Weekly Gbpjpy Breakout System

Wow... Thanks... I didn't know that. Back to the drawing board... and yet, I've gotten favorable results based on my analysis so far. Maybe it won't matter in the end... Time will tell. So, what do you have for a high and low on the channel, if I may ask?

Okay, I just downloaded a North Finance demo. Looked, and for this week there isn't much difference. The only difference would come if the 1 extra hour candle at the end of the NF week holds a higher high or lower low than those defining the channel earlier, NF and IBFX would be about the same. A quick scan of the last 5 weeks showed that final bar did not make a change. But I realize that it might. I'll look back more completely in the near future to see if I can quantify how often this has happened. Thanks Billbss for point this out.

Short has been triggered

Looking back through July 2007, the influence of the additional 1 hour bar that IBFX is missing each week, comes into play, or is a factor in this system about 1 time every 4 weeks. When this bar changes the channel size, about half the time it is a significant factor in how trades play out the following week, and half the time it does not. The majority of the times it is a factor it is to the advantage of the trader, but not always.

In the long run, it would seem to be worth considering, but not factoring this is probably not going to change the system from a winner to a loser. It changes the results significantly about once ervery 2 months. Thanks again, Billbss

HI guys,
Have been following this system with great interest. I am relatively new to forex and am looking for a mechanical type system. I have spent most of the weekend backtesting this system and although the results looked promising to start with when I applied the system and stepped through every trade it unfortunately doesn't look to good for the way I would trade it over the 2005 and 2006 years. Before I go on I just want to say I am new to this and there is always the possibility that my backtesting is flawed in some way.
I have attached my spreadsheet and MT4 template to show the setups for each trade.

I trialed 2 systems side by side. I have used harolds rules.

System 1 - Take profit at 1x channel.
Use the 2 bar, then back to 3 bar. If channel less than 100pips. Make a 100 pip channel.

Add 7 pips on to upper level of channel (allowance for spread).
Buy and sell stop orders placed 3 pips outside of upper and lower channel lines to allow for channel to be actually broken before entry.
Stop levels also set at 3 pips outside of channel again to allow for channel to actually be broken before being stopped out.
Take profit at 1x channel band.

Gaps - Ignore smaller gags - less than 20 pips.) For larger gaps If gap is filled within the channel trade as per normal. If it gaps outside of channel, don't take the trade. If price returns to channel and closes gap, trade as per normal. (I actually backtested ignoring gaps vs waiting for them to close and it didn't really make much difference to the bottom line)

If the first trade is successful then thats it for the week.
If the first trade fails, Take a trade in the opposite direction if triggered.

System 2 - Take profit at 2x channel.
Same as above but take profit at 2 x channel band.


Risk

Risk is 1% of account on each trade ( compounded).
I see the system profitability is sometimes measured in pips instead of return on account. I'm still trying to figure out how that accurately measures profitability. (I'm still a newbie, remember).
For me I will only risk a set % per trade. Whether the channel is 100 pips wide or 400 pips wide it will still be 1% risk. This means that some pips will be worth $1 (100pip chan) or 25 cents (400 pip chan). Again, if I am missing something here I would appreciate more experienced people pointing me in the right direction.

I have also subtracted 5% off of the winning trades. This is to allow for the 3 pip buffer on either side of the channel. (probably more like 3% but I like to allow for the worst case scenario)


RESULTS


System 1 - Take profit at 1x channel.

2005 - 7.95% return on account
2006 - 9.62% return on account

System 2 - Take profit at 2x channel.

2005 - 8.78% return on account
2006 - 2.04% return on account


I hope some of the more experienced guys will look at my spreadsheet and tell me I am wrong because I like the look of this system.


ps 2007 results look a lot better than this. Maybe 20% + on account but haven't finished test yet.

BigRicky, Firstly, thank you for your hard work. No matter what anybody ends up saying about it down the road, good bad or indifferent, I applaud your contribution.

I'll have to look at what you've got more closely and do some deep thinking. I'm no expert, but I work on it a little bit everyday, really. Someday, when I grow up...

That said, I can't see if you accounted for sort of the "final scenario". That is, under original rules, you try for 2X, and if that fails, you try for 3X. If your 3X attempt doesn't get stopped out, you hold to the close of the week. So, in some cases you might end up with 2.25X or 2.75X. Indeed, I suspect there are 3X weeks as well; and I'm not seeing this accounted for in your spreadsheet.

You might get a fair bit of feedback on Net Return vs. Pips. If I think on it some I might be able to write something that makes sense, but one thing that jumps out right away is that people are using different account types, and obviously different account amounts, as well as different % of account at risk. If we had the raw numbers for each week in your 3 year study we could do a fair bit more analysis too. For example, many people will factor the size of their position based upon the max drawdown. The way you've calculated things that kind of statistic is filtered. Finally, we can't really see if there is a trend in the numbers, such as, Profit targets with channels greater than X (250 pips, 300 pips, etc.) are very rarely achieved (stated here only as an example, not as a fact).

A natural question that comes up in the course of these things is the viability of a trailing stop, or stops in general, and what the correct levels should be. I'm experimenting with this right now, as a matter of fact. This week, I opened 2 positions. One with a 1X TP, one with a 2X TP. I know from looking back over the weeks that I have studied that more than half the time the market will not make the 2X mark. I also know, 2X seems more plausible on the second move of the week. I suspect that the 2X on the first move this week could be a bit tough, seeing how far we had to go, AND take out a new low... So, on my 1X position, once it was 1/2 the way to profit, I put on a TS. As things go, I was just stopped out (Tuesday's Asian market surged down, and has retraced some -- enough to hit my stop). Well, I now feel safe to let the 2X position run... if it will, because I've got money off the table. If it reverses all the way to the channel without hitting the 1X TP, I can close it and watch to see what happens next. If it pushes forward to the 1X, I figure to put on a channel crossing reverse position if it gets there (I'll probably put a TS on it once it gets half way across). The market will do what it wants to do, but I've seen enough to consider that a move up might have a better shot at the 2X -- at least for this week. If it doesn't do it, that's fine. If it finds 2X going down on the first move of the week, so be it (I'm already in profit, and I'll make some more too).

Okay, I know I'm rambling on... Sorry. My main point is that there is much to consider about how to trade this channel breakout. When you work off your account balance we miss the details of the trade, the nuts and bolts. There could be a big difference between making 1% on a 100 pip move, and making 1% on a 400 pip move. The market that moves 100 pips if fundamentally different from the market that is able to move 400 pips. I hope you see my point here. When to expect which, and the probability of each are things to consider. At some point I think one has to look at the upcoming week and evaluate possible news influences on the market. A market with virtually no news is less likely to run 500 or 600 pips, no matter what your channel size is telling you.

Your money management approach looks very prudent, and is an intelligent, if conservitive (for me), way to approach trading this system. But to refine the system we need to look at the raw numbers, and THEN eventually see if your money management is too conservative for this system, just right, or outrageously risky. Again, what you are doing is probably fine based on what we already know about this system. But in my thinking, a money management approach is applied after you've analyzed what you can expect from the system.

On your template, as I was reading your post I got the impression I'd somehow see all the plots for all the weeks, but the only thing showing on my screen is the current week. Just checking to see if that is what you packaged.

Thanks again for your contribution.

results for 2005 were 2200 pips 2006 1300 pips 2007 4100 pips thats how you determine profitablity

In a way.

All pips aren't equal, though.

This week, for example, if my trade hits profit, it will have gained 240 something pips.

I will have made the same money as a week when it only took 125 pips to hit profit. I size my trades to risk 2% if I lose. I also make 2% if I win.

Qman,
Thanks for the reply. I haven't yet tested the 2x / 3x system. Just the 1x / 1x and 2x / 2x. I will look at this. (Your'e right this was the original and I should have tested it.)

If you scroll back to 2005 and 2006 on the template you will see the drawn in channels (Haven't done 2007 yet)

I only used 1% account system, as this is the only way I know how to test a system (Yes, I am a conservative trader).
I would love to know how to convert the positive pip total that this system returns into a repeatable R/R type system but as of yet I don't know how to do this.

Harold, Thanks for your reply. I sense that you were a little frustrated by my post. Please be kind, I am just new and trying to learn. I realise you must get lots of newbies making inaccurate posts here, but the only reason I spent time testing this system is because it looked very sound.
Would you be able to explain to me what money management, R/R system you would use to convert the good annual pips returned into a trading system.

Thanks for your help guys.

bigricky, You can see what Billbss has posted. He is a seasoned trader, and he does this similarly to you, except he is doing 2%. I think that is very reasonable for this system, based on what we know about it so far. One might be able to risk more depending on what all the numbers work out to, and whateverelse someone has going on in their account. As I was trying to write earlier, I think you have a solid approach. It's just that, in evaluating the strengths and weaknesses of this system, you have to look at the pip counts to see what percentage of your account makes sense to trade (and in my case, when, since I expect to take positions on a reverse cross of the channel at some point during the week). And that gets to the point of my interest.... I really like this little system, but I think there are additional opportunities lurking in here. I'm a very analytical person by nature, but I do detailed analytical work in my real job all day long. So, when it comes to diving into a hard look at FOREX I'm often just running on fumes. Sort of ticks me off, cause if I could apply myself when I'm at 100% I know I'd make better progress... So, hang in there. People like Billbss are really cool. They don't send you in a wrong direction very often. If you pay attention you'll learn something everyday.... Little things, like, "Hey, your broker closes before everybody else ... (you dolt!)..." Okay, I just added that last part. He's usually pretty tactful ....

Billbss,
Thats how I see it aswell when taking one of these trades I risk a certain amount of account (either 1% or 2%) on each trade. The number of pips doesn't really matter. (Unless backtesting proves that smaller or larger pip channels are more sucessful etc..).

This is what I based my backtest on and came up with the above figures(1% RR).

2% RR per trade
2005 1x chan 17.93%
2x chan 19.71%

2006 1x chan 21.70%
2x chan 5.36%

If anybody else has backtested this with a fixed % RR system, or can show a different system that shows a different result I would be keen to swap spreadsheets and compare notes

Thanks guys

Qman,
Looked at a 2x chan trade, and if it fails go for 3x on the next swing (like the original system).



2005 Only went to 3x chan 3 times (3 % to previous figure)
2006 Didn't make it to 3% chan at all.( Actually closed under 2% creating a lower profit)


RESULTS (2% account per trade)

2005 T/P 1x (2nd trade 1x) 17.93%
2005 T/P 2x (2nd trade 2x) 19.71%
2005 T/P 2x (2nd trade 3x) 26.86%

2006 T/P 1x (2nd trade 1x) 21.70%
2006 T/P 2x (2nd trade 2x) 5.36%
2006 T/P 2x (2nd trade 3x) 1.07%

Have just finished the backtest for 2007. As expected, very good results


2007 (2% risk per trade - compounded)
1x chan T/P (2nd trade 1x) 46.09%
2x chan T/P (2nd trade 2x) 65.21%
2x chan T/P (2nd trade 3x) 47.16%


Updated SS and MT4 template attached

thats a great job..thx for your effort...
since now your work shows us a more stabil and profitable way of this..
I'd like to see you opening a new thread with optimization of this strategy you made..
thx again..nice work...

Bigricky, thanks for your continued hard work. That was a brilliant idea of yours to use the fibo tool to draw all these lines.

Since you've done your analysis for 2007, I was wondering if you had a template with the plots for that year. The last one I downloaded didn't have anything for 2007. If you've got the work done it would be a time saver for all of us if you can provide it.

Nomask, good to see you checked back in here. Wonderful little system you have here. Thank you very much for sharing.

Qman,
last post has the latest MT4 template attached to it.

Thanks for the feedback Nomask. I am looking at some optomisation idea's now. If you guys have any ideas of the things that might be good to trial please post them.

Idea's so far.

After we take a 1x profit and price returns to the channel, probability of it crossing the chan to the otherside and taking another position.( This was mentioned earlier in the thread by you guys)

Seasonality? (A new concept for me, But system seems to perform better at the end of the year)

Compare results from 2bar, 3bar and 100 pip entries.

Possibility of a 3rd trade if the first 2 failed

+ a few other things.

Basically I have the time to backtest but am lacking some of the experience. So as not to waste my time testing things that may be pointless please give me some feedback and any other ideas that might be worth testing.

Nomask, I am happy to start a new thread if that is the normal thing to do in this situation (Am quite new to posting and don't really know the right protocols), although I would prefer to carry on posting here if that is alright with you.

Thanks
Richard

As fond as I am of the idea that price will re-cross the channel after getting 1x -- if it returns to the channel, that is... In looking at your earlier years just briefly, it seems that this is a more recent phenomena, probably related to the greater volatility of the market. No one can say where we are headed with any certainty, but it seems to me that as more people start "playing" in the forex markets the more volatility we are seeing. Just an opinion, and that may not be what is really going on at all. But, if volatility continues to increase then I think this crossing trade will work. In quieters times, it would seem it won't. So, to be more confident in this trade we would want to have some sort of tip-off that the market was "generally" volitile enough... How to determine? Lots of people use ATR to try to quantify volatility. I'm not sure that is best, but it would be a start.

In that same vein, it might be worth looking at ATR for the previous week vs what your channel height is. It seems logical that a bigger ATR and a smaller channel might lead to a higher succes ratio, but the market sometimes resists such logic. But I think it is worth a look.

I quickly glanced at about 8 months last week, without the benefit of your lines, mind you, and thought I saw something about Fridays. If from 0000 Friday morning to the close, price was up significantly, Sunday and Monday, longs were favorable, and vice versa. If there was no significant difference (say price opened and closed within 40 or 50 pips) neither is favored (even chance).

There is the regular host of lagging indicators, of course. One thing I started looking into this morning was where a Stochastic(7,5,5 for no particular reason) was when price crossed the trigger. Seemed like if Stochastic was at 80 for long, or 20 for short, there was a higher failure rate or that cross was early -- and one could have waited for a later crossing in the same or opposite trade direction. Also, seemed like a higher success rate if Stochastic was below 50 when price crossed for the buy, and above 50 when price crossed for the sell.

I've taken about a 10 minute look at trendlines based on the fractal indicator... sort of thinking along the Moteki system but applied to 1 hour charts, using the weekly trigger points to jump into a trade if the trendline was favorable, and closing out when the trendline was violated. Would have worked like a charm this week... but that's just 1 week. Food for thought though.

You may be opening Pandora's box here. It seems the system trades very well with the original rules, and nobody should lose sight of that fact. I'm all for optimising, but too often people are looking for something that is 100% correct. That is just never going to happen.

Those are some thoughts that might get you started. There are some far greater minds than mine that visit here. I hope they will chime in.

Oh, and don't forget that Hayseed stopped by here. He has some really interesting tricks he does with his EAs. So, if he eventually decides he wants to work on this, I know he will get some thing going that we'll all find amazing... But I know he is a busy man.

IBFX had price break the trigger line for a long by 1 pip before retreating. Anybody else get in long?

I'm out at 1x profit on the short. That's me for the week. (It came within 2 pips of stopping me out before heading south again).

I have looked at a few more options with the system.

Looked at taking the cross channel trade after hitting the 1x profit. As you had already observed Qman this doesn't look to be very consistent or profitable so dropped that idea.

Also noticed that Jan - April are the most profitable months across all 3 years.
Could be coincidence or seasonal!!! I don't know enough about this subject to comment. I don't think I would consider trading only certain months.

Qman - I would rather not look at introducing indicators. I know they would probably help but I am trying to go "Indicator Free" these days whenever possible. And I don't want to get too involved with fundamentals either.


One thing I didn't look at initially was look at going for a 2x target after the first 1x target failed. (This is a bit more in line with Nomasks original 2x / 3x system).


Results are favorable and consistent across all 3 years.

With 2% account risk. (Compounded)

2005 Increase from 17.93 to 19.83%
2006 Increase from 21.70 to 26.11%
2007 Increase from 46.09 to 61.27%


I'm not going to dick around with this system anymore at this stage. Like you say it looks reasonable sound.

The other thing is I am very conscious that my backtesting skills are at an early stage of development and it is possible I am missing something fundamental here.

So I will demo trade this for now and focus on improving my backtesting knowledge.


Richard