So you want a system that actually works?

[JonnyT]

So far the systems posted here have been pie in the sky.

I'm going to give you details of one for free that does actually work.

The system is as follows:

1) Trades the EUR/USD future on Globex (could use a SB or Forex Broker account

2) Compile hourly charts from 07:00 UK time to 23:00 UK time

3) Enter a trade in the direction of a breakout of an hourly candle

4) Use a stop loss of 30 pips

5) Move the stop to Break Even at 30 pips profit

6) Take a profit at 120 pips or at EOD

Therefor there will not be a trade prior to 08:00 UK time and you will have closed your position at 23:00 UK time.

Using the EUR/USD future on Globex the gross return up to and including the 23/1/4 is 596 pips from 46 trades.

If you use Stop Limit orders then slippage is in the order of 1 pip every 2 completed trades, IB as a broker would cost $4.80 per trade.

At $12.5 per pip the gross is $7450 per contract
Allow IBs costs of 4.8 * 46 = $220
Allow for Slippage 23 * 12.5 = $287.50

Net profit per contract is $6942.5

The worst drawdown is circa 180 pips i.e $2000

I would recommend a minimum $15,000 account to trade this via the Globex future.

This has given a return on account of 46% since the New Year

This blows away any of the systems mentioned here (they haven't worked in backtesting, this one has).

Of course this is very limited backtesting and any trading of this system is your own responsibility. I have merely offered this for discussion.

Here are the trades:
02/01/2004 Long 1.2573 1.2543 -0.0030
02/01/2004 Long 1.2580 1.2564 -0.0016 EOD
05/01/2004 Long 1.2571 1.2656 0.0085 EOD
06/01/2004 Long 1.2662 1.2782 0.0120
06/01/2004 Short 1.2737 1.2698 0.0039 EOD
07/01/2004 Long 1.2708 1.2678 -0.0030
07/01/2004 Short 1.2631 1.2661 -0.0030
07/01/2004 Long 1.2662 1.2632 -0.0030
07/01/2004 Short 1.2627 1.2568 0.0059 EOD
08/01/2004 Long 1.2578 1.2548 -0.0030
08/01/2004 Short 1.2549 1.2579 -0.0030
08/01/2004 Long 1.2618 1.2738 0.0120
08/01/2004 Long 1.2758 1.2728 -0.0030
08/01/2004 Short 1.2736 1.2730 0.0006 EOD
09/01/2004 Short 1.2729 1.2729 0.0000
09/01/2004 Long 1.2731 1.2797 0.0066 EOD
12/01/2004 Long 1.2821 1.2821 0.0000
12/01/2004 Short 1.2804 1.2743 0.0061 EOD
13/01/2004 Short 1.2735 1.2765 -0.0030
13/01/2004 Long 1.2770 1.2740 -0.0030
13/01/2004 Short 1.2732 1.2732 0.0000
13/01/2004 Long 1.2748 1.2718 -0.0030
13/01/2004 Short 1.2707 1.2737 -0.0030
13/01/2004 Long 1.2733 1.2733 0.0000
13/01/2004 Short 1.2714 1.2714 0.0000 EOD
14/01/2004 Short 1.2707 1.2707 0.0000
14/01/2004 Short 1.2693 1.2633 0.0060 EOD
15/01/2004 Long 1.2636 1.2606 -0.0030
15/01/2004 Short 1.2600 1.2630 -0.0030
15/01/2004 Short 1.2619 1.2557 0.0062 EOD
16/01/2004 Short 1.2555 1.2435 0.0120
16/01/2004 Short 1.2389 1.2389 0.0000
16/01/2004 Short 1.2369 1.2360 0.0009 EOD
20/01/2004 Long 1.2363 1.2483 0.0120
20/01/2004 Long 1.2501 1.2601 0.0100 EOD
21/01/2004 Short 1.2569 1.2599 -0.0030
21/01/2004 Short 1.2625 1.2625 0.0000
21/01/2004 Long 1.2629 1.2658 0.0029 EOD
22/01/2004 Short 1.2649 1.2649 0.0000
22/01/2004 Long 1.2707 1.2677 -0.0030
22/01/2004 Short 1.2674 1.2704 -0.0030
22/01/2004 Short 1.2682 1.2712 -0.0030
22/01/2004 Short 1.2696 1.2687 0.0009 EOD
23/01/2004 Long 1.2700 1.2700 0.0000
23/01/2004 Short 1.2656 1.2686 -0.0030
23/01/2004 Short 1.2657 1.2569 0.0088 EOD



JonnyT

 

[FetteredChinos]

Looks good JT..

i presume you have this coded for Automated Trading, or else 7am-11pm is a pretty long day by anyones book..

if so, how reliable is your hardware and software to cope with 16 hour days on a regular basis?


FC

 

[TheBramble]

Well JT, nice post and good detail.

How long have you been running this system?

 

[JonnyT]

Yes I have coded the system.

Yes my Hardware and Software are very very stable.

JonnyT

 

[darrenf]

Ah JT, I've been wondering about your mystery system. Glad youv'e decided to share it with us although I have to say I'm surprised this has generated so little interest so far, particularly in comparison with some of the other non tested, unproved systems banded about.

Anyway, a question if I may.

Does the system enter a trade only on the first breakout of an hourly bar or does it also enter a trade on every subsequent breakout?

ie if first breakout is 8.15am long on breakout of the 7 to 8am bar, and then say at 9.05, you get a further breakout of the 8 - 9am bar, is that just one long entry at 8.15?

Also, if first trade in the day is stopped out/ breakeven, do you take subsequent entries in the same day?

I suspect from the trade log, the answer is that the system continues to take trades throughout the day, but only if the previous one has been stopped out at -30 or breakeven?

Would be grateful if you could confirm

Cheers

Darren

 

[JonnyT]

Hi Darren,

This isn't my main system!!!

I haven't traded this one live yet, the drawdowns on this system are magnitutes higher than my 'live system' but the rewards are about 70% greater. I have developed code to trade it and will probably be live with this one aswell sometime in the next month if the forward testing is positive.

It only allows one trade per Bar. If the trade fails you will not take another trade against that Bar.

HTH

JonnyT

 

[zebra]

Hi JonnyT

I'm not familiar with a couple of the terms you use, can you explain a little further.

What do you mean when you say you compile hourly charts, is this on excel spreadsheet or is this on your trading platform, does this mean you only ever trade on the outbreak on the hourly bar?

When you refer to drawdown c180 pips, is this the maximum sum required to fund the trade? Sorry but I'm not familar with the term drawdown!

Thanks.

Zebra

 

[JonnyT]

1) Yes you only trade on the outbreak of an hourly bar

2) Say you started trading and subseqently built up to 1000 pips profit. Say you then had a bit of a bad streak, some winners, some losers and you now have a net 800 points. (i.e. a drawdown from equity high of 200 points).

So in theory yes the minimum you need to trade a system is the max drawdown + margin per contract. You would be mad to do so.

Personally I trade with at least 5 times max drawdown and cut back the contracts when I am losing.

JonnyT

 

[ColinRiche]

JT

Have you got some hourly OHLC data to hand so we can do some backtesting for ourselves ?

 

[75again]

Hi JonnyT
Firstly – thanks for the idea and I’m sure others will be grateful too. A couple of early questions…

You wrote:

…have developed code to trade it and will probably be live with this one aswell sometime in the next month if the forward testing is positive.

Can you give some general details about this (language, takes signals from where?) and how the API links in with IB?

Zebra wrote

What do you mean when you say you compile hourly charts, is this on excel spreadsheet or is this on your trading platform, does this mean you only ever trade on the outbreak on the hourly bar?

And I’d be interested in a reply to this too.

Cheers Jonny

75

 

[Rubicon River]

Hi,

I would be grateful if you could explain what you constitute as a breakout and how you calculate it.

thnx

 

[donaldduke]

You might want to base your stop and profit target around
the ATR or something based on how volatile the markets are
rarther than magic numbers like 30.

I dont trade forex but im sure they go through periods where
volatility expands and contracts, just like the other markets do..

 

[Crossroads]

This looks like a variation on the common 27/23 forex system. It's been discussed at length on moneytec. Somebody deleted the long running thread over there, but there is plenty more discussion, here is a good place to start for people who want to read up on this type of system http://www.moneytec.com/forums/_showthread/_threadid-5246

__________
Xroads

 

[BBB]

Well done Jonny.

 

[eureka]

Johnny - great title and intro for your thread! - how can you say that something "actually works" if you have not traded or backtested it properly ? i have backtested a zillion systems only to find out they don't work long-term - i guess rather than making claims, i'd ask members of this forum who have time and resources to properly backtest this - it's a faily straightforward system - good luck

 

[Bluewave]

Hi,

I would be grateful if you could explain what you constitute as a breakout and how you calculate it.

thnx

Ditto, how are you determining what is a "breakout"? Are you judging this from previous support/resistance levels, if so, for how many days/weeks back?

It's easy to spot a breakout in hindsight, at the end of the day with the enitre chart before you, but on a candle by candle basis (hourly ones at that!), it's hard.

I would also be interested to know how you're generating trades automatically, because as FC says, this system would mean a long day infront of the screen

 

[JonnyT]

A few replies!

Bluewave: A breakout on hourly candles/bars is very easy. The market simply moves beyond the high or low of the prior hour. You don't have to sit in front of the screen all day as you can set up limit/stop orders with conditional orders attached.
Yes I have written software that can trade this system fully automated without me being present

PS are you the guy who develops systems on www.bluewavetrading.com such as zones?

Eureka: A very subjective question. How far back do you go? Do markets change? People should always but always verify any claims themselves as there have been many systems posted here that do not work. Thats why I put this one up for debate.

Crossroads: I'm aware of the 27/23 system, but this is nothing like it. This is better. The 27/23 took a single trade a day when the market moved 27 pips one way or 23 pips the other from the 07:00 GMT open. I guess they are both breakout systems but thats where the similarity ends.

DonaldDuck: Agreed the figures may not be etched in stone but every Forex system I backtest likes the 30 pip stop for some reason! Seems to be a good pointer whether you have caught the trend or not. Very few systems do not change as time progresses, I knocked this one up in an hour and haven't developed it further as yet. I just used the KISS principle.

75Again: Have a look at the IB website. If you open an account you can download the sample apps in Java, C++ and Excel. There is a discusion forum on IN and a great helpful yahoo group on http:/groups.yahoo.com/groups/TWSAPI

ColinRichie: Drop me a PM and I'll send the Hourly data. However this is tested using 1 minute data not the hourly data as you cannot say that a stop hasn't been triggered from the hourly data as this market can swing abruptly.

Is anyone develops this idea further, carries out more backtesting etc then please share the information with me. Its only fair isn't it?

HTH

JonnyT

 

[Bluewave]

A few replies!

Bluewave: A breakout on hourly candles/bars is very easy. The market simply moves beyond the high or low of the prior hour. You don't have to sit in front of the screen all day as you can set up limit/stop orders with conditional orders attached.
Yes I have written software that can trade this system fully automated without me being present

PS are you the guy who develops systems on www.bluewavetrading.com such as zones?

Dang! There goes my future idea to set up a "Bluewave Trading" site. lol. Alas, no Jonny, I am not the "Randy one" (sorry, couldn't resist!) who runs that company and website. In fact, on quick perusal their systems flummoxed me somewhat - great system names though: "Tidal Wave", "Tsunami" and "Area 51"

Re: Candle breakout. Yes, ok, of course, a breakout must be when the previous hours low/high is breached. But on a typical FOREX market, doesn't that produce a lot of false breakout signals? Why not use a low-period EMA to catch the trends and then stay with the trade until the price jumps the EMA for 2 hourly bars?

Re: limit orders/stop orders. As a user of Finspreads, unfortunately their online system doesn't allow for conditional limits to be set up beforehand. You have to start a trade first, then set the closing stop. I think u have to telephone such orders which is a bit silly when everything else can be done via the website! I understand other SB companies do allow this though, such as City Index and IG Index.

Anyway, at least you are clever enough to have written your own software to trade for you

 

[JonnyT]

Re: Candle breakout. Yes, ok, of course, a breakout must be when the previous hours low/high is breached. But on a typical FOREX market, doesn't that produce a lot of false breakout signals? Why not use a low-period EMA to catch the trends and then stay with the trade until the price jumps the EMA for 2 hourly bars?

That would be a different system. Does it work?

JonnyT

 

[garyp75]

Johnny,

Looks good thanks - inevitably a few questions:

- Why the future and not the underlying? What difference does trading the spot make?

Bluewave I use IG and they don't allow conditional orders online, don't know about City though? D4F does but giving them my money makes me choke...