Swing momentum trading - the robots are taking over

Here is the next step in my bid to become filthy, stinking rich.


The scooby-doo and Boris Schlossberg threads have attracted some seriously clever people with fantastic trading ideas. Last week, octomanoak directed my attention to this site: http://www.invest-fx.com/forex-advan...se/market-flow


The site set my creative juices flowing. This thread is all about swing trading. My basic premise is this:

 

1. A higher-frame swing high means the trend is down.
2. By switching to a lower time frame and waiting for a swing high. I can sell into the trend after a retrace.

In the opposite direction:

 

1. A higher-frame swing low means the trend is up.
2. By switching to a lower time frame and waiting for a swing low. I buycan into the trend after a retrace.

I use ZigZag (takes longer to establish that a swing has occurred) or ZigZagMe (much quicker) to spot swings on hither time frames, then Fractals on the trading time frame to time the trade entry. Drag a zz amd a frac onto the same chart and you will see why.


For those of you unfamiliar with my work, I am a serial coder of muli-pair, multi-time frame trading robots. So far, there are two in the attached zip:

 

1. Swing trend continuation trader
   1. Uses the ZigZagMe indicator to spot the trend after a high time frame swing. I am using the D1 tf.
   2. Confirms the trend after a lower time frame swing in the same direction. I am using the H4.
   3. Trades an opposite-direction swing on a low time frame. I am using the H1 tf.
   4. Exits:
      1. trade hits an Atr-based take profit or stop loss.
      2. use mptm to micro-manage the trades.

2. Zigzag swing break
   1. Uses the ZigZag indicator to identify the most recent swings high and low of a higher time frame.
   2. Regards a break of either the high or low of the swing as establishing market momentum.
   3. Uses ltf swings to identify trading opportunities.
   4. Exits:
      1. trade hits an Atr-based take profit or stop loss.
      2. use mptm to micro-manage the trades.

There is a third robot. This one is a single pair trader that has made a surprisingly encouraging start on demo:

 

1. Steve Hopwood's Super_Signal auto trading robot
   1. Uses D1 Rsi to establish the trend direction
   2. Uses the Super_Signal indi to catch the latest swing and trigger the trade

   AND

2. squalou's version of the Super_Signal robot: squalou is a fantastic programmer whose skills make mine look infantile. He has recoded a version of SS in his own style and introducing a number of extra features. The robot and details of its use can be found at http://www.forexfactory.com/showpost...&postcount=304 Give this robot a try. It will be better than my version and the only reason I cannot endorse it directly is because I do not understand what he has done.

Each robot has its own user guide.



The engine inside the multi-pair robots is essentially the same. It is also very simple, so new robots can be coded at the first appearance of a good idea.


My code is open-source. Coders can adapt it and pinch it for their own use. The only restriction I place on its use is this: please make any robot you code that includes any of mine, free to traders to download and to use.


Posting restrictions
I have restricted the ability to post in this thread to: traders with at least 1 voucher; traders on my 'buddy' list. I have done this because I do not want the thread cluttered with 'newbie' questions about the basics. New/inexperienced traders, this is not intended as a direct insult. What I want here is discussion with traders who fully understand what it is they are doing; when you newbs have been around a while, you will come to understand what a difference experience makes to the posts a member writes and blanche at the recollection of some of the questions you asked. I do. Go to your profile page; there is a link to the FF page that describes the voucher process.


Traders with experience who have not yet been vouched for, can get around this restriction be being added to my 'buddy' list; all you have to do is pm me to ask me to do this.


For anyone not fully conversant with the idea of swing high/low and Atr can read 'The Swing and Atr' pdf included in the zip.


There are two great threads that new traders might like to explore.

 

1. http://www.forexfactory.com/showthread.php?t=204646 – the Scooby-doo thread. Scooby-doo is a professional trader with a large commercial bank and has been sharing his insights with us. There are some promising robots being tested there; some are being traded live.
2. http://www.forexfactory.com/showthread.php?t=216557 - this is an offshoot of the scoobs thread, dedicated to coding robots that trade strategies posited by Boris Schlossberg; you can read the details in the thread. The 2 MA MACD multi-pair combo robot is the most promising there, although traders are trying to tweak others to make them more successful.

A warning to new traders
Although I describe these robots as 'auto-traders', they should not be treated as fire-and-forget bots. Just because they are robots does not automatically mean they are successful traders. Read, mark, learn and inwardly digest what I have to say now:

 

1. At best, a robot is only 90% as good as the strategy it is trading.
2. If the strategy is rubbish, so is the robot.
3. Until you understand the strategy, you will not be able to trade the robot. You will panic when trades look as though they might go bad, and prematurely close trades that would have turned in your favour had you given them time.
4. You should never trade a robot live until you have:
   1. demo-traded the strategy manually so that you understand it
   2. demo-traded the robot so that you have seen what it can do

5. If you do get to the stage where you are prepared to live trade a robot, do so using tiny lot sizes. Most criminals use different data feeds for their demo and live accounts. A robot may trade like a master on demo, then come completely unstuck live. Live trading is the only true test of a strategy and its trading robot.

Let's see if we can develop some pip-monsters here.

Why use fractals instead of zigzag?

Fractals have the potential for consecutive highs/lows. At least with zigzags, yes they repaint, but only the most recent leg repaints, and as I see it, are slightly more usable.

Only because I have never used them. I will look into it.

For your testing pleasure...

Ahhhhhhhhhh. I see. Yes, much more useful.

Could you do me a favour, please, and save me some time by posting the code to get the latest zigzag hi-lo?

Cheers

Agree 100% I use

ZigAndZag, - I find that one be the best in that matter

Sorry I Forgot to add Indy

And some more, to make sure you not catch fake (re-pain) ZigZag signals

Place it on H4 Add BBand and only take Buy/Sell signals near low

or high Band + Add Parabolic SAR, if you want to get "sure" signals

or otherwise since you get 1 Signal and it gets off Average down/up

next time signal came up. usually don't take more than 2-4 fake ZigZag Signals at most

But, if you wait for P SAR to confirm, then you might don't need average at

all, but that could be not Exact Button/Top..

"Fractals" are useless, besides it's also will take few bars before

Fractal gets activated, compare to ZigAndZag which is immediate

Here is that Example H4 don't need no 15M TF BS for Large Swings:

hi Steve,

there are many ways to get those legs. simple EA below will print Swing High/Low of MT4 standard Zigzag

Congrats on the new thread Steve! Already there are some great contributions.

The above code finds the last three zigzag points which will not repaint.

And I just checked that ZigZagMe

It's not even close as good as ZigAndZag with real Errors and Signals

ZigZagMe just simple zigzag Indy... try mine instead, much better,

Or use both..

zigzagme does show there a regular zigzag would have painted.

Hehe. I think 'stunning' is probably the word.

Cheers guys. I did not expect so much fantastic information in such a short time.

Here in the UK it is 10.00 pm Saturday, in round figures. The wine has been good tonight and I know that wine and programming do not mix.

Sundays are busy teaching days for me, so it will be Monday before I can start to make real sense of all this.

Do not take my upcoming silence as lack of interest, is what I am saying here. I simply did not expect such fantastically helpful responses, otherwise I would have held off starting this thread for a couple of days.

Hokay, guys. Here is a real plea for help.

I do not understand how indicators work.

I especially do not understand how custom indis work.

I mean, I can send a call to the moving average indi and return a result I can work with. Rsi - I am your man. Macd - yawn; is there anyone out there actually not able to send a call via code to macd? Oh dear.

Custom indis? Forget it. You would not believe how many invites to code trading systems I have turned down because I have never yet been able to extract any useful information from them.

This time last week (18th April 2010) I had not met market flow. I knew about it, I simply had not been introduced to/had the brains to realise it is such a simple concept.

Up until Thursday last week (22nd April 2010), I had not met Fractals and was hard-coding swings hi-lo.

Until today (25th April 2010) I had not met ZigZag, except in passing and with no idea what it represented. Apparently there are umpteen variants. I don't care.

So, would someone with far more understanding (or even just a bit more understanding) of of these things please:

 

1. Look at the attached pic. It shows the most recent major swing low, then the previous major swing high. Invaluable information.
2. Then provide the code that will extract from the attached indi:
   1. the prices of these two swings and the candle times at which they occurred.

Given time, I will work this out for myself. Guys, if you are unfamiliar with my work then you have no idea how thick I can be and how long this might take. You will do us all a favour if you can simply provide the code I need.

Cheers

Steve, the code I gave you does that already.

Basically....

 

1. Check bar. If ZigZag value matches high (swing high) or low(swing low)
2. check the next bar to the left.
3. Scan as many bars as needed until you have all the points you need.

from my last code example, z2 show the price of the recent swing (in your pic, it is swing low), and z3 show the price of the previous swing, ie. swing high

code below shows the times the last two swings occurred:

Subscribed!

Hi Steve,

I must admit that your swing momentum auto-trading robot got off a very good start last week. Just one night it caught several good trades with equity up as high as 6%. When trading week ended, it hovers around 4% floating profit. Hope it goes the way it is. Thanks for putting the fractal article together simple and concise. I have never thought of using fractal the way you mentioned. Very interesting concept. Love this thread. Thanks.

Good plan, looking at the way things have started here.

Cheers sangmane. As you can see from the attached pic, I have got your code up and running.

This is all very interesting and most enlightening. It is well worthwhile dragging both a fractals and a zigzag onto a chart to see the difference in the power of the two indis.

I just spent a few minutes running strategy tester in visual mode, just to watch zz's behaviour. It takes a few candles to establish a swing. Once established, the momentum away from that swing looks strong. A swing high, for example, marks the start of a trend downwards that can last quite some while.

I am thinking it might be useful to combine zz and frac. Take the Trend Continuation robot:

 

1. upper tf (D1) zz finds a swing high to establish that there is a down trend in place.
2. medium tf (H4) finds a swing high to confirm the down trend
3. fractals on the trading tf (H1) finds a swing high that triggers the sell

It will not take long to adapt the robot, so I will try to do this tonight in time for the markets opening and we will be able to play with it on demo.