Another simple breakout system

Hello Merlin and the other forex experts,

I have been lurking in this forum for a while and like the friendly atmosphere.

I am a newbie and have been looking at various systems to trade forex in a mechanical manner without getting my emotions involved. I have been looking at breakout systems, where I can place orders and not look at them for the rest of the day.

Here is one system I have been looking at for use on the Euro USD pair, but would like Merlin or some other expert test it. It would be a great help.

1. The OHLC is based on 5pm EST - the day starts at 5 pm EST and ends at 4:59:59 EST and the OHLC is based on the prices within this time period.

2. After 5 pm EST, calculate the following based on the OHLC of the day's data:

a) If Close < Open then (High + 2 * Low + Close) = X

b) if Close > Open then (2 *High + Low + Close) = X

c) if Close = Open then (High + Low + 2 *Close) = X

Then, calculate
Resistance = X - Low

Support = X - High

3. At 8am EST next morning, check if the price and Hourly RSI (14). If the price is between the Resistance and Support Values, then do the following based on RSI values -

a) If RSI is above 50 place a buy stop (Good for Day Only) 5 pips above Resistance, or
b) If RSI is below 50 place a sell stop (Good for Day Only) 5 pips below Support.

The following are the possible exits:
i) A stop loss of 50 pips
ii) A profit target limit order of 150 pips
iii) Exit at 5 pm EST is neither of the above was hit during the day.

It should be fairly simple, and the projected values are from Demark's work, for people who are familiar with his work.

Thanks a lot,
Julie

I'm a little familiar (and generally quite impressed) with DeMark's work, and recognized this 'pivot' system because I tested it a while ago. Alas, it didn't seem to perform any better (and often worse) than a standard pivot point system. The addition of the RSI rules, however, may help. I'd be interested in what results anybody came up with.

yes this sounds like a spin off of the standard pivot system. i like his use of the RSI, i will try to get it programmed soon and let you know the outcome. i have a few others i have to get at first....

Hi Julie...

I've been lurking as well for the past few weeks...and I'm new to FOREX...having been introduced to it at a "4X Made Easy" seminar...didn't buy the software...but have been studying chart reading and practicing with a GFT Demo Account...

I think I like the system you describe...but I'm confused about the formulas to calculate "X"...for example...let's say a pair's high was 1.5200...low was 1.5100...open was 1.5125...and close was 1.5175...this falls under section 2b...X = (2 * 1.5200 + 1.5100 + 1.5175)...giving X the value of 6.0675...making resistance (6.0675 - 1.5100) = 4.5575...and support (6.0675 - 1.5200) = 4.5475...this can't be right...can it???

It seems that scenarios 2a and 2c would suffer from similar afflictions...am I missing something here???

Thanks in advance...
Tony

If memory serves me right it should be:

Projected High = X/2 - low
Projected Low = x/2 - high

Somebody please correct me if I'm wrong...

good evening traders

hi crazyaboutforex

if anyone would like a look at my interpretation of the above formula in spreadsheet style-- with ammendments to the formula from fxflash52

email tag
[email protected]

I had a spare couple of hours today so I implemented and tested this breakout system. Attached are my results. Its a zipped CSV file (tab separated columns).

My test period is 2003-12-08 - 2004-08-21 with a 2 week gap at the start of February (I don't have tick data for this period). I've used tick data to simulate entries and exits. The times are in GMT (or UTC if you prefer) timezone. So 5pm EST is either 22:00 or 21:00 depending upon whether daylight savings is active or not. The results assume the broker spread is 3 pips for EURUSD. And I've used FXFlash52's updated formula for calculating resistance/support levels.

The results: 561 pips over the nearly 8 month period.

I'd appreciate some feedback on the results and/or confirmation of them. One small observation on the stops. Sometimes the trade stops out at -50 pips, sometimes -51, sometimes -52. This is because the price sometimes gaps 2 or 3 pips (or more) and the stop is activated at the earliest available price.

Cheers,
Patrick

I think you are right. Sorry for the stupid typo.

Patrick,

Thank you for the results. It appears to be what I was expecting. However, I think we will need to test it for a longer time period, to get some reliability as to whether the system can be traded. It is good to keep a system simple for robustness, but a couple of changes may help.

Alternate 1.

How about if we use a stop loss of say 0.4*ATR(10)? The ATR should be based on the daily values. The profit target can similarly be adjusted to say 1.6*ATR(10) to capture the real long range days.

Alternate 2 - This will no longer be a day trade.

Instead of closing out the trades at the end of each day, the following modifications can be made:

a) Use a trailing stop based on the same multiple as the stop loss ie, move the stop up only if the trade shows some profit. Do this at 8 am EST only, but based on the 5pm to 4:59:59 pm data.
b) As for profit taking, use a larger profit target, maybe something like 5 or 6 times the stop loss value, ie, if we are using 0.4*ATR(10) for stop, use 2.5*ATR(10) for the profit target. However, the ATR(10) is the value that was calculated during entry, and should not be updated.
c) If the trade was closed out, then enter at 8am again.

The Alternate 2 will be a longer term trade, and perhaps the ATR multiple need to be increased. I have not demo tested this strategy, so I can't say how it works.

Alternate 3 - Use weekly data for the support/resistance calculations and 4 hour data for RSI. Enter only on Monday mornings and exit on Friday evening.

Just something to mull about, but for newbies like me, Alternate 1 should be a good start. I would really appreciate if Patrick or Merlin or any one else can back test these strategies.

Now questions for Patrick - what software did you use for the test? Do you have any more data, maybe just use 5 min or even 15 data and test for a longer period of time? The system is currently in a drawdown, and it will be interesting to see what is the maximum historical drawdown that can be expected. It is important for all mechanical systems so that the trader can adequately finance her/his account.

Thanks again.

Patrick (pdcruze) wrote:

"My test period is 2003-12-08 - 2004-08-21 with a 2 week gap at the start of February (I don't have tick data for this period). I've used tick data to simulate entries and exits. The times are in GMT (or UTC if you prefer) timezone. So 5pm EST is either 22:00 or 21:00 depending upon whether daylight savings is active or not. The results assume the broker spread is 3 pips for EURUSD. And I've used FXFlash52's updated formula for calculating resistance/support levels.

The results: 561 pips over the nearly 8 month period. "
--------------------------------------------------------------

I received similar results ( 635 pips from feb.3,04 to today)with the setup as decribed in the 1st post and an adr based trailing stop.

However -----

I changed the setup time from 8am est to 10PM London time, eliminated Monday trades and got spectacular results. The set up was the same as above ( 50pip initial stop, 150 pip takeprofit, trailling atr stop, enter 5 pips past sup/res depending on rsi <> 50, max 1 trade per day, close trade at end of day, delete any pending orders at end of day.

From April 26, 2001 to Aug 27, 2004 an average of 494 pips per month. /max dd 259 pips/ 65% winrate/ 2.4 win to loss size ratio .

The most interesting thing was the performance in 2001 & 2 tested better than most of the pivot break systems that I tested for that period. I am sure that tweaking can further improve the performance.

big caveat - I do not have good software for backtesting. I am strongly suspect of anything that backtests well until I have demo tested for long enough to have sufficient live demo data to compare with backtest data for the same time period.

Merlin - I am looking forward to reading your results from a proper test of this simple system!

Flash,

Thank you for taking the time to run the tests. I am glad that you were able to better the system. I see that through our collaborative efforts, we have come across systems (simple GBP system for example) that are superior to commercially available systems. This makes me feel good, and I am sure that it makes all of us participants very satisfied. I have a few questions for you, based on your modifications:

1. What kind of ADR based trailing stop did you use... I mean what multiple and time period and time scale of ADR did you use?

2. If you did not use a trailing stop, what would the result be? I would like to know that, because if I am trading this system, I wouldn't be able to use trailing stops, as I can't follow the markets all the time. The set it and forget it, is the only way I can trade.

3. What is 10pm London time in terms of EST? I think it is 4pm without Daylight Savings, but if this time varies a couple of hours, does that hamper the results?

4. If you elliminate the RSI, and just put in a buy order above the Resistance and a sell order below the support (ie, straddle) and use the same strategy, how is the performance? There is a chance that both the stops are hit the same day, but sometimes, the false breakouts maybe offset. Just a thought that may need some testing.

5. How does the system work for USDCHF, GBPUSD and USDJPY pairs? In my mind, USDCHF should be the best. Usually, breakout systems like this show best performance with USDCHF.

Thanks again for your help.

is this system similar with other system on other thread ( http://www.forexfactory.com/forexfor...read.php?t=760 ) ???

Dave,

I think this system is very similar to the other system you referred to. Only the calculation of the support/resistance levels are different. The daily Hi/Lows, floor pivots etc., are all ways to calculate S&R levels. I chose to use Demark level's because the Floor Trader's Pivots are overused, and probably used by the brokers to run stops. The daily Hi/Lows are too prone to false breakouts, imo.

The purpose of this method is catch those long range days, which occurs not often, but frequently enough, in forex. As with any mechanical system, this system will suffer from drawdowns due to whipsaws. Case in point is the USDJPY pair. I followed that pair for a while, and the BOJ interventions just killed the system. Still, using a basket of pairs with good volatility, adequate capitalization (atleast twice the historical drawdown plus margin), money management and not to forget discipline, these simple systems will outperform the market.

It is my strong belief that less is more, when it comes to trading systems. Maybe getting rid of the RSI and just going with straddles will make the system more robust. It may increase the drawdown, but again, proper capitalization should take care of that.

I hope experts like Merlin et al can backtest the system and find out the historical drawdowns so that we can adequately capitalize our account. And then, the rest is up to the individual trader.

Thanks again.

comments like this tell me you are a mature system developer. how long have you been at it?

First of all, thank you Merlin for the compliments. However, all the knowledge is theoretical... collected from my hours of reading books on trading. That is in addition to other courses and garbage I have glanced over.

I have been trying my hand at discretionary trading, albeit unsuccessfully, on and off for about 4 years. About a year ago, I realized that I need a mechanical system to trade, as it will hopefully help me to keep my own emotions out of trading. I looked hard at commercial systems available, at least reading what the descriptions were, and following forums. When I read a couple of simple breakout systems, like the 2730 or something similar at Moneytec, I came up with this system. Having no way to backtest, I created a spreadsheet and followed a few pairs manually by demo trading for the last few months. I summarized my observations in the above post. The results were encouraging enough to make me think to jump back into real trading. However, I wanted to know the drawdowns to be able to fund my account properly. That is why I decided to post the system here, and ask for your help.

Hopefully, you and the other great guys can help all of us to implement our ideas into trading systems.

Thanks a lot.

you could have fooled me. i too "heard" that the most simple systems are the most robust, but i never really believed it until months after i started developing systems. the more complex, or the more "degrees of freedom" a system has, the worse it is, i dont care what the backtest results show. people throw 6 indicators together and get an incredible backtest result and think, "wow, this must be the holly grail because what are the chances i could have just thrown some indicators around and got this kind of result!". well, the odds are actually pretty good that will happen i never really appreciated that fact until i saw it with my own eyes. but when i saw it, it was like i was reborn.

ouch! you dont know how bad this hurts me when i hear people say this.

dont you realize that what took you a few months to do, you can be doing in less than 10 minutes? there is software written specifically for you to do this! just buy one and learn it!

I just found this. Is this alive!
I am going to program this into metatrader (free). If anybody has already done this please let me know.

Hi, can metatrader backtest system? I want to learn more about backtesting my current strategy.

Yes it can backtest and more.