Menu
Sister Sites

Options

Bookmark Thread

First Page First Unread Last Page Last Post

Print Thread

Similar Threads

Calendar as Fundamental Source at Forex Factory 14 replies

Forex Factory Store 0 replies

Journal Factory posters - how much money are you making? 17 replies

Newbie Factory Journal 12 replies

Attachments: EA Factory
Exit Attachments

EA Factory

I will try it.
Quoting aha
Hi guys,

Orange PM me about his system here:

http://www.forexfactory.com/forexfor...61&postcount=1

I think it's really smart and interesting. Does anyone get the time to make an EA for it? It should be less than 30 minutes.

Here is what I think about the input parameters:

Int EntryOffset: the entry point away from the high/low of previous day
Int HardSL: the initial stop loss
Bool Trailing
Int TrailingSL:

That's it. Any comments are welcomed.

aha
Done in 10 minutes.
http://www.forexfactory.com/forexfor...chmentid=13335
But I don't think this simple method would work.
Quoting GP2X
I will try it.
Thanks GP2X! How was the backtesting result?

Quoting GP2X
Done in 10 minutes.
http://www.forexfactory.com/forexfor...chmentid=13335
But I don't think this simple method would work.
Target: Trading naked.
Hi Aha,

Thank you for your TD-Seq and TD-Combo indicators. I'm trying them for 5-min charts.

Here were something I found last night.

1. Both indicators didn't update screen/numbering automatically so I had to manually open them in indicator lists every 5 minutes. This is troublesome for 5-min trading. See 1st attached.

2. They didn't update last completed bar which is on left of current bar, even with manual updates. You can see from 1st attached. No problem for 1-hour chart.

3. I'm not sure but I think TD-Combo updated 1-hour chart automatically but not for TD-Seq.
One small numbering glitch was one of them (or both?) skipped number 4 when updated automatically. I can't remember this was for setup or countdown. For example, it showed 1,2,3,5,6 and turned to 1,2,3,4,5 after I updated it manually.

4. According to TD Countertrend pdf, perfect sell setup is high of sell setup bar #8, #9 or subsequent one has to be higher than highs of both bar #6 and #7. This is not in your indicators. See 2nd attached.

5. According to same pdf, validated sell countdown is high of bar #13 must be higher than or equal to close of bar #8. See 3rd attached.

I plan to do a 1 to 2 weeks' manual backtest for your indicators and two other coders' during this week.

Thanks again.

Regards,
Hoh
Attached Images
One more thing:

Both TD-Seq and TD-Combo have had same setup numbering but different countdown numbering when they are overlapping?

Hoh
Attached Image
Aha,

Thanks to you and the team behind these indicators. I have been following this thread on these indicators and have been trying it on FXDD demo for the last 2 days. I have the same problem as Hoh. Both these indicators do not automatically place numbers on the chart.

Would appreciate your help and guidance. Thanks.

Payal

Quoting hoh
Hi Aha,

Thank you for your TD-Seq and TD-Combo indicators. I'm trying them for 5-min charts.

Here were something I found last night.

1. Both indicators didn't update screen/numbering automatically so I had to manually open them in indicator lists every 5 minutes. This is troublesome for 5-min trading. See 1st attached.

2. They didn't update last completed bar which is on left of current bar, even with manual updates. You can see from 1st attached. No problem for 1-hour chart.

3. I'm not sure but I think TD-Combo updated 1-hour chart automatically but not for TD-Seq.
One small numbering glitch was one of them (or both?) skipped number 4 when updated automatically. I can't remember this was for setup or countdown. For example, it showed 1,2,3,5,6 and turned to 1,2,3,4,5 after I updated it manually.

4. According to TD Countertrend pdf, perfect sell setup is high of sell setup bar #8, #9 or subsequent one has to be higher than highs of both bar #6 and #7. This is not in your indicators. See 2nd attached.

5. According to same pdf, validated sell countdown is high of bar #13 must be higher than or equal to close of bar #8. See 3rd attached.

I plan to do a 1 to 2 weeks' manual backtest for your indicators and two other coders' during this week.

Thanks again.

Regards,
Hoh
Not good, I have to raise the stoploss to 200-300 to tune it to profitable.
Quoting aha
Thanks GP2X! How was the backtesting result?
Hi Hoh,

Thank you for sharing your testing result.

Both indicators wasn't intended to be used at a timeframe lower than 30M, so I did not integrate the logic for that. But if your manual testing result is good on low timeframe (I doubt it), I will add it in.

If you find a problem on updating, you may try to switch timeframe and see what happens, i.e., change timeframe from one timeframe to another and then change it back. In that way you at least don't need to reopen the indicator.

I am aware of the updating problem with TD-Seq and in TD-Combo the problem has been fixed (I think.. ).

The perfect setup/countdown are indicated by larger arrows. The arrows in your charts were small ones, which means the setup/countdown weren't perfect.

aha


Quoting hoh
Hi Aha,

Thank you for your TD-Seq and TD-Combo indicators. I'm trying them for 5-min charts.

Here were something I found last night.

1. Both indicators didn't update screen/numbering automatically so I had to manually open them in indicator lists every 5 minutes. This is troublesome for 5-min trading. See 1st attached.

2. They didn't update last completed bar which is on left of current bar, even with manual updates. You can see from 1st attached. No problem for 1-hour chart.

3. I'm not sure but I think TD-Combo updated 1-hour chart automatically but not for TD-Seq.
One small numbering glitch was one of them (or both?) skipped number 4 when updated automatically. I can't remember this was for setup or countdown. For example, it showed 1,2,3,5,6 and turned to 1,2,3,4,5 after I updated it manually.

4. According to TD Countertrend pdf, perfect sell setup is high of sell setup bar #8, #9 or subsequent one has to be higher than highs of both bar #6 and #7. This is not in your indicators. See 2nd attached.

5. According to same pdf, validated sell countdown is high of bar #13 must be higher than or equal to close of bar #8. See 3rd attached.

I plan to do a 1 to 2 weeks' manual backtest for your indicators and two other coders' during this week.

Thanks again.

Regards,
Hoh
Target: Trading naked.
Yeah, SEQ and Combo have the same setup logic but different countdown logic.

Both countdown shown in your charts are for a previous setup, not the ones shown in the charts.


Quoting hoh
One more thing:

Both TD-Seq and TD-Combo have had same setup numbering but different countdown numbering when they are overlapping?

Hoh
Target: Trading naked.
Are you saying both of them? I am using TD-Combo on 12 charts and they seems to be fine. I know there is a glitch in TD-Seq on updating numbering when I was developing TD-Combo on the top of it. I will fix it when I get the time.



Quoting payal
Aha,

Thanks to you and the team behind these indicators. I have been following this thread on these indicators and have been trying it on FXDD demo for the last 2 days. I have the same problem as Hoh. Both these indicators do not automatically place numbers on the chart.

Would appreciate your help and guidance. Thanks.

Payal
Target: Trading naked.
Quoting aha
Thanks GP2X! How was the backtesting result?
i took a look at td carrie using VB on excel.
i had to change the original TD carrie method as it uses a percentage of the stock prices for buy\sellstops, stoploss levels and take profit levels.

I used a lowest historical low as my "zero" level and subtracted this from the current open of the daily bar. I used this difference in calculating my 1% buy\sellstops, 4% stoploss levels and 12% take profit levels. (Using the absolute value of the currency would make the stop loss, and take profit and buy\sellstops extremely large).

for most of the settings i tried, the method was not profitable.
however there were some settings which broke even or were profitable.
it was all dependent on the stoploss and take profit levels that were set.
(for me this really drove the point home about money management and how a system can be profitable or unprofitable from that alone).

i did most of the backtesting using a VB macro on excel. i put in a dialog box for you guys to make it easier to use. forgive the crudeness of the dialog box--it was my first time using a dialog box.


attached is the excel spreadsheet with EURUSD daily data from alpari.
scroll to the bottom for Net profits, largest profit for a given trade, and largest loss.

if you want to try different stoploss, takeprofit and buy\sell stop percentages. press alt+f8. select the "show user form" macro. run it.
a dialog box appears and you can enter whatever percentages you want.

I think it's time I begin learning MQL4 in earnest.
I'll take on the coding for TD REI. After reading it in demark's book, i think it will be a great indicator to use to help determine entry when TD combo or TD sequential signifies there will be a reversal.

note: demark extensively uses price action (albeit his versions of them) to determine entry for many of his indicators and trendline breaks. something as simple as price action and money management in conjunction with TD Combo\Seq can make a very profitable system

phoenix

p.s. i had to change the extension to .jpg as the forum would not allow me to upload a ".xls" file. change the extension back to .xls to open the excel file.
Yes. I tried both of them on H1 and M5 charts for GBP/USD. I have to refresh in order to get the numbering working and when I do it the second time, it gets changed. Past data's numbering seems not affected. It is only the current days value gets changed.

BTW, I am using the 0_5 for reversal and 0_1 for the combo. I guess these are the most recent versions, right?

Let me know. Thanks.


Quoting aha
Are you saying both of them? I am using TD-Combo on 12 charts and they seems to be fine. I know there is a glitch in TD-Seq on updating numbering when I was developing TD-Combo on the top of it. I will fix it when I get the time.
Quoting aha
Both indicators wasn't intended to be used at a timeframe lower than 30M, so I did not integrate the logic for that. But if your manual testing result is good on low timeframe (I doubt it), I will add it in.
Thanks for your reply.

I'd done a quick manual backtest for EUR/JPY for September and October 2006 by using your TD-Combo with Left Brain Trading's Absolute Fibo (LBT) of 56 this morning. My target was 2 LBT levels which was about 8pipsx2=16pips. I entered 2 lots:1st lot for 8pips and 2nd lot for 16pips. Below was result from my memory because I didn't have time to do complete backtest:

1.Winning rate was around 95% for stop lost of 100pips.
2.Winning rate was 95%-99%(?) for stop lost of 200pips.

Only one case of drawback which was more than 100pips i.e. 185pips for these two months. This only case hitted its original target of 16pips after 15 days. So this was theoritically winning trade but practically losing trade.

I also did a quick TD-Combo along with LBT Ab Fibo 146 for 1-hour chart. The winning rate was very close to that of 5-min except all target and stop lost were 3-fold larger than 5-min. For example, target was 50-70pips.

My other find out was there were 2-5 TD-Combo trades per day for 5-min chart. For 1-hour chart, there was 1 trade for every 1-7 days. Time to close a (winning) trade were 15 minutes to 8 hours for M5 and 4 hours to 4 days for H1.

So my conclusion was risk/reward and profit for M5 and H1 were basically the same if taking account of trade opportunity and completion time.

My main purpose of trying M5 was because my capital is not large enough to hold drawback for half to few days as in H1 trades. Also, I can put my capital into 3 different pairs at the same time (if they have valid setups) in M5 and take profit after few hours for next trades.

Last thing is I can gain more experiences by trading M5 instead of H1 during same period.

So I'll be very pleased if you can make automatic update for shorter term. Maybe a setting input which can be changed by user for updating every x minutes?

I plan to test 1-min chart for last few days because this morning I saw similar price movement for M1 and M5 but there were about 10-20 valid trades per day for M1

Quote
If you find a problem on updating, you may try to switch timeframe and see what happens, i.e., change timeframe from one timeframe to another and then change it back. In that way you at least don't need to reopen the indicator.
Thanks and I'd tried it just now but the time to switch over different TFs was similar to update through indicator list. I even changed number of bar for history and chart to 5K in option.

Quote
I am aware of the updating problem with TD-Seq and in TD-Combo the problem has been fixed (I think.. ).
Can TD-Seq be fixed?

Quote
The perfect setup/countdown are indicated by larger arrows. The arrows in your charts were small ones, which means the setup/countdown weren't perfect.
Now I can see the different arrow size...
I think this arrow size is not obvious. Can you change to other arrow type for perfect setup/countdown?

Sorry for long post and thanks for your great indicators.

Regards,
Hoh
Quoting hoh
...
1.Winning rate was around 95% for stop lost of 100pips.
2.Winning rate was 95%-99%(?) for stop lost of 200pips
Aha,

I found answer for my own result of almost 100% winning rate (too good to be true?):

I monitored 1-min chart real-time this afternoon and found that TD-Combo (and also TD-Seq) erased incomplete countdown whenever there was opposite setup was formed or for other reasons.

Now I explain why this real-time repaint of indicator effected my backtest result (this is normal for all real-time repainting indicators):

In my backtest for sept. and oct., I checked every complete setup which was not accompanied by countdown and made counter-trend trade. TD-Combo did well for my 8-24pips target and I got almost 100% winning rate.

Now for those complete setups which were accompanied by complete countdown, the first setup acted like trend maker which led to second to forth setups in same direction and formed a complete countdown finally. So more than 70% of my counter-trend trades for these setups failed.

Since most of the time TD-Combo creates setup along with countdown, incomplete countdown confused me in real-time because it may or may not be completed later.

This confusion did not happen in my backtest because all incomplete countdown had already been erased.

I hope you can get me.

Can you please add a setting input which lets user to choose whether incomplete countdown is shown. If can, also shows a "x" mark at the bar which TD-Combo decides to cancel incomplete countdown so that I can decide what next step to be taken during backtest and forward test.

Regards,
Hoh
Attached Image
Thanks for the hard work, phoenix. Lots of methods seems to be reasonable but once test it overtime, it's not profitable.



Quoting phoenix
i took a look at td carrie using VB on excel.
i had to change the original TD carrie method as it uses a percentage of the stock prices for buy\sellstops, stoploss levels and take profit levels.

I used a lowest historical low as my "zero" level and subtracted this from the current open of the daily bar. I used this difference in calculating my 1% buy\sellstops, 4% stoploss levels and 12% take profit levels. (Using the absolute value of the currency would make the stop loss, and take profit and buy\sellstops extremely large).

for most of the settings i tried, the method was not profitable.
however there were some settings which broke even or were profitable.
it was all dependent on the stoploss and take profit levels that were set.
(for me this really drove the point home about money management and how a system can be profitable or unprofitable from that alone).

i did most of the backtesting using a VB macro on excel. i put in a dialog box for you guys to make it easier to use. forgive the crudeness of the dialog box--it was my first time using a dialog box.


attached is the excel spreadsheet with EURUSD daily data from alpari.
scroll to the bottom for Net profits, largest profit for a given trade, and largest loss.

if you want to try different stoploss, takeprofit and buy\sell stop percentages. press alt+f8. select the "show user form" macro. run it.
a dialog box appears and you can enter whatever percentages you want.

I think it's time I begin learning MQL4 in earnest.
I'll take on the coding for TD REI. After reading it in demark's book, i think it will be a great indicator to use to help determine entry when TD combo or TD sequential signifies there will be a reversal.

note: demark extensively uses price action (albeit his versions of them) to determine entry for many of his indicators and trendline breaks. something as simple as price action and money management in conjunction with TD Combo\Seq can make a very profitable system

phoenix

p.s. i had to change the extension to .jpg as the forum would not allow me to upload a ".xls" file. change the extension back to .xls to open the excel file.
Target: Trading naked.
i found 2 versions for TDREI. But i did not test it or verify the code to comply with Demark instructions
Attached File(s)
File Type: mq4 Phat_TD_REI.mq4   2 KB | 562 downloads
File Type: mq4 REIMT4.mq4   3 KB | 578 downloads
The trend isyour friend untilis about to end...
Man, 100 pips SL and 8-24 pips profit will get you no where. That means you will have to win about 4-11 times more than loss to break even. I would maintain a healthy 1:2-3 risk return ratio.

What I am getting at is that TD-Combo is more suitable for mid-term ranging market, where trends are strong enough to bring in nice profit but not too strong to break your setup a couple of times. Like right now GBP is in a trend and TD-Combo has signaled reversal but BoE is expected to raise interest rate once more in Nov so the trend may continue. In this kind of situation, are you going to position yourself in the other side of the trend?

About the confusion of the repainting, you may turn the numbering off in the parameter list. You will only see arrows on the charts and those arrows only appear when the setup and countdown complete.




Quoting hoh
Aha,

I found answer for my own result of almost 100% winning rate (too good to be true?):

I monitored 1-min chart real-time this afternoon and found that TD-Combo (and also TD-Seq) erased incomplete countdown whenever there was opposite setup was formed or for other reasons.

Now I explain why this real-time repaint of indicator effected my backtest result (this is normal for all real-time repainting indicators):

In my backtest for sept. and oct., I checked every complete setup which was not accompanied by countdown and made counter-trend trade. TD-Combo did well for my 8-24pips target and I got almost 100% winning rate.

Now for those complete setups which were accompanied by complete countdown, the first setup acted like trend maker which led to second to forth setups in same direction and formed a complete countdown finally. So more than 70% of my counter-trend trades for these setups failed.

Since most of the time TD-Combo creates setup along with countdown, incomplete countdown confused me in real-time because it may or may not be completed later.

This confusion did not happen in my backtest because all incomplete countdown had already been erased.

I hope you can get me.

Can you please add a setting input which lets user to choose whether incomplete countdown is shown. If can, also shows a "x" mark at the bar which TD-Combo decides to cancel incomplete countdown so that I can decide what next step to be taken during backtest and forward test.

Regards,
Hoh
Target: Trading naked.
Quoting hoh
Thanks for your reply.

I'd done a quick manual backtest for EUR/JPY for September and October 2006 by using your TD-Combo with Left Brain Trading's Absolute Fibo (LBT) of 56 this morning. My target was 2 LBT levels which was about 8pipsx2=16pips. I entered 2 lots:1st lot for 8pips and 2nd lot for 16pips. Below was result from my memory because I didn't have time to do complete backtest:

1.Winning rate was around 95% for stop lost of 100pips.
2.Winning rate was 95%-99%(?) for stop lost of 200pips.

AHA: SL is too big for your TP target... which basically says that we can't take every signal we get.

So I'll be very pleased if you can make automatic update for shorter term. Maybe a setting input which can be changed by user for updating every x minutes?

Can TD-Seq be fixed?

Now I can see the different arrow size...
I think this arrow size is not obvious. Can you change to other arrow type for perfect setup/countdown?

AHA: Unfortunately I don't really have the time to fix it right now. But for TD-SEQ, you may try the version from GP2X, Clam and Hans. They may have the functions you want.



Regards,
Hoh
Thank you for sharing your testing results, Hoh.
Target: Trading naked.
AHA.... I have read some of your posts and like where you are coming from....

Over in the daily range thread myself and a few other traders have come up with 4 different variations of a daily range strategy which we would like backtested over a period of possible years. we would like to get a summary month on month of hwat each system achieves pip wise and then obviously what the drawdown is.

The strategies are very easy to understand and are completely mechanical in nature.

I have written detailed specs for all of these and am just finalising the finer details.

Would it be possible for you guys here to back test it for us and then possible write an EA to go with it?

I Guarentee you that these variation are all profitable in the short term but we need to know about the long term and then have to be able to automate it. This is something that everyone will benefit from :-)

Please let me know if this is possible and I will post the specs.
Micardo,

I am very interested in the system you guys are developing and I don't think there is a problem to implement the system in an EA (or EAs). But before we start coding, can you guys provide a manual backtesting result of about two monthes? In that way we can have an idea of the profitability, and in the same time it may help you improving the maturity of the logic behind the system.

You may post your specs here too, some of the testers here may be interested in manual testing them.

aha


Quoting Micardo
AHA.... I have read some of your posts and like where you are coming from....

Over in the daily range thread myself and a few other traders have come up with 4 different variations of a daily range strategy which we would like backtested over a period of possible years. we would like to get a summary month on month of hwat each system achieves pip wise and then obviously what the drawdown is.

The strategies are very easy to understand and are completely mechanical in nature.

I have written detailed specs for all of these and am just finalising the finer details.

Would it be possible for you guys here to back test it for us and then possible write an EA to go with it?

I Guarentee you that these variation are all profitable in the short term but we need to know about the long term and then have to be able to automate it. This is something that everyone will benefit from :-)

Please let me know if this is possible and I will post the specs.
Target: Trading naked.
0 traders viewing now