Mister ED's 4 hour Heikin Ashi system | Page 55 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

EURUSD 15min. Heikin-Ashi System 31 replies

Heikin-Ashi 2 replies

Heikin-Ashi Technique 2 replies

Regarding Smoothed Heikin Ashi indicator 0 replies

Trading with Heikin Ashi 9 replies

Attachments: Mister ED's 4 hour Heikin Ashi system
Exit Attachments

Mister ED's 4 hour Heikin Ashi system

  • #1,081
  • Edited 5:28pm Oct 26, 2009 5:22pm | Edited 5:28pm
Quoting hello
I think I may have found an issue OR maybe a better way to use the RAVI filter. Look at the circles in my attachment, which a buy was triggered and stopped out.

I think we should use the first HAS candle against the RAVI filter? What do you guys think?
Longbar,

Post #1059 I was referring to the HAS candle that RAVI was filter against. As you can see the first blue HAS candle should go against the filter.

For example in my attachment post #1059 you see the circle with the two BUY HAS candles. It seems that the RAVI is being executed on the Second BUY HAS candle instead of the first.

Kind of hard to explain
  • #1,082
  • Oct 26, 2009 5:24pm Oct 26, 2009 5:24pm
Quoting LongBar
Note that when the filter takes place you see this message:
[size=1][color=Blue]2009.10.26 12:55:38...
I thought it was the first HAS We can test the first HAS.
  • #1,083
  • Oct 26, 2009 5:26pm Oct 26, 2009 5:26pm
Quoting LongBar
When I bug-fixed the v1.6.0 to v1.6.1 there was one minor change made to the RAVI FX Fisher indicator evaluation:
Instead of doing this

 

  1. -RAVIFX_Trigger < Indicator reading < +RAVIFX_Trigger

the expression became this

 

  1. -RAVIFX_Trigger <= Indicator reading <= +RAVIFX_Trigger

observe the equal sign.

So if you get different results using the different versions then it is likely down to the RAVIFX_Trigger parameter... of course assuming that you get 90% modelling quality for...

BTW->I believe this option tested better.
-RAVIFX_Trigger < Indicator reading < +RAVIFX_Trigger
  • #1,084
  • Oct 26, 2009 5:36pm Oct 26, 2009 5:36pm
Quoting hello
thanks very much for the explanation. How do you improve model quality? Mines always shows N/A but I have a lot of unmatched data errors.

Edit**Which version did you test with?
Uhh, that almost falls into the magic part

The hard way is if you try to download 1M and other TF data from your broker platform using the chart (scroll back) or scripts. Probably you would go for 1M and then convert that to the higher TF using the period_converter script provided with MT. Best if you delete all your TF data before start this procedure. This would give you very limited test data though as most of the brokers offer you only a few months 1M data to download off the demo/live accounts... but at least you would have "live" data off your broker.

The other hard way of you just download some sort of data via the History Center (see Tools menu). Here you can quickly have several years worth of consistent data - from somewhere, but you have no idea how that compares to your actual broker's data...

The even more harder way if your broker offers TF data for download. You can download those files and import using the History Centre. This would be the best though very limited brokers offer the data...

So once you cleaned up and loaded data into the History Centre then restart MT to store the data to disks and then you can try backtest within the dates you now have data for. Watch out for 90% quality as below that the results completely unreliable, with the 90% modelling the results would still be just an approximate though...
  • #1,085
  • Oct 26, 2009 5:39pm Oct 26, 2009 5:39pm
Read this for 90% modeling quality:

http://alansforexblog.com/2008/02/27...pert-advisors/


I use fxdd; for M1 history go here:

http://www.fxdd.com/en/mt1m-data.html

History data is available from other brokers but I don't have the links...

Jack
  • #1,086
  • Oct 26, 2009 5:44pm Oct 26, 2009 5:44pm
Quoting LongBar
Uhh, that almost falls into the magic part

The hard way is if you try to download 1M and other TF data from your broker platform using the chart (scroll back) or scripts. Probably you would go for 1M and then convert that to the higher TF using the period_converter script provided with MT. Best if you delete all your TF data before start this procedure. This would give you very limited test data though as most of the brokers offer you only a few months 1M data to download off the demo/live accounts... but at least you would have "live"...
"The other hard way of you just download some sort of data via the History Center (see Tools menu). Here you can quickly have several years worth of consistent data - from somewhere, but you have no idea how that compares to your actual broker's data."

I've done this before but still no positive results. I will try to delete data and download again.

Can you have RAVI filter go against the first HAS bar and keep the filter syntax as this -RAVIFX_Trigger < Indicator reading < +RAVIFX_Trigger.

This seems to make a big difference to me. What do you think?
  • #1,087
  • Oct 26, 2009 5:45pm Oct 26, 2009 5:45pm
Quoting jack h
Read this for 90% modeling quality:

http://alansforexblog.com/2008/02/27...pert-advisors/


I use fxdd; for M1 history go here:

http://www.fxdd.com/en/mt1m-data.html

History data is available from other brokers but I don't have the links...

Jack
Thanks Jack! I will test this out
  • #1,088
  • Oct 26, 2009 6:12pm Oct 26, 2009 6:12pm
Quoting jack h
Read this for 90% modeling quality:

http://alansforexblog.com/2008/02/27...pert-advisors/


I use fxdd; for M1 history go here:

http://www.fxdd.com/en/mt1m-data.html

History data is available from other brokers but I don't have the links...

Jack
Ah hah! got 90% with FXDD

Thanks for the help.
  • #1,089
  • Oct 26, 2009 6:19pm Oct 26, 2009 6:19pm
Quoting hello
Ah hah! got 90% with FXDD

Thanks for the help.
Your welcome!
  • #1,090
  • Oct 26, 2009 8:02pm Oct 26, 2009 8:02pm
Quoting hello
Thanks Jack! I will test this out
Attached are the statements for the GBPJPY h4 with 90% model quality. Still 1.6 outperforms 1.6.1by more than 100% profit.
Attached File(s)
File Type: zip testa.zip   36 KB | 357 downloads
  • #1,091
  • Oct 26, 2009 9:30pm Oct 26, 2009 9:30pm
Long Bar,

I have one request if possible. Can you add some option for break even positions. For example if both positions are in profit for an X amount of pips, the stops would move to break even on both (X pips beyond entry point). My point is; if a signal is generated the momentum should be good enough for the next bar to move in our direction.

For example : Entry Long 1.2000

Minutes Later price reach 1.2025 (+25pips)

Stops would be placed at 1.2005 or whatever.

This way we could lock at least some profits.

I don't not if I made clear myself. My English is not very good.
  • #1,092
  • Edited 11:17pm Oct 26, 2009 10:18pm | Edited 11:17pm
Quoting Forexaccount
Long Bar,...
Good Idea! Like a trailing stop after a certain amount of pips is reached. We already use TS, but only after TP1 has been reached. I like your suggestion as most of the time the market moves in our favor, but TP is not reached.

Something like this would also work.

For Example: exit option 4: Custom TP && TS at HAS(both positions)?

However there is exit method 1 "1:Trailing SL on HAS bar" TS the HAS bars or wicks. I still like your suggestion better
  • #1,093
  • Oct 27, 2009 4:47am Oct 27, 2009 4:47am
yep!

2 winners!
Attached Image (click to enlarge)
Click to Enlarge Name: 2.gif Size: 23 KB
Attached Image
  • #1,094
  • Oct 27, 2009 6:59am Oct 27, 2009 6:59am
Be aware that if you use TP=SL=ATR and you exit before the TP is reached you are then altering your R/R and one stop loss can wipe away many winnings!

Quoting hello
Good Idea! Like a trailing stop after a certain amount of pips is reached. We already use TS, but only after TP1 has been reached. I like your suggestion as most of the time the market moves in our favor, but TP is not reached.

Something like this would also work.

For Example: exit option 4: Custom TP && TS at HAS(both positions)?

However there is exit method 1 "1:Trailing SL on HAS bar" TS the HAS bars or wicks. I still like your suggestion better
  • #1,095
  • Oct 27, 2009 7:52am Oct 27, 2009 7:52am
Quoting bali2002
Be aware that if you use TP=SL=ATR and you exit before the TP is reached you are then altering your R/R and one stop loss can wipe away many winnings!
Thanks Bali.


Which ea are you trading?

Thank you
  • #1,096
  • Oct 27, 2009 10:00am Oct 27, 2009 10:00am
Longbar,

Sorry for the time it took to respond to your post #977. I have been out of touch the past ~2 weeks. I am just catching up.

These are the news filters I am currently using with other EA's and they work well. I placed it on one pair and it worked so please continue on when you get time.

Looks like the thread took a Ravi turn so I have more catching to do.

Thanks for all the work I look forward to paper trading 1.6.1 in the next few days.

Alsea
  • #1,097
  • Oct 27, 2009 11:07am Oct 27, 2009 11:07am
Quoting hello
Thanks Bali.


Which ea are you trading?

Thank you
I'm trading manually for now, with good results.
  • #1,098
  • Oct 27, 2009 1:55pm Oct 27, 2009 1:55pm
LongBar,

I just started using the EA live but I just noticed something. The Exit option # 3 is not really a Break even option. The stop is moved where the 1st order was closed giving very little space for price movement. I thought that break even was to move the stop to the entry position. Let me know if you get what I am saying. The overall EA is just perfect.

Thanks
  • #1,099
  • Oct 27, 2009 4:20pm Oct 27, 2009 4:20pm
Quoting hello
Attached are the statements for the GBPJPY h4 with 90% model quality. Still 1.6 outperforms 1.6.1by more than 100% profit.
hello,

Seems to me what you are really after is to aply the RAVI indicator to current (last) bar, even though the strategy checks the closed bar (second last). I am goign to add an another EA parameter so you can select which bar you want to read the indicator off - this is the main difference between ver 1.6.0 and 1.6.1.

Note that the indicator at the current bar will repaint itself so any decision the EA might make could make no sense later once the current bar (price) progresses over time. The only way of decision validation is to review the EA logs to see what was the indi reading at that point in time.

Thanks for sharing out your test results!

Longbar
  • #1,100
  • Oct 27, 2009 4:46pm Oct 27, 2009 4:46pm
Quoting Forexaccount
LongBar,

I just started using the EA live but I just noticed something. The Exit option # 3 is not really a Break even option. The stop is moved where the 1st order was closed giving very little space for price movement. I thought that break even was to move the stop to the entry position. Let me know if you get what I am saying. The overall EA is just perfect.

Thanks
Had a look at the code and I see no obvious bug there. The following logic is applied (well at least the intention was):

 

  1. When 1st order closes move 2nd order's SL to the 2nd order's entry point (break even)
  2. However if trailing on HAS bar would move the SL beyond the break even point (ie into profit) then use that as SL

This is because ExitMethod 3 applies BE & HAS trailing logic.

Would you double check that your experienced SL movements follow the above logic?

LB

0 traders viewing now