Simple Breakout/Straddle EA required | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Breakout Straddle! 19 replies

Breakout long straddle with fixed odds 19 replies

very simple indicator required 0 replies

Simple Explanation Required 1 reply

Breakout EA (can probably be used as straddle too...) 0 replies

Simple Breakout/Straddle EA required

  • #1
  • First Post: Oct 1, 2009 9:30am Oct 1, 2009 9:30am
Hi

I'm looking for a very simple Breakout or Straddle EA so I can do some backtesting. I've looked through hundreds of threads and tried countless EAs but I can't find anything that meets or can be modified to meet my requirements:

1. Breakout based on one hour. Stoploss for long or short is the opposite side of the breakout box.
2. No take profits, trailing stops or anything like that. Simply close out the position before the breakout reference hour the following day.

I have a feeling that this type of strategy could be very profitable based on the fact that it is psychologically difficult to implement. It will have lots of losers but the winners will be very big.

Thanks

Pete
  • #2
  • Oct 8, 2009 1:41am Oct 8, 2009 1:41am
I noticed nobody replied to this posting and I was wondering if the request is still open. I have finished writing several EAs and might be able to reuse some of the code I developed. Let me know.
  • #3
  • Oct 8, 2009 3:43pm Oct 8, 2009 3:43pm
Quoting CodeMeister
I noticed nobody replied to this posting and I was wondering if the request is still open. I have finished writing several EAs and might be able to reuse some of the code I developed. Let me know.
I was also thinking of a breakout method myself. My head has been bashed so many times with different systems. I am finally stopping the pain and going with simple systems.

I have an EA that I have hacked and slashed to try and make it do what I want. It is nowhere complete, but it does take trades (sort of). I am not a programmer so please forgive me.

It is a modified version of the speed trap system on this forum.

If anyone has any interest I can explain what I am doing and what is wrong with it.

Everett
  • #4
  • Oct 9, 2009 3:00am Oct 9, 2009 3:00am
Quoting CodeMeister
I noticed nobody replied to this posting and I was wondering if the request is still open. I have finished writing several EAs and might be able to reuse some of the code I developed. Let me know.
Yes - please post.

I'm currently live trading this strategy and I'm getting some spectacular returns. I'm straddling the USD/JPY candle that starts at 21:00 GMT and finishes at 22:00 GMT (22:00 to 23:00 London time) and am getting some spectacular results. Stops are very tight and it can move 7 or 8 times in you favor during the day.

I'm playing around with entries as I'm asleep when the trades are triggered so an EA will be really helpful. Thanks guys.
  • #5
  • Oct 9, 2009 8:42am Oct 9, 2009 8:42am
Maxx, I was offering to code your strategy not post my code. There are some common features between the two so I wouldn't have to start from scratch. This makes it easier to do of course and I should have something ready for testing by Monday or Tuesday.

My strategy (15m JPYUSD channel breakout) is still a work in progress and is not ready to post. As far as I can determine, the code is bug free. It has done quite well over the past month, but my backtesting showed only small gains. It is quite sensitive to market conditions and picking the right settings is crucial for it. I need to put some more effort into researching the reasons for this.
  • #6
  • Oct 9, 2009 1:17pm Oct 9, 2009 1:17pm
CodeMeister,

I was wondering if you can help me with is piece of code. I am not a programmer, but I am trying to learn.

I am trying to set a start hour and end hour for the breakout range. Once the EA has found the range I would like to take the highest and lowest of the range and store them in global variables. Once the trade executes I would like to reset range and wait for the next day range period.

The code below only allows me to set the end hour and then I have to look back to get range. Also the global variables is not working properly since high and low change during the day, once the range is found it should stay until reset for next day.


extern int Range_Hours = 7; // End hour to start look back for range
extern int Look_back = 3; // Number of hours to look back to determine range


if(Hour()== Range_Hours && Minute()<5)
{
double H =High[iHighest(NULL,0,MODE_HIGH,Look_back,1)];
double L =Low[iLowest(NULL,0,MODE_LOW,Look_back,1)];
GlobalVariableSet("BoxL",L);
GlobalVariableSet("BoxH",H);
}

double BL=GlobalVariableGet("BoxL");
double BH=GlobalVariableGet("BoxH");

Thanks if you can help.
  • #7
  • Oct 10, 2009 5:16am Oct 10, 2009 5:16am
Oh OK - I thought you might have come across some previously posted code that you could repost.

OK I'll tell you my ideas and you can make a decision whether you would like to code the strategy and share it with me.

Idea 1 (as I mentioned in a previous post)
Take the USD/JPY candle that starts at 21:00 GMT and finishes at 22:00 GMT (22:00 to 23:00 London time). Buy order is the high of the candle and stoploss is the low of the candle. Sell order is the low of the candle and stoploss is the high of the candle.

If you have code on a 15 minute candle it should take you only a few minutes to modify your code and backtest this strategy.

Idea 2
Take the USD/JPY candle that starts at 21:00 GMT and finishes at 22:00 GMT (22:00 to 23:00 London time). Buy order is the 10 pips above the high of the candle and stoploss is the high of the candle. Sell order is 10 pips below the low of the candle and stoploss is the low of the candle.

General thoughts
After trading this strategy for a few weeks I can say that I have been manually moving my stop. It seems like if there is a big move in one direction there is usually a retrace. For example, a few days ago the trade triggered sharply long and there was a number of bars of consolidation. I moved my stop below those bars so that if there was a retrace I would hold on to some profit, but I gave it plenty of room if it was going to go higher. Maybe a manual stop moving is essential to the profitability of this strategy so an EA won't work?

My second thought is perhaps there should be an automatic take profits based on the average true range of the pair or a similar calculation based on the movement of the pair. I don't think there should be a take profit based on a multiple of the size of the trade but I'm not sure the best way to calculate an appropriate take profit figure.

My third thought is that the second untriggered trade should be canceled at some point in the day but I'm not sure what should decide this. I don't think it should only be based on the time of the day - maybe it should be canceled on the basis of the distance between it and the current price when compared to the average true range (or something similar).

What do you think?


Quoting CodeMeister
Maxx, I was offering to code your strategy not post my code. There are some common features between the two so I wouldn't have to start from scratch. This makes it easier to do of course and I should have something ready for testing by Monday or Tuesday.

My strategy (15m JPYUSD channel breakout) is still a work in progress and is not ready to post. As far as I can determine, the code is bug free. It has done quite well over the past month, but my backtesting showed only small gains. It is quite sensitive to market conditions and picking the right...
  • #8
  • Oct 10, 2009 2:51pm Oct 10, 2009 2:51pm
I can Modify my box breakout if you need it...

PM me if you want it.
  • #9
  • Oct 10, 2009 6:56pm Oct 10, 2009 6:56pm
Your comments are very helpful and I have few things of my own to add. I looked at several daily charts and drew lines on them at the high and low of the bar. The thing I noticed was that most of the day's activity happened away from the box after the first few bars got moving. This makes me think that winning trades have to able to run with a wide trailing stop and profit target rather than grabbing the first 20 to 30 pips. I think we are both in agreement on this point. I am thinking of a dynamic trailing stop and profit target that starts wide at 40 to 50 pips and are tightened with each new bar keeping a risk/reward ratio of 1:1.5. I have experience with an ATR like stop loss and that may be a good alternative.

Both your ideas 1 and 2 can be incorporated with an input parameter for Gap = 0, 10 (or something else). I definitely see the benefit of the gap as the first bars are very near to the high/low range. I can see the very real possibility of both sides of the straddle being triggered and with both trades ending up losing. Providing a gap before the trigger price should reduce that risk at the expense of some profit. I was wondering how often you have seen this situation.

The conservative approach would be to do as you suggested and cancel the opposing trade at some determined point. A more aggressive approach probably won't be worth the risk.

The other thing I thought of was not holding any orders open on weekend and not trading on the first Friday of the month (NFP). There should also be minimum and maximum range (say 8 to 20 pips) to avoid the unusual narrow and wide range situations.

I can see enough potential in this strategy to proceed with a version that can be backtested to determine how it performs on a long term basis. The results you are seeing could of course be due to the market situation at this time and may not continue.
  • #10
  • Oct 12, 2009 2:37pm Oct 12, 2009 2:37pm
I have coded the EA and partially tested it. The results are not very encouraging. Over the past month or so, they were strong with a 46% win ratio and a 10% payoff. I also backtested it during the summer and it had a 78% loss rate and a -14% payoff. So there is definitely dependence on markets conditions and input settings which should not be surprising.

My main findings about this strategy were:
Stops and profit targets need to be left wide open otherwise the trades get stopped too soon.
Closing the opposite order when the first is triggered seems to work as well as leaving it open as long as possible.

Let me know if you want to test the code.
  • #11
  • Oct 13, 2009 6:27am Oct 13, 2009 6:27am
Quoting CodeMeister
[size=2]Your comments are very helpful and I have few things of my own to add. I looked at several daily charts and drew lines on them at the high and low of the bar. The thing I noticed was that most of the day's activity happened away from the box after the first few bars got moving. This makes me think that winning trades have to able to run with a wide trailing stop and profit target rather than grabbing the first 20 to 30 pips. I think we are both in agreement on this point. I am thinking of a dynamic trailing stop and profit target that starts...
I'm having trouble visualizing how this will work. Could you post an example?

Quoting CodeMeister
Both your ideas 1 and 2 can be incorporated with an input parameter for Gap = 0, 10 (or something else). I definitely see the benefit of the gap as the first bars are very near to the high/low range. I can see the very real possibility of both sides of the straddle being triggered and with both trades ending up losing. Providing a gap before the trigger price should reduce that risk at the expense of some profit. I was wondering how often you have seen this situation.


Very often. Yesterday I was triggered in one direction and stopped out. Today I was triggered in both directions (by a couple of pips) and stopped out. The original idea behind my strategy was to have a very tight initial stop and get stopped out a lot. When the strategy works it really makes money. For example, on Friday it triggered and took off. On a 10 pip stop it made 135 pips for the day. Unfortunately I moved up my stop at around 8am london time so I only made around 60 pips. But if you're risking 20 pips per day (10 pips each straddle) a good day covers you for at least 6 days, potentially more (assuming both sides of the straddle aren't hit). The hard part is not moving your stop at all and letting the trade run. I could have doubled my profits on Friday if I hadnt moved my stop, but at one point I would have been roughly only 10 pips ahead (down from 70). Can you psychologically take that drawdown in your profits? If you can it seems that the profits are pretty good.

Quoting CodeMeister
[size=2] The conservative approach would be to do as you suggested and cancel the opposing trade at some determined point. A more aggressive approach probably won't be worth the risk.

The other thing I thought of was not holding any orders open on weekend and not trading on the first Friday of the month (NFP). There should also be minimum and maximum range (say 8 to 20 pips) to avoid the unusual narrow and wide range situations.

I can see enough potential in this strategy to proceed with a version that can be backtested to determine how it...
These are great points. I'm thinking of widening my initial stops and the NFP is a good idea.
  • #12
  • Oct 13, 2009 6:28am Oct 13, 2009 6:28am
Quoting CodeMeister
I have coded the EA and partially tested it. The results are not very encouraging. Over the past month or so, they were strong with a 46% win ratio and a 10% payoff. I also backtested it during the summer and it had a 78% loss rate and a -14% payoff. So there is definitely dependence on markets conditions and input settings which should not be surprising.

My main findings about this strategy were:
Stops and profit targets need to be left wide open otherwise the trades get stopped too soon.
Closing the opposite order when the first is triggered...
What was your initial stop and did you have a trailing stop? Which bar did you base the straddle on?
  • #13
  • Oct 13, 2009 9:34pm Oct 13, 2009 9:34pm
Your comments got me thinking that I should take a different approach to testing and input settings. So I opened everything up and the results are more encouraging. I was able to get July's results to break even. Unfortunately September's results were also went to break even. This just proves that with enough back testing, you can find settings that match what the market does. Of course that is almost useless for forward testing and eventual real world implementation. The reason I say things are encouraging is that we still relatively early in the testing and revising stage and with more insight and effort, we might find the secret yet.

I do disagree with you about an initial tight stop because I think that would prevent the big moves from developing, unless we were to always have pending orders in place and take 3 or 4 losers on some days.

I think I will PM you my test results with the different settings. I will also send you the code and you can try some things for yourself.
  • #14
  • Oct 15, 2009 7:38am Oct 15, 2009 7:38am
What about instead of basing the breakout off the high and low of the candle, basing it off the close of the candle and using the size of the candle to determine the size of the range required for the breakout?

For example, take the close of the 22-23 hours London time candle. Long breakout is Close + Spread + Size of candle. Long stop is Close + Spread. Short Breakout is Close - Spread - Size of can dle. Short stop is close - spread. This makes more sense as the high and low of the candle are irrelevant and the current price at the time of the trade is placed (the close) is more important. And you're just using the size of the candle to give you a rough approximation of the current volatility.

Or if you wanted a slightly wider breakout you could make it close + spread + 1 & 1/2 times size of candle.

What do you think?
  • #15
  • Oct 16, 2009 2:42am Oct 16, 2009 2:42am
Hi Guys,

Been testing this theory manually - I have a couple of points:

GMT runs out in a few days, so don't worry about adding this to the EA.

From what I have seen on the few tests I have done, the margins on the hour candle are quite large, and it seems to work better with big stop losses in place, although keeping an eye on trend is important as occasionally this would incur huge losses.


I really think you have something here though, and I have even had some success doing this on every hourly candle with shorter stops and profits.


I am not a professional coder, but if I can offer any help or support just let me know!
  • #16
  • Oct 16, 2009 3:11am Oct 16, 2009 3:11am
This is an easy change to make to the code, but the back testing will take several days. As far as the rationale goes, basing the box on the closing price versus the high/low doesn't seem to be a big difference maker. The same problem exists - namely the bars have to stay inside the box until the big move happens and if the box is too big we give up too much profit. If you think the closing price will position the box better and/or make it smaller, I guess we will see during testing.

I will definitely make the multiplier (1.5 or whatever) an input setting.

Another idea to consider is make the box adjustable every 3 hours or so. This would allow the small noisy moves to stay within the box without triggering the pending orders and then hitting the stop loss. By adjustable I mean both the size and the position of the box relative to the recent price activity. This would allow early moves to happen as before and slower moves to develop before the wrong trade was triggered. I did this with my channel breakout strategy and it seems to have given me a small edge although I sometimes sacrifice part of the early move.
  • #17
  • Oct 21, 2009 3:02am Oct 21, 2009 3:02am
Codemeister how did you go?
  • #18
  • Oct 21, 2009 7:58am Oct 21, 2009 7:58am
Quoting clocksley
Hi Guys,

Been testing this theory manually - I have a couple of points:

GMT runs out in a few days, so don't worry about adding this to the EA.

From what I have seen on the few tests I have done, the margins on the hour candle are quite large, and it seems to work better with big stop losses in place, although keeping an eye on trend is important as occasionally this would incur huge losses.


I really think you have something here though, and I have even had some success doing this on every hourly candle with shorter stops and profits....

Clocksey, sorry but I don't remember reading your comments before. Thanks for the interest and encouragement. Eyeballing this strategy will give you some idea about its potential but you have to run it in backtesting to understand the real challenge. I will PM you the executable and test results.

We are at breakeven and I would like to think that there is one thing missing that will tip this into profitable results. In reality, it will probably be a bunch of little things. I really like your idea about trading with the trend. Could you expound on that, even in non-technical terms. I have a few ideas of my own and I need more time to clarify them, but I think you are onto something.
  • #19
  • Oct 21, 2009 8:15am Oct 21, 2009 8:15am
Quoting TheMaxx
Codemeister how did you go?
Peter, I have been working on something else. I did have time to look at the adjustable box idea. Out of the 40 or so trades, it would have turned one small loser into a nice win and with 2 others it might have been effective, but they were too close to call. Most trades would not have been affected. I haven't looked real close at your idea of using the closing price, but I think it would gain more pips in some of the winning trades and save a few pips on some losers rather than turning losers into winners.

One idea I had in working on the other code was to identify the type of market situation and try to take trades to be in sync with what is going on. By this I mean that the strategy I was working on worked well in sideways market and not very well in a trending market, so I want to take less trades when the market is trending to make a winner out of the strategy. Do you have some thoughts on what influence the type of market conditions would be on your strategy? Should we only straddle in sideways consolidation conditions and trade with the trend otherwise? How do we identify these conditions?
  • #20
  • Oct 21, 2009 8:35am Oct 21, 2009 8:35am
What's wrong with any of the Generic Breakout EAs?

http://www.forexfactory.com/showthre...=155239&page=9

Kore also coded a generic breakout EA

http://www.forexfactory.com/showpost...&postcount=113
0 traders viewing now