4H Box Breakout | Page 110 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Box Breakout indicator 197 replies

Breakout EA based on New York Box 3 replies

4H Weekly Box Breakout [TD] 4 replies

EURUSD Box Breakout 6 replies

Breakout Box 5 replies

Attachments: 4H Box Breakout
Exit Attachments

4H Box Breakout

  • #2,181
  • Aug 21, 2009 6:07am Aug 21, 2009 6:07am
Quoting Vikyng
Karl Popper argued that a theory can only be considered scientifically valid if it is falsifiable. This was based on his contension that there is a logical asymmetry between scientific verification and falsifiability.


There was nothing faulty in my statement. We used different criteria for trades.
After 1 loss the probability of second loss is higher with those rules according to which I tested this strategy.

This strategy can perform in ranging markets but not with current rules.
  • #2,182
  • Aug 21, 2009 6:16am Aug 21, 2009 6:16am
Quoting danno96
Hi Dibanos

These are variations which I was playing around with my backtesting to see if...

I think you have been doing great job, but adding more moving parts to an excellent strategy is a waste of time. it will only break the system performance.
The key to this system is money management.
It is where I want t concentrate and I am starting trading live 3-5 versions of money management starting this Sunday. including your rules you posted before.
Very slight modifications and I would like to see results. Plus, taking partial profit before TP1 could be good idea for time of ranging markets. When you cannot get TP3-4 it would be better to pick something on each trade to make it work.
  • #2,183
  • Aug 21, 2009 6:58am Aug 21, 2009 6:58am
I have a word of caution regarding backtesting. There is HUGE confusion about which candle to use. The backtests are not valid if you dont use the right candle. Please post what candle you use, should be 1700-2100 EST.
  • #2,184
  • Aug 21, 2009 7:02am Aug 21, 2009 7:02am
Quoting LegoFx
Hi guys.

I'm pretty new to EAs, so I have a few basic questions:

1) Once I attach the EA to the chart, what happends? I mean, I see in the top lef corner some info and the two yellow lines, but what is really happening behind the scenes? for exmple, Will the EA put an arrow in the chart when is time to enter?
2) If I check live trading, will it open orders and move the TP to BE for me?
3) Is there an initial post in which Ronald describes the different variables in the EA and how to use them?

Thanks in advance.
Behind the scenes:

 

  1. The EA first scans each bar going back until it finds the start of the week (as defined by you). This start could be Sunday, Monday, Tuesday, etc.
  2. The EA then counts forward however many bars you want, to establish the time range.
  3. The EA then check each bar in the range for the highest ever price, and the lowest ever price
  4. It then just waits for the price to break either end of the range and will enter an order/send a signal when that happens.
  5. Rinse and repeat.

Question #2
Not completely. You need to make sure that you set that in the settings.

Answer #C

There was, but I can't find it.

extern string Remark1 = "== Main Settings ==";
extern int MagicNumber = 0; This is a unique identifier that the EA uses to determine which trades belong to it. If you use multiple EAs, make sure that each one has a different magicnumber, so that they don't interfere with each other's trades.
extern bool SignalsOnly = False; If set to true, the EA will only send signals (based on what you select below), and not place trades.
extern bool Alerts = False; Is set to true, the EA will give you a popup alert.
extern bool SignalMail = False; If set to true, the EA will send an e-mail (need to configure this in the terminal).
extern bool PlaySounds = False; If set to true, the EA will play the soud ("Alert.wav") in the /Sounds folder.
extern bool ECNBroker = False; If set to true, the EA will first place a trade with no SL/TP, then modify that SL/TP to whatever it's supposed to be. Thsi feature is in here because some ECN brokers require that trades be placed that way.
extern bool EachTickMode = True; If set to true, the EA will take trades immediately. If set to false, the EA will take trades only at the end of the bar.
extern double Lots = 0; Fixed lot size
extern bool MoneyManagement = False; Enables/disables MM
extern int Risk = 0; % of available equity to use (enter values between 0 and 100).
extern int Slippage = 5; Maximum slippage allowed
extern bool UseStopLoss = True; Enable/disable stoploss function
extern bool UseFixedStopLoss = True; Enables/disable fixed stoploss
extern int StopLoss = 100; Fixed stoploss value
extern bool UseStopLossMultiplier = True; Enables/disable stoploss multiplier
extern double StopLossMultiplier = 3; Multiplier value, this will use the box value (high/low + buffer).
extern bool UseTakeProfit = False; Enables/disables takeprofit
extern bool UseFixedTakeProfit = True; Enables/disables fixed takeprofit
extern int FixedTakeProfit = 60; Fixed takeprofit value
extern bool UseTakeProfitMultiplier = True; Enables/disables takeprofit multiplier
extern double TakeProfitMultiplier = 3; Multiplier value based on a multiple of the box.
extern bool UseTrailingStop = False; Enables/disables trailing stop
extern int TrailingStop = 30; Trailing stop value in pips. Note: The trailing stop only applies to trades in profit.
extern bool MoveStopOnce = False; Enables/disables the movestoponce function
extern it MoveStopWhenPrice = 50; Trigger price in pips for the move stop once function
extern int MoveStopTo = 1; Price in pips relative to breakeven to mvoe the stoploss to when the trigger price is reached

Note: In the above example, when the trade hits 50 pips profit, the stoploss will be moved to Breakeven + 1

Note on Stoploss:
You can enable and use all stoploss possibilities. Keep in mind that the EA will use the tightest stoploss available.

Note on takeprofit:
You can enable and use all takeprofit possibilities. Keep in mind that the EA will use the largest takeprofit available.


extern string Remark2 = "";
extern string Remark3 = "== Breakout Settings ==";
extern int DayStart = 0; (Sunday = 0, Monday = 1, etc) The day you choose to start your week.
extern int BarsFromWeekStart = 1; The number of bars to start from the week. If you use the 4H bars, put in 1, if you use 1H bars, put in 4, etc.
extern int BreakoutBuffer = 0; The number of pips above/below the range to place a trade
extern int MaxTradesPerWeek = 10; Maximum trades allowed per week
extern int MaxLongsPerWeek = 8; Maximum long trades allowed per week
extern int MaxShortsPerWeek = 8; Maximum short trades allowed per week

Note: MaxLong + MaxShort DOES NOT have to equal MaxTrades. I put in this feature so I can allow the EA to give a bias. If--for example--the market enters me in 8 short trades, and they all lose, there is still a chance for that runaway long trade which makes up for all the losses.

extern string Days = "Sunday is 0, Monday 1, Friday 5, etc";
extern bool UseCloseOnWeekEnd = False; Enable/Disable the close trades by week-end function
extern int ClosePositionsDay = 5; Day of the week to use
extern string ClosePositionTime = "23:00"; Time to use.

  • #2,185
  • Aug 21, 2009 7:13am Aug 21, 2009 7:13am
Quoting Telac
I just want to confirm the resourse problem. I'm running a P4 3.2ghz pc with 2 gigs ram and I get the same thing - 100% cpu usage with the ea running on 11 charts. Just a further note; even when I only run 1 or 2 charts I simply can't open the properties menu of the ea. I have to close remove the ea and bring it back onto the chart to be able to access the input. I also cannot use the stategy tester with this ea - it just doesn't work.

I'd like to thank Ace for sharing this remarkable strategy with us and Ronald Raygun for the EA implementation,...
Get a fresh install of metatrader. Put only my EA on that, and tell me what happens.I'm not sure if you have other indis in the terminal.
  • #2,186
  • Aug 21, 2009 7:32am Aug 21, 2009 7:32am
Quoting shiko
I think you have been doing great job, but adding more moving parts to an excellent strategy is a waste of time. it will only break the system performance.
The key to this system is money management.
It is where I want t concentrate and I am starting trading live 3-5 versions of money management starting this Sunday. including your rules you posted before.
Very slight modifications and I would like to see results. Plus, taking partial profit before TP1 could be good idea for time of ranging markets. When you cannot get TP3-4 it would be better...

Hi Shiko,

Totally agree with you, this system is simple, mechanical and objective with no indicator, - KISS. The more the variations, the more the time require to manage the trades and more room to made errors. The key to this system is good money management and maintain the R:R. Taking partial profit may be good idea in a ranging market but not as such a good idea in trending one. The key in this system is to be able to capture those long trend run when they occur. I've not done any backtesting on taking partial profit but is something on my to do list.

I'm think of running a different system for a ranging market together with this system leaving this one as simple as possible.

  • #2,187
  • Aug 21, 2009 8:19am Aug 21, 2009 8:19am
Quoting shiko
After 1 loss the probability of second loss is higher with those rules according to which I tested this strategy.

This strategy can perform in ranging markets but not with current rules.
Although it could be noted that the probability of a loss is greater on each and every trade, not just after a losing trade. That's based on Ace's statement that this system has more losing trades than winning trades. That's o.k., though, because the secret of this system is the risk to reward ratio. A trader only needs to win 20-30% of trades on this system, depending on TP, to be profitable.

In regard to ranging markets, very few systems perform well in these conditions. However, after the market ranges while looking for direction, a strong move in one direction or the other often follows, so it's usually worth the wait.
Greetings from Guanajuato, Mexico
  • #2,188
  • Aug 21, 2009 8:20am Aug 21, 2009 8:20am
Quoting aajduk
When i talk about being fussy look at todays gj. Even though it is friday and i would not trade anyway, there are 2 reasons why i would not take the trade. If you zoom out there is a long term trendline just touching the box candle and there is a support line under it as well. Also it is a very big candle, 180 pips. If you used the previous candle as the box candle like ace does, it would have been a good trade. That can happen for and against you at times.
It meets all my other rules, but like I said, rules are for the blind following of fools...
Anthony thanks for the additional details on your daily trading. I agree with you about blindly following rules.

You are using this daily system for GJ, EJ, EU and GU?

Thanks and have a great trading week next week!

Sandy
  • #2,189
  • Aug 21, 2009 8:25am Aug 21, 2009 8:25am
nubchai,

The limited backtesting i have done also indicates the ucad and au are ok as well.

Anthony
  • #2,190
  • Aug 21, 2009 8:46am Aug 21, 2009 8:46am
Quoting danno96
Hi Shiko,

Totally agree with you, this system is simple, mechanical and objective with no indicator, - KISS....


It sounds great. Unlike you I have no programming skills and do everything manually week by week watching screen
I agree that partial profit taking is bad idea when market is so volatile when you can get 3:1 reward. But in ranging markets, each pip counts.
Well, pips actually have no meaning. i count only % added or lost.
I am going to trade live few versions starting this Sunday and I would really like to add few more pairs that work with this strategy. for now I stick to G/J will risk 1% on it. E/J will risk 0.25% on it. But I thing G/U and U/Can might work well as well. Oil is volatile and Canadian buck is bound to it.
I also want to run 2% risk on another account. Money management rules will have minor differences. Like move to B/E at TP1 and at 50 pips in profit for G/J and 40 for E/J. I actually back tested yesterday since January 1. Return was slightly bigger on TP1,4-5 % difference, but, there were far less losses on 50 pips rule and also very good return. I think account safety is a paramount thing and is more important than return.
waiting for Sunday. watch G/J, support 153.50 has held 3 times while huge triangle is forming from above on 3 hours time frame. The range narrowing... good sign before break out
  • #2,191
  • Aug 21, 2009 8:53am Aug 21, 2009 8:53am
Quoting yen44x
Although it could be noted that the probability of a loss is greater on each and every trade, not just after a losing trade. That's based on Ace's statement that this system has more losing trades than winning trades. That's o.k., though, because the secret of this system is the risk to reward ratio. A trader only needs to win 20-30% of trades on this system, depending on TP, to be profitable.

In regard to ranging markets, very few systems perform well in these conditions. However, after the market ranges while looking for direction, a strong move...

I agree, but I think it would work in 2003 as well, just with partial profit taking. I have checked 2003 by glance. G/J looked funny, many weeks were like break out and then slow painful move by this pair in break out direction the whole week in straight line.
With current rules it won't work but this Sunday candle holds some significance like Pivot point and other similar stuff. so, strategy build around it might work well over time but..., rules gotta be changed from time to time. there should be rules versions for different market conditions.
  • #2,192
  • Aug 21, 2009 9:28am Aug 21, 2009 9:28am
Quoting Ronald Raygun
Behind the scenes:

 

  1. The EA first scans each bar going back until it finds the start of the week (as defined by you). This start could be Sunday, Monday, Tuesday, etc.
  2. The EA then counts forward however many bars you want, to establish the time range.
  3. The EA then check each bar in the range for the highest ever price, and the lowest ever price
  4. It then just waits for the price to break either end of the range and will enter an order/send a signal when that happens.
  5. Rinse and repeat.

Hi Ronald Raygun,

I looked into your EA code, if I'm not wrong the logic condition for breakout is:
......&& Open[Current + 0] < LongEntry && Close[Current + 0] >= LongEntry
......&& Open[Current + 0] > ShortEntry && Close[Current + 0] <= ShortEntry

is not better to add also :
....|| Close[Current + 1] < LongEntry && Open[Current + 0] >= LongEntry
....|| Close[Current + 1] > ShortEntry && Close[Current + 0] <= ShortEntry

?

It would be great if could be added conditional cycle to automatically set the SL to every TP when price reaches the next TP level.

Thanks.
Lubo.
If I know how, I tell You!! Few but Good!!
  • #2,193
  • Aug 21, 2009 9:42am Aug 21, 2009 9:42am
Quoting Vikyng
Karl Popper argued that a theory can only be considered scientifically valid if it is falsifiable. This was based on his contension that there is a logical asymmetry between scientific verification and falsifiability.
I haven't heard of Karl Popper, but I may have met a relative by the name of Jalapeno Popper.

Actually, you're going to have to "dumb down" your post (way down) for me to have any idea whether to agree or disagree.
Attached Images
Greetings from Guanajuato, Mexico
  • #2,194
  • Edited 10:56am Aug 21, 2009 9:56am | Edited 10:56am
Quoting LuboLabo
Hi Ronald Raygun,

I looked into your EA code, if I'm not wrong the logic condition for breakout is:
......&& Open[Current + 0] < LongEntry && Close[Current + 0] >= LongEntry
......&& Open[Current + 0] > ShortEntry && Close[Current + 0] <= ShortEntry

is not better to add also :
....|| Close[Current + 1] < LongEntry && Open[Current + 0] >= LongEntry
....|| Close[Current + 1] > ShortEntry && Close[Current + 0] <= ShortEntry

?

It would be great if could be added conditional cycle to automatically set the SL to every TP when price reaches the next...
Your solution says the EA will enter the trade if the bar opens outside the box.

My solution says the EA will enter the trade if the price moves outside the box. If you set EachTickMode to false, my solution says the EA will enter the trade if the price closes outside the box. (I think this is exactly what you're trying to do in your code)

Your code isn't wrong, it just uses a different order entry philosophy.
  • #2,195
  • Aug 21, 2009 10:12am Aug 21, 2009 10:12am
Quoting aajduk
nubchai,

The limited backtesting i have done also indicates the ucad and au are ok as well.

Anthony

Thanks Anthony

Sandy
  • #2,196
  • Aug 21, 2009 10:43am Aug 21, 2009 10:43am
Quoting yen44x
I haven't heard of Karl Popper, but I may have met a relative by the name of Jalapeno Popper.

Actually, you're going to have to "dumb down" your post (way down) for me to have any idea whether to agree or disagree.

I agree Yen...that post was full of $10 words!!
  • #2,197
  • Aug 21, 2009 12:17pm Aug 21, 2009 12:17pm
Ronald Raygun - Octomanoak is having trouble posting so he has asked to post this for him.

[quote=Ronald Raygun;2973230]That EA runs on a relatively underpowered Intel Little valley (~1GHZ) with 30 of its buddies with no problem. What are the specs of your computer?

As for getting the EA to work...

1) How many bars from the week's start on your timeframe makes 4 hours?
2) What is the lot size? I set the lot size to 0 by default. Back in the day, I used a standard size of 1 lot, and some idiot attached my EA to their chart and blew their account. That same idiot decided to sue me for "negligence".

Hey Ronald. I'm using FXDD, so the first 4 hours of the week make 4 hours. Using a 2 ghz duo with 4 gb ram. Essentially using droland's settings. My settings below:

Remark1=== Main Settings ==
MagicNumber=0
SignalsOnly=0
Alerts=0
SignalMail=0
PlaySounds=0
ECNBroker=0
EachTickMode=1
Lots=0.10000000
MoneyManagement=1
Risk=2
Slippage=5
UseStopLoss=1
UseFixedStopLoss=1
StopLoss=180
UseStopLossMultiplier=0
StopLossMultiplier=3.00000000
UseTakeProfit=0
UseFixedTakeProfit=0
FixedTakeProfit=60
UseTakeProfitMultiplier=0
TakeProfitMultiplier=3.00000000
UseTrailingStop=1
TrailingStop=125
MoveStopOnce=1
MoveStopWhenPrice=90
MoveStopTo=1
Remark2=
Remark3=== Breakout Settings ==
DayStart=0
BarsFromWeekStart=1
BreakoutBuffer=20
MaxTradesPerWeek=10
MaxLongsPerWeek=8
MaxShortsPerWeek=8
Days=Sunday is 0, Monday 1, Friday 5, etc
UseCloseOnWeekEnd=0
ClosePositionsDay=5
ClosePositionTime=23:00

Does it matter which timeframe chart the ea goes on?
  • #2,198
  • Aug 21, 2009 12:50pm Aug 21, 2009 12:50pm
Can anyone guide me on this....about ATR.
I have an ATR (16) reading of 1.0978 on the chart.Is it meant 978pips for the week or otherwise. Thanks
  • #2,199
  • Aug 21, 2009 1:03pm Aug 21, 2009 1:03pm
[quote=droland;2976165]Ronald Raygun - Octomanoak is having trouble posting so he has asked to post this for him.

Quoting Ronald Raygun
That EA runs on a relatively underpowered Intel Little valley (~1GHZ) with 30 of its buddies with no problem. What are the specs of your computer?

As for getting the EA to work...

1) How many bars from the week's start on your timeframe makes 4 hours?
2) What is the lot size? I set the lot size to...

If you really are using droland's settings, you need to use the 1HR chart.

Everything looks ok except for the following:

BarsFromWeekStart: needs to be 4 if you're using droland's settings (4 bars from the week's start).
  • #2,200
  • Aug 21, 2009 1:39pm Aug 21, 2009 1:39pm
Quoting Ronald Raygun
I actually have neither. I just watch MythBusters too much.

As for everyone else, there is a new minor feature added to this EA and for all future EAs I put in. It's the ability to trade with ECN-style brokers who require a trade and the stoploss/takeprofit to be added separately.

I have also added a feature to close all trades by a certain time of the week (like friday?).


Days = "Sunday is 0, Monday 1, Friday 5, etc";
UseCloseOnWeekEnd = False; If set to true, the EA will close all trades and not open new ones after the time specified...
HI Ron.

Your concept of nesting the OrderSend in an OrderModify function I had never seen before. I reached out to some really top notch programmers like HighAchiever and Bluto (Craig Stoltz) (Phd Computer Science - Nasa consultant). The weakness of the MT4 platform apparently will cause some problems if on the OrderSend part of the Order we get a requote or some other error.

Here are their comments:
Craig Stoltz:
Dave,

Yes, this is a clever trick that will work but it doesn't adhere to very good programming design practices.

It works because you've enclosed the OrderSend() function inside the OrderModify() function which is completely acceptable and efficient. The OrderSend() function implicitly returns a Ticket Number value from it's call to the broker server to place the order. Therefore, the OrderSend() function acts as a proxy for the Ticket Number value which is the required first parameter of the OrderModify() function.

Now here's the rub. To accommodate ECN brokers, you typically want to code the OrderSend() & OrderModify() steps as two separate processes. You do this because you need to first interrogate the Ticket No. return value of the OrderSend() function. If it's greater than zero, then the OrderSend() was successful. However, the OrderSend() return value can be less than zero which would indicate an Error Code that you should check for and report via an alert. Additionally, if you do receive an Error Code, in the consolodated OrderModify/OrderSend edition, you'd be attempting to perform an OrderModify() on a bogus ticket number (error code < 0) which would in turn generate another error.

Just my two cents worth. For a reliable EA, I'd stick to the two-stage traditional method. It's not that much more effort to code properly.

Craig

Here is another comment from Highachiever:

Dave,

I full agree with what Craig says.

Whilst this might be a fancy trick, you are still not sending both orders at the same time. This is still a 2 step process. What is occuring is the order send is activating first and then upon the return of the order send the order modify is executed.

This would mean that the price server is in fact sent 2 orders in very quick succession, and if this is wrapped in a loop to check for successful execution without any sleep statements then the price server will be hit in very quick succession.

Many of the brokers are these days checking to see if EAs act in this inappropriate way, and if they do they will either ban, or introduce their own latency when that EA / account sends orders (read context busy errors).

This is not desirable, so best solution is as per Craigs post in that 2 seperate processes correctly programmed to be price server compliant is the best approach.

Cheers,
Tim
_________________________
I have seen some code that really works well as it loops up to 25 times to make sure the OrderModify has been recieved. I think it also loops for the entry as well to accomodate those damn requotes etc. Let me look and I will post it here.

Dave R.
8 traders viewing now