Mister ED's 4 hour Heikin Ashi system | Page 36 | 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

  • #701
  • Aug 19, 2009 4:23pm Aug 19, 2009 4:23pm
Mr Ed & LongBar,

Thank you for all of the work you have put into this system!

Just a suggestion for listing what trade days are valid. In post #389 (an indicator to show the system visually before an EA was written), I used the follow code to list what days to trade:

extern string desc = "enter 5 y or n letters";
extern string Trade_GBPJPY = "yyyyy";
extern string Trade_GBPUSD = "yynnn";
extern string Trade_EURJPY = "ynyyy";
extern string Trade_EURUSD = "yynyy";
extern string Trade_USDCAD = "ynnny";
extern string Trade_EURGBP = "yyyyy";
extern string Trade_AUDUSD = "ynnnn";
extern string Trade_AUDJPY = "yyyyy";
extern string Trade_NZDUSD = "ynnny";
extern string Trade_USDJPY = "yyyyy";


This might be a way to wrap up all of the trade days at one time.

Keep up the great work!

Mark
  • #702
  • Aug 20, 2009 5:03am Aug 20, 2009 5:03am
Quoting Niktesla
Hello. i'm just finished the quick overview of the thread.
Mr. ED, Mr. LongBar - Thanks for your hard work !

I haven't yet tried the EA, but I'd like to propose a feature to implement.
Instead of one string parameter with the day listed, you could make ~7 integer parameters which takes binary values 0 or 1.
Should look somehow like ~
TradeOnMonday 1
TradeOnTuesday 1
TradeOnWednesday 0
ect.

The point is - with binary variables like that you can make optimizations with mt4 tester - on which days you should trade and on which you should not.
I thought that the approach would be to use the backtest to trade all opportunity, ie all days and candle formations. The EA generates statistical files which would provide details for the user to see which day/candle setups are the most profitable ones and use those on foward/live trading with the corresponding settings.

This means no need to backtest one-by-one the day/candle setups and the single run would generate all the stats. Considering some level of cross-impact of trades this may not be perfect and may involve some additional manual runs to validate results though.

So to everyone: if you want the above parameter changes to be done please provide your feedback!
  • #703
  • Aug 20, 2009 5:06am Aug 20, 2009 5:06am
Quoting beepblip
Hi LongBar... any thoughts about this? I really believe in this sort of swing trading strategy. It should perform better than real-life on a back test, but it is failing to show a profit. Is there additional information you need about my backtest runs? Should I run on a different broker?

Cheers,

beepblip
Hi beepblip,

I would suggest to consider Mister ED's other optimisation suggestions described in his document. Mr ED suggests a dynamic system considering which days and candles are to be traded for a particular pair. Those may improve your results...

LB
  • #704
  • Aug 20, 2009 5:08am Aug 20, 2009 5:08am
Quoting mt2414
i join in the chorus of thanks to Mr. Ed and Longbar for their generosity and hard work!

with no EA experience, i'm at a loss for trying to get "MisterED_v0.13.0" to take auto-trades....
Could you post a screenshot of the chart the EA is running on?
  • #705
  • Aug 20, 2009 8:23am Aug 20, 2009 8:23am
Quoting LongBar
I thought that the approach would be to use the backtest to trade all opportunity, ie all days and candle formations. The EA generates statistical files which would provide details for the user to see which day/candle setups are the most profitable ones and use those on foward/live trading with the corresponding settings.
I think I really missed this file with stat-info. I definitely need to look at it properly
Of course, if there's such an implementation already exist, then we just don't need the workaround I mentioned

I really hope we will see some backtest results & stats soon.
  • #706
  • Aug 20, 2009 10:34am Aug 20, 2009 10:34am
Quoting LongBar
Hi beepblip,

I would suggest to consider Mister ED's other optimisation suggestions described in his document. Mr ED suggests a dynamic system considering which days and candles are to be traded for a particular pair. Those may improve your results...

LB

That's bang on longbar. People ignore the optimization part, I think that's crucial and could be the difference between making pips and losing pips. Mister Ed does this by analyzing all the pairs and gold every day for a month to see their results. At the end of the month he analyzes which days he was profitable and which days he weren't.
  • #707
  • Aug 20, 2009 10:45am Aug 20, 2009 10:45am
Quoting LongBar
So to everyone: if you want the above parameter changes to be done please provide your feedback!

Hello LongBar,

Thank you for donating your skills and dedication to this project - the progress has been amazing.

You ask for feedback on changes. While I think that Mark has a very valid point, I feel that the priority should be to first include all essential components of Mr Ed's trading - because he is making money with this system.

Thus I would suggest as the next step the retracements.

Thanks again for your lead in this.
  • #708
  • Edited 11:15am Aug 20, 2009 10:56am | Edited 11:15am
Hello Longbar,

I agree with Bubo. Mr Ed has a very good system here and if it can be automated to the way he trades it then it will benefit everyone. Like everyone else, I appreciate your working on the ea and helping people.

Regards,
Jeff


Quoting bubo
Hello LongBar,

Thank you for donating your skills and dedication to this project - the progress has been amazing.

You ask for feedback on changes. While I think that Mark has a very valid point, I feel that the priority should be to first include all essential components of Mr Ed's trading - because he is making money with this system.

Thus I would suggest as the next step the retracements.

Thanks again for your lead in this.
  • #709
  • Aug 20, 2009 11:10am Aug 20, 2009 11:10am
Quoting LongBar
I thought that the approach would be to use the backtest to trade all opportunity, ie all days and candle formations. The EA generates statistical files which would provide details for the user to see which day/candle setups are the most profitable ones and use those on foward/live trading with the corresponding settings.

This means no need to backtest one-by-one the day/candle setups and the single run would generate all the stats. Considering some level of cross-impact of trades this may not be perfect and may involve some additional manual...

Please leave as is.
It covers the backtest well and is clearer to read when setting the EA.

JR
  • #710
  • Aug 20, 2009 12:00pm Aug 20, 2009 12:00pm
Quoting LongBar
All,

This is a major revamp of the EA to use 2 orders per trigger when trailing is configured. The trading logic and ticket handling have been changed big time, also full support for ECN type brokers should be present now, ie recovery of any kind of trading errors now added in.

Due to these major changes please be cautious and use this EA on demo account only until bugs discovered and fixed

Changes:[list][*]The TP1CheckPoint parameter has been removed: TP1 is checked per each incoming tick[*]Two orders are submitted to the account when...
Hi Longbar,

V0.14 acted like the Energizer bunny last night, opening 337 trades, closing each one in the same minute it opened, despite the price being nowhere near the SL/TP. It also opened as many as 20+ trades for a pair in the same minute. Drained a 23k FXCM demo account to 2k in less than 3 hours.

The strange part is it only happened with the EA set for a fixed 1 lot, and exit 2. The other five V0.14 demo combinations (fixed, exits 0,1 and %, exits 0,1,2) appear to be functioning fine. Only this combo produced the hypertrading. All the trades appear to have been closed right after they were opened, resulting in a loss on each one due to the spread. 4 of 337 trades were profitable and I assume it was due to a lucky price spike as the order was opened.

In the MT4 report, the Comment field showed a few entries I've never seen: "so: 99.7%/1993.0/2000.0" in place of the usual "133.057 MisterED" type entries.

I'm going to download a fresh MT4 and start over in case it's just a corrupted MT4 file. Will keep you posted.

Let me know if you want the .set file, MT4 report or journals.

JR
  • #711
  • Aug 20, 2009 2:06pm Aug 20, 2009 2:06pm
Quoting JRpips8
Hi Longbar,

V0.14 acted like the Energizer bunny last night, opening 337 trades, closing each one in the same minute it opened, despite the price being nowhere near the SL/TP. It also opened as many as 20+ trades for a pair in the same minute. Drained a 23k FXCM demo account to 2k in less than 3 hours.

The strange part is it only happened with the EA set for a fixed 1 lot, and exit 2. The other five V0.14 demo combinations (fixed, exits 0,1 and %, exits 0,1,2) appear to be functioning fine. Only this combo produced the hypertrading....
Hi JR,

I've seen the corrupted comment field earlier when two or more EA instances were attached to same pair on different charts, or when multiple MT instances were started off simultaneously off the same MT installation, hosting multiple EA instances on the same chart at the same time. This latter one can "easily" happen on hosted windows environments.

Could you check out if any of the above was applicable to your case?

Cheers
longbar
  • #712
  • Aug 20, 2009 2:47pm Aug 20, 2009 2:47pm
Quoting LongBar
Hi JR,

I've seen the corrupted comment field earlier when two or more EA instances were attached to same pair on different charts, or when multiple MT instances were started off simultaneously off the same MT installation, hosting multiple EA instances on the same chart at the same time. This latter one can "easily" happen on hosted windows environments.

Could you check out if any of the above was applicable to your case?

Cheers
longbar
Hi,

Mine is running locally on a single PC, multiple monitors, with 2 other PC's on the domain running different MT4 accounts. No VPS or other hosted account.

Hmmmm, there was an extra chart of EURCAD on the profile, with the EA enabled, hiding off to the side. Too many charts, too few monitors. It was an H1 chart with the EA set to H4, like the other EURCAD chart EA.

Most of the trades were on other pairs, not eurcad, but I assume that's what did it. None of the other 5 demo accounts have duplicate charts and they appear to be fine. Good to know not to have duplicate pair/EA chart.

Thanks for your help. Glad it wasn't a bug!

Cheers,

JR
  • #713
  • Aug 20, 2009 3:12pm Aug 20, 2009 3:12pm
Mister Ed, Could you check your history and see if the last few days set a new record for the biggest drawdown. If not then I'm sure its close!!!

Thanks
  • #714
  • Aug 20, 2009 5:09pm Aug 20, 2009 5:09pm
The leading change this time is the addition of retracement trading trigger. This happens when trade gets triggered by HA colour change but the signal candle vs ATR ratio is over the 200% limit. In this case the EA will wait until a certain retracement level and then submits the pending orders in the direction of the trading signal.

RetraceTriggerPct defines the retracement percentage level triggering the pending orders to the account at the top/bottom of the signal candle. The pending orders will expire just before the close of the current bar in the case those would not get fired.

Also there is a new parameter ATRMultiplierTP to scale up/down target profit level based on the ATR value. The existing ATRMultiplier parameter has been renamed to ATRMultiplierSL just to keep them in pair and clarity.

Again: please test test and test first before consider the EA for real money

Longbar
Attached File(s)
File Type: mq4 MisterED_v0.15.0.mq4   44 KB | 488 downloads
  • #715
  • Aug 20, 2009 9:42pm Aug 20, 2009 9:42pm
Quoting Pipquest
Mister Ed, Could you check your history and see if the last few days set a new record for the biggest drawdown. If not then I'm sure its close!!!

Thanks
Hi Pipquest

The unfortunate truth is that you are right. This month has been the worst in 2 years for the system. The market is giving trades and then reversing and closing things out. There have been many large candles that have given set ups.

I think that this is likely due to a general lack of direction and volume that occurs during the August holiday.

My gut feeling is that it may be best to sit out the rest of the month or at least wait until some direction is being shown. The market is too choppy at the moment.

My stats using optimised trades shows I am 132 pips down so far this month. Not a disaster but could do better.

Please note that I have not traded live this month as I have other things on my plate at the moment. I intend to return to live trading in September once some normality returns.

All the best,

Mister ED
  • #716
  • Aug 21, 2009 11:20am Aug 21, 2009 11:20am
V14 Continued to take and close out trades until it decimated a 25,000 demo account. I had it set to close 1 position and let the other run, but it kept opening and closing positions as fast as it could.

Something is definitely wrong with the EA.
  • #717
  • Aug 21, 2009 1:37pm Aug 21, 2009 1:37pm
Quoting rcbarlow
V14 Continued to take and close out trades until it decimated a 25,000 demo account. I had it set to close 1 position and let the other run, but it kept opening and closing positions as fast as it could.

Something is definitely wrong with the EA.
rcbarlow,

Were you using fixed lot or % risk?

I had similar experience with v 0.14 but only when using fixed 1 lot risk with exit 2. (More details in posts #710-711)
At first I thought it was because I had a duplicate pair/EA on that profile and several corrupted Comment entries. However, after removing the duplicate, opening a new account and starting fresh, it again repeated the rapid open/ close trades.

Good news is that when I replaced it with V0.15 last night it appears to be working correctly: no Energizer bunny trading.

I don't know if Longbar fixed something in v0.14 or if the change to v0.15 just cured the problem, but so far so good.

JR
  • #718
  • Aug 21, 2009 2:39pm Aug 21, 2009 2:39pm
Yes, I was running fixed lots with #2 exits.

I will try V15 and see if this fixes the problem. I had no other EAs running at all.

Thanks.
  • #719
  • Aug 21, 2009 6:22pm Aug 21, 2009 6:22pm
hi longbar,

will the ea still function as normal if the atr used is the standard one on mt4 and changed from 14 to 20?
  • #720
  • Aug 22, 2009 9:08am Aug 22, 2009 9:08am
Quoting 4freedom
I've studied this thread and really like the system. It's clean, simple and can be traded without emotion or subjectivity. Thanks, Ed.

I ADDED a 10 period EMA on the close to my charts and, looking back in history, this seems to filter out the most of the trades that would have lost. Same trade triggers as outlined by Ed with the additional rule on not taking the trade on a new HA candle close unless it also closes above/below the EMA.

Check it out...
The use of 10-period EMA would be useless here since the system introduced here may be replaced by the 5-13 EMA crossover system which produces same signals as the system.
0 traders viewing now