BTM - Expert Advisor (EA) | Page 113 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

MT4 expert advisor 2 replies

Expert Advisor 8 replies

Vegas Daily MT4 Expert Advisor results 29 replies

Can someone create for me an easy Expert Advisor? 0 replies

Interbank FX Expert Advisor how to program? 15 replies

Attachments: BTM - Expert Advisor (EA)
Exit Attachments

BTM - Expert Advisor (EA)

  • #2,241
  • Feb 2, 2020 12:23am Feb 2, 2020 12:23am
Just sharing...
I just think of that in the original BTM thread, BTM has once said that he didn't trade this method with CHF pairs and EURGBP.
And I just turn to look at my trading report in fxbook and waow, I think there's something true (as pictured).
For sure I know there aren't even enough sample size yet, and mayve these pairs just need some different setup and filter,
but 3 months of trading orders may have a little value for reference. (my setting: r=0.25, no filters applied)

And also truly appreciate what Joeef has done right here...salute
You may be the only one here in FF that I dare to put your project immediately in a live acct for even testing but not demo,
(of coz I would put the risk on my back haha)
Green pips to y'all!
Attached Image (click to enlarge)
Click to Enlarge Name: trading report.png Size: 122 KB
  • #2,242
  • Feb 2, 2020 4:15am Feb 2, 2020 4:15am
But I have one newbie question right here
as pictured, what are those little red and blue cross mean?
Thanks for help
Attached Image (click to enlarge)
Click to Enlarge Name: Screenshot1.png Size: 56 KB
  • #2,243
  • Feb 2, 2020 4:28am Feb 2, 2020 4:28am
Last week JFC
RR > 0.4 setup
PercentRangeFilter="65"
Attached Image (click to enlarge)
Click to Enlarge Name: New Bitmap Image 2.jpg Size: 56 KB
Don't NO-SL and drive! And remember: Always keep your eyes on the trend!
  • #2,244
  • Feb 2, 2020 10:02am Feb 2, 2020 10:02am
Quoting ots931010
Just sharing... I just think of that in the original BTM thread, BTM has once said that he didn't trade this method with CHF pairs and EURGBP. And I just turn to look at my trading report in fxbook and waow, I think there's something true (as pictured). For sure I know there aren't even enough sample size yet, and mayve these pairs just need some different setup and filter, but 3 months of trading orders may have a little value for reference. (my setting: r=0.25, no filters applied) And also truly appreciate what Joeef has done right here...salute...
Your post let me have an interesting idea when combined with correlation coefficient, my idea is, when two highly correlated currencies (increase or decrease together) are usually due to the same or similar fundamental reason (or simply do not care), then the pair will mainly size way, choppy and unclear trend move, so the trend follow systems will not work well in this situation.
I tested by recording the correlation coefficient (based on USD) of the pair before 1 day of opening the orders, then recalculated the profit factor of adding cases that removed all orders with correlation coefficient >0.7 (70% correlation). Detailed results in attached excel file, preliminary as follows:
Correlation coefficient <1 (default):
Profit: 1175.405
Loss: (474,815)
PF: 2.48
Correlation coefficient <0.7:
Profit: 1032.09
Loss: (174,265)
PF: 5.92
The limitation of this test except that the number of samples is only within 3 weeks of my results is:
- Because cannot use USD to compare with USD, the pairs of USD I have not considered
- Available correlation coefficients are calculated based on a certain currency, this currency can increase and decrease so the limit <0.7 of each day is not the same, if applicable, there must be 1 calculation algorithm A relative margin based on the highest and lowest of each day.
@Joeff: You may consider to add this filter ...
Attached File(s)
File Type: xlsx BTMEA_correlation_coefficient_test.xlsx   31 KB | 292 downloads
  • #2,245
  • Feb 2, 2020 8:19pm Feb 2, 2020 8:19pm
Quoting Thgoldenstar
{quote} Your post let me have an interesting idea when combined with correlation coefficient, my idea is, when two highly correlated currencies (increase or decrease together) are usually due to the same or similar fundamental reason (or simply do not care), then the pair will mainly size way, choppy and unclear trend move, so the trend follow systems will not work well in this situation. I tested by recording the correlation coefficient (based on USD) of the pair before 1 day of opening the orders, then recalculated the profit factor of adding...
Oh waow, what a huge finding for me, much appreciated!
Thanks a lot for sharing.
by the way, what's the tool you used to calculate correlation?
  • #2,246
  • Feb 2, 2020 11:17pm Feb 2, 2020 11:17pm
Quoting ForexKing998
‘When the EA start losing stop trading ’ does that mean we have to have 2 accounts : first one (demo) for dummy trading to detect the losing one and wait till it starts winning . Only then we allow the Second (real ) account to start trading (winning) till it loses and stop trading it again ?
You can try that (this is a work in progress) so anything that helps. You may need to run a few accounts to try different results. If you look at the posts Vicente placed using default settings on a long term demo you can see the account draw down periods. JoeeF has recently posted different configurations giving different results. So for now that would be one way, or cut your lot size back to reduce the account draw down or disable the Ea for a few days to reduce the percentage of account draw down. Once that period is over the gains seem to be there.
Here are my results for last week on a small cent account. The NZDCAD had bounced off the new next day tp1 before I got to it so I closed the first leg manually. The second leg to the new day pivot did then complete also. The GBPAUD, CADJPY,USDCAD did take profit also but for some reason did not highlight green. This has happened before and I have been meaning to ask the broker about this. Anyway results below same settings as post 2,166.
Attached Image (click to enlarge)
Click to Enlarge Name: Trades 27-31 Jan 2020.jpg Size: 218 KB
  • #2,247
  • Feb 2, 2020 11:33pm Feb 2, 2020 11:33pm
Quoting ots931010
{quote} Oh waow, what a huge finding for me, much appreciated! Thanks a lot for sharing. by the way, what's the tool you used to calculate correlation?
I use correlation coefficient indicator in tradingview
  • #2,248
  • Feb 3, 2020 2:53am Feb 3, 2020 2:53am
Quoting Thgoldenstar
{quote} Your post let me have an interesting idea when combined with correlation coefficient, my idea is, when two highly correlated currencies (increase or decrease together) are usually due to the same or similar fundamental reason (or simply do not care), then the pair will mainly size way, choppy and unclear trend move, so the trend follow systems will not work well in this situation. I tested by recording the correlation coefficient (based on USD) of the pair before 1 day of opening the orders, then recalculated the profit factor of adding cases that removed all orders with correlation coefficient >0.7 (70% correlation). Detailed results in attached excel file, preliminary as follows: Correlation coefficient <1 (default): Profit: 1175.405 Loss: (474,815) PF: 2.48 Correlation coefficient <0.7: Profit: 1032.09 Loss: (174,265) PF: 5.92 The limitation of this test except that the number of samples is only within 3 weeks of my results is: - Because cannot use USD to compare with USD, the pairs of USD I have not considered - Available correlation coefficients are calculated based on a certain currency, this currency can increase and decrease so the limit <0.7 of each day is not the same, if applicable, there must be 1 calculation algorithm A relative margin based on the highest and lowest of each day. @Joeff: You may consider to add this filter ... {file}
The impact of correlation might be huge. It's true. Definitely. Anyway, it's important to find proper approach to solve this issue. Thank you for your suggestions.

Take care.
And you will leave pain to take care of itself.
  • #2,249
  • Feb 3, 2020 4:26am Feb 3, 2020 4:26am
January 2020 - JF4XTrader v1.05 - JFXC variant ...

Really wide settings. No restrictions. Zero possibility of over-fitting.
Attached Image(s) (click to enlarge)
Click to Enlarge Name: Image1.png Size: 32 KB
Click to Enlarge Name: Image2.png Size: 19 KB
And you will leave pain to take care of itself.
  • #2,250
  • Feb 3, 2020 8:11am Feb 3, 2020 8:11am
Quoting JoeeF
{quote} The impact of correlation might be huge. It's true. Definitely. Anyway, it's important to find proper approach to solve this issue. Thank you for your suggestions. Take care.
Thanks for your interest, my idea is:
1. Compare all currencies based on 1 common denominator, can use gold (XAUUSD, XAUEUR, XAUGBP, XAUNZD, XAUAUD, XAUCAD, XAUJPY, XAUCHF ...)

2. The formula for calculating the correlation coefficient (CC) can be found here
https://www.statisticshowto.datascie...cient-formula/

3. The work cannot be done here, because when gold increases or decreases sharply, the above pairs CC will increase or decrease, the more gold increase or decrease, the closer the correlation coefficient will be.

For example, in January the price of gold did not change much, the correlation coefficient between XAUAUD and XAUEUR was -0.5, after in February the price of gold increased sharply and although AUD and EUR completely did not change the trend, the correlation coefficient increased. to 0.8 (just example)

As noted above cc is quite approximate and depends on the common denominator increasing or decreasing. My proposition is to eliminate that relativity by comparing CCs.

For example:
CC of XAUAUD and XAUCAD is 0.7
CC of XAUAUD and XAUCHF is 0.71
...
CC of XAUUSD and XAUJPY is 0.98

So 100% of the change is 0.98-0.7 = 0.28
so XAUAUD and XAUCAD (AUDCAD) is 0% correlated
XAUAUD and XAUCHF (AUDCHF) is 3.6% correlated
...
XAUUSD and XAUJPY (USDJPY) is 100% correlated
Now apply the filter (eg <70%) which will filter out USDJPY and pairs above 70% correlation.

Hope to help.
  • #2,251
  • Feb 3, 2020 11:09am Feb 3, 2020 11:09am
Quoting JoeeF
January 2020 - JF4XTrader v1.05 - JFXC variant ... Really wide settings. No restrictions. Zero possibility of over-fitting. {image} {image}
I forgot the equity curve "plot by day" ...
Attached Image (click to enlarge)
Click to Enlarge Name: Image1.png Size: 18 KB
And you will leave pain to take care of itself.
  • #2,252
  • Feb 3, 2020 11:17am Feb 3, 2020 11:17am
Quoting JoeeF
January 2020 - JF4XTrader v1.05 - JFXC variant ... Really wide settings. No restrictions. Zero possibility of over-fitting. {image} {image}
Position sizing has been changed from fixed size (0.1 lot) to fixed fractional (1%) ...
Attached Image (click to enlarge)
Click to Enlarge Name: Image1.png Size: 18 KB

Attached Image (click to enlarge)
Click to Enlarge Name: Image2.png Size: 17 KB

Attached Image

Attached Image

Attached Image (click to enlarge)
Click to Enlarge Name: Image5.png Size: 19 KB

Attached Image (click to enlarge)
Click to Enlarge Name: Image6.png Size: 16 KB
And you will leave pain to take care of itself.
  • #2,253
  • Edited 11:56am Feb 3, 2020 11:43am | Edited 11:56am
Quoting JoeeF
January 2020 - JF4XTrader v1.05 - JFXC variant ... Really wide settings. No restrictions. Zero possibility of over-fitting. {image} {image}
Equity crossover rules have been applied. Position sizing: Fixed size - 0.1 lot ...
Attached Image (click to enlarge)
Click to Enlarge Name: Image1.png Size: 19 KB

Attached Image (click to enlarge)
Click to Enlarge Name: Image2.png Size: 18 KB

Attached Image (click to enlarge)
Click to Enlarge Name: Image3.png Size: 24 KB

Attached Image

Attached Image (click to enlarge)
Click to Enlarge Name: Image5.png Size: 19 KB

Attached Image (click to enlarge)
Click to Enlarge Name: Image6.png Size: 16 KB
And you will leave pain to take care of itself.
  • #2,254
  • Feb 8, 2020 9:45am Feb 8, 2020 9:45am
Quoting JoeeF
January 2020 - JF4XTrader v1.05 - JFXC variant ... Really wide settings. No restrictions. Zero possibility of over-fitting. {image} {image}
Up and down ...
Attached Image(s) (click to enlarge)
Click to Enlarge Name: Image1.png Size: 32 KB
Click to Enlarge Name: Image2.png Size: 18 KB
Click to Enlarge Name: Image3.png Size: 19 KB
Click to Enlarge Name: Image4.png Size: 17 KB
And you will leave pain to take care of itself.
  • #2,255
  • Feb 9, 2020 3:33am Feb 9, 2020 3:33am
Slow week, most of the variations = small loss or break even.
We look forward to next week.
Don't NO-SL and drive! And remember: Always keep your eyes on the trend!
  • #2,256
  • Feb 9, 2020 11:41am Feb 9, 2020 11:41am
JF4XTrader_PM_1: Move the second leg to BE after the first leg has been closed at TP.
JF4XTrader_PM_2: Move both legs to BE after PercentLong/PercentShort of tp1 has been reached.

I'm curious ...
Attached Image (click to enlarge)
Click to Enlarge Name: JF-2020-02-09-001.png Size: 18 KB
And you will leave pain to take care of itself.
  • #2,257
  • Feb 9, 2020 12:55pm Feb 9, 2020 12:55pm
Quoting Ched
{quote} You can try that (this is a work in progress) so anything that helps. You may need to run a few accounts to try different results. If you look at the posts Vicente placed using default settings on a long term demo you can see the account draw down periods. JoeeF has recently posted different configurations giving different results. So for now that would be one way, or cut your lot size back to reduce the account draw down or disable the Ea for a few days to reduce the percentage of account draw down. Once that period is over the gains seem...
Can i ask what settings do you use for the ea? if you dont mind may you please send me the xml file?
  • #2,258
  • Feb 10, 2020 1:43am Feb 10, 2020 1:43am
Quoting sibupholwane
{quote} Can i ask what settings do you use for the ea? if you dont mind may you please send me the xml file?
The best results with cautious settings I have found so far are the same as in post 2,163. You will find the HTML there with the information. But you need to be sure the broker pair suffix is the same. So best look back at post one to add a currency pair. My list may not be the one you want also.
As you can see with JoeeF's secreenshot in post 2,254 another red week. Maybe due to Trumps Impeachment results, the aftermath of Brexit and NFP made things nervous. I disabled the EA early in the week to try to avoid some of the losses and wait for things to come good again. So I will post results later.
  • #2,259
  • Feb 10, 2020 1:56am Feb 10, 2020 1:56am
Quoting Ched
{quote} The best results with cautious settings I have found so far are the same as in post 2,163. You will find the HTML there with the information. But you need to be sure the broker pair suffix is the same. So best look back at post one to add a currency pair. My list may not be the one you want also. As you can see with JoeeF's secreenshot in post 2,254 another red week. Maybe due to Trumps Impeachment results, the aftermath of Brexit and NFP made things nervous. I disabled the EA early in the week to try to avoid some of the losses and wait for things to come good again. So I will post results later.
A nice post ...

Take care.
And you will leave pain to take care of itself.
  • #2,260
  • Feb 10, 2020 2:05am Feb 10, 2020 2:05am
Quoting sibupholwane
{quote} Can i ask what settings do you use for the ea? if you dont mind may you please send me the xml file?
I'm running and testing a lot of EAs. Anyway, the last results are form "wide XC" settings: JF4XTraderX.xml

https://www.forexfactory.com/showthr...1#post12689231

Please, be aware that in most cases you cannot simply take that XML config file without some modifications. You need to check symbol names at least.

Please, read the first post.

Take care.
And you will leave pain to take care of itself.
5 traders viewing now