mouse Tail" - breakout strategy for daily charts

Dyer

New member
Apr 19, 2020
6
0
1
Isaiah said:
People, wash down the owl. It is very interesting what will happen at dolgosrok..
I wrote the basis for screwing filters along the length of the candles, and so on. Hands did not reach as right now the AI topic in development opened.
 

Quentin

New member
May 5, 2019
9
0
1
The impact of the quality of quotes on the results of testing an Expert Advisor with identical parameters. On the left - Dukas ' quotes uploaded to MT4 (adjusted for Alpari trading time), on the right - quotes uploaded via MT4 Alpari.
 
Last edited by a moderator:

Enoch

New member
May 6, 2019
9
0
1
And there is no adviser under MT5? There, it seems, quotes are automatically loaded for any instrument and with high quality.
 

Quentin

New member
May 5, 2019
9
0
1
To speed up calculations when optimizing the Expert Advisor, you need to set the volatility period to a minimum (somewhere at least 10 for varying the size of the tails by%). If the optimization is based on static parameters (Min/Max of the candle tail in pips), then this parameter is not needed at all. After optimizing for dynamic parameters, you can then play with the influence of the period of volatility (there is a suspicion that this parameter will only affect the stability of forward and back testing, and will not help much to increase the profit factor).
 

Enoch

New member
May 6, 2019
9
0
1
Quote
MT5 is a very niche terminal, none of its 10 fans have seen this topic, so you should not wait for an owl...
Then you'll have to do it yourself :)

In general, I get a different story, not like the previous testers :)
I also ran it for the last 5 years (from May 2015 to May 2020). Story quality = 99%. Orders were placed with a minimum candle body of 5 p. and a minimum shadow of 5 p. At a gap of more than 10 p. orders were not placed. Pending orders are placed at 0 o'clock and are held until 22 o'clock of this day, and at 23 o'clock orders are closed, if some have not yet closed.
Average for 5 years:
- it turns out that the most profitable period began only in October 2019, and before that there was almost no return, i.e. in quiet times, the strategy does not work very well;
- Mondays are usually unprofitable, and Fridays are 50/50;
- from May to September - unprofitable months and January (well, that is, in the midst of holidays, probably, when it is also quiet in the markets).

Spoiler alert
 

Brock

Member
Customer
Sep 1, 2011
40
0
6
Enoch said:
it turns out that the most profitable period began only in October 2019, and before that there was almost no return, i.e. in quiet times, the strategy does not really work;
That's right! Breakout strategies work well only when there is increased volatility. Now it is such a period, there is an opportunity to earn money.
 

Enoch

New member
May 6, 2019
9
0
1
Below is an expert Advisor for those who want to change, add, and check something...
And under the spoiler, its parameters.
Spoiler alert
The Expert Advisor is written for MT5, for tests, because MT5 has more features for testing and analyzing results.
The EA is made for new quote points, but the size of the candle, tails and gap in the source data of the EA are set in the old points, because it is more familiar (1 old point = 10 new points-this is so that everything is completely clear :)).
Install the MT5 terminal, for example, from the official website, connect the demo account (MT5) of the broker and test it on various tools. Quotes are loaded by themselves. Any timeframe, the Expert Advisor will still take data from D1. (This is if anyone thinks that MT5 is something complicated.)

Expert Advisor parameters.
D1_Body_Min - the minimum size of the candle body (in old points).
D1_Body_Max - the maximum size of the candle body (in old points).
D1_Tail_Min - the minimum size of the candle tail (in old points).
D1_Tail_Max - the maximum size of the candle tail (in old points).
Gep_Max - maximum gap size (in old points).
HourOpen - the hour of placing orders at the beginning of a new day (according to the time of the server (broker)).
HourClose - the hour of closing all orders (according to the time of the server (broker)). which are in the market and have not yet managed to close on a take or stop.
Trade_Mon - allow trading on Monday.
Trade_Tue - allow trading on Tuesday.
Trade_Wed - allow trading on Wednesday.
Trade_Thu - allow trading on Thursday.
Trade_Fri - allow trading on Friday.

Pending orders remain placed (waiting for activation) until the time that is 1 hour less than the HourClose parameter.
 

Enoch

New member
May 6, 2019
9
0
1
For GBPUSD, the optimal parameters for testing over the past year (from May 2019 to May 2020) (it may not make sense to look further into the history, since there were other conditions than now) were the following:
- min body of the previous day's candle = 5 p.;
- max body of the previous day's candle = does not matter;
- min shadow of the previous day's candle = 5 p.;
- max shadow of the previous day's candle = 200 p.;
- on Mondays, it is better not to trade, because there are more losses (this is not related to gaps);
- maximum gap size = doesn't matter.
As a result, we received 114% of the profit for the year, but the main profit was made in March, when there was a special shock in the markets (oil collapsed).

Spoiler alert
 

Quentin

New member
May 5, 2019
9
0
1
Added a check on the medjikam of deposits - now the bot has become multi-currency and will not touch the deposits of other owls.
 

Attachments

  • mousetail_v1.02.ex4
    22.5 KB · Views: 0

Quentin

New member
May 5, 2019
9
0
1
Try this version. If it doesn't help, look for jambs in your tester settings (correct testing time, loaded history n bars back from the initial testing date (where n is the volatility period used in the set), etc.
 

Attachments

  • mousetail_v1.02_for_tester.ex4
    22.7 KB · Views: 2
Last edited by a moderator:

nigger

New member
Jan 11, 2018
11
0
1
Quentin said:
Try this version. If it doesn't help, look for jambs in your tester settings (correct testing time, loaded history n bars back from the initial testing date (where n is the volatility period used in the set), etc.
I understood.It does not test on tick data.
 

Quentin

New member
May 5, 2019
9
0
1
I also test it on tick data. Judging by the error that the bot gave you, the parameter of the volatility period in the test set is greater than the number of quote bars loaded before the test date. For example, by default, the average annual volatility parameter is 260. If you have been testing since 2014, and your loaded quote history starts later than the end of 2019 or the beginning of 2020, you will get an error. Then either you need to load more bars deep into the history, or reduce the period of volatility in the test set. For example, judging by the optimization data, for GBPUSD, the volatility period (VolatilityPeriod) is enough for 10, if the parameters of the length of the candle tails are set in % of the volatility. If the parameters of the length of the candle tails are set in pips, then the period of volatility is not necessary at all (it can be set, for example, 1).
 
Last edited by a moderator:

Enoch

New member
May 6, 2019
9
0
1
So far, everything is working well. The strategy makes a profit from the moment it is published =d>
 

Case

Member
Customer
Mar 29, 2018
73
0
6
Quentin said:
For GBPUSD, the optimal parameters for testing over the past year (from May 2019 to May 2020) (it may not make sense to look further into the history, since there were different conditions than now)
So if history repeats itself for 2019 or 2020?
Quentin said:
I also test it on tick data.
And share the test?
 
Last edited by a moderator:

Quentin

New member
May 5, 2019
9
0
1
Case said:
And share the test?
Yes, please. The attachment contains the test results for the GBPUSD pair from 01.01.2010 to 30.04.2020. Dukas tick quotes of the history since 2008 were downloaded through the TickDownloader program and fed to the MT4 terminal with an adjustment for the trading time Alpari. The set used for the test, I have already laid out in this branch, but I will attach it here.
 

Attachments

  • mousetail_gbpusd_2010-2020_0,1lot10000_nomm.zip
    180.5 KB · Views: 1

Case

Member
Customer
Mar 29, 2018
73
0
6
Enoch said:
2. Forex is not a Swiss reliable bank. You need to be prepared for everything, including anticipating the onset of different "if" situations." :)
Then it is easier to flip a coin, if you do not calculate a large period of different markets.
 

Enoch

New member
May 6, 2019
9
0
1
Reed said:
that's it? only 2 pages of discussions about this strategy?
What would you like to discuss, for example?
The effectiveness of the strategy does not depend on the number of discussions)