Simple DAX Strategy???

[cd173]

Hi All

Now would this work???
Watch the DAX for the first hour then buy/sell if it breaks out/down of the range after this time.
I'm thinking 20 to 30 point target, but unsure of the stop loss which might be 5-10 point range??
Had a glance back over the last couple of weeks and looked a possiblility, but not sure how it would pan out longer term?
Does anyone use a similar system or could kindly provide some backtest data maybe?

All comments welcomed!!

Many thanks
C

 

[Roberto]

I've been trading the Dax on and off for about 4 years and, to my shame, can't give you the facts and figures that would answer your quesitons. I can offer some impressions, though: I suspect that you might be better looking at the first 2 hours rather than the first hour (would doubtless reduce the number of trades but greatly increase the success-rate), not have a fixed target thereby depriving yourself of the chance to catch a big move (seems counter-intuitive to me, anyway), and using a slightly wider stop-loss than you suggest because the Dax is certainly noisier than, say, the FTSE-100. I would instinctively want to look at a 2-hour breakout system with a 20-point stop and let profits run until a moving average crossover and/or a divergence in the MACD and/or the RSI or whatever suggested otherwise.

 

[Roberto]

Having thought a bit more and looked at hourly (and half-hourly) candlesticks for the last couple of months of the Dax, I realise that what I've suggested above is such a completely different system from what you were asking about that arguably there's little point in comparing the two ideas. Specifically, using a 2-hour breakout rather than a 1-hour breakout means that on several days one would be shorting rather going long, and vice-versa. All of which perhaps says something about the arbitrary nature of some of these systems.

Nevertheless, both the "systems" proposed above appear to show healthily positive returns over the last couple of months. A 30-point target is very rarely achieved anyway, as it turns out (not that 2 months' back-testing is sufficient to prove this point, obviously!). But there are some occasions when a potential 100+ point move could have been caught without the arbitrary close.

I've looked at all the usual things I currently display on charts for intraday trading (which are too many!) and still believe that using TA signals to close would be the way forward - but of course that probably says more about the system-tester than the system. Something so subjective makes back-testing cumbersome (to put it mildly) so I should leave others with better facilities/information to answer your interesting question! But thanks for providing the stimulus for my 100th post here

 

[cd173]

Hi Roberto

Many thanks for the feedback. It was just an idea that popped when I remembered that the DAX will generally go in a particular direction after 9am. Having a quick look back, it seemed a positive strategy. I think a 10 point stop would be enough with a 20-25 point target. Trailing stops are v difficult for this sort of thing so as you suggest an MACD reversal may be a good exit point. Perhaps I'll try with a breakeven stop at 10points gain!
I'll probably let this run for the next couple of months to get a fuller picture of the results/stops etc and post the P/L. I'll also have more DAX data by that time!
Happy 100th post BTW!

Cheers
Chris

 

[Roberto]

I'll look at it in more detail when I have time (probably over the next few days) and see if I can produce some back-testing results for entering the first hour breakout (a) with a 20-point target and a fixed 10-point stoploss, and (b) without a specific target but exiting at the first MACD reversal, also with a 10-point stoploss. I suspect that these systems can sometimes have quite good results over a specific short period, but then lose their shirt if you run them indefinitely; but it might be interesting anyway.

 

[cd173]

Hi Roberto

Thanks for that. I intend to do the same this weekend. Probably go back 3 months and see if I can generate a 25 point target with a 10 point (trail maybe?) stop. I remember being told quite a few months back that the DAX has this particular characteristic, so it might be more robust than you think.
We'll have to wait and see!!

Thanks
Chris

 

[Roberto]

We'll have to wait and see!!

Hi Chris,
Very interesting. Ok, we'll see what we come up with. Certainly worth knowing about if there's anything in it.
Roberto

 

[cd173]

And the results are...............

Well, nothing too spectacular, but profitable.

I used a system with a 20 point target and a 12 point stop. After a gain of around 12 points I moved the stop to breakeven. I only have 3 months worth of back data to work with, so results posted with caution. Not including slippage or commissions, it ends up with around 80 points per month. However, the worst draw down was around 47 points (around £800) and I'm not sure I could stomach that!
Not doubt this system could be improved upon!
Results posted in attachment.

Roberto, don't know if you came up with anything better or more consistent??

BTW, if anyone has any DAX intraday data I could import, I'd be eternally grateful!! :cheesy:

Cheers
C

 

[Roberto]

Very interesting ... many thanks. Will try to post something along the lines I mentioned later today.

 

[Roberto]

A disappointing update: I'm afraid it turns out that I can't, after all, produce the figures I wanted to produce. I'm so sorry. I realised only this afternoon when I sat down to start work on this! The reason is that I've just switched datafeed suppliers (to ProRealTime, with which I'm generally delighted, by the way) and I have very much less backdated information available with it than I'd realised.

I did manage to look at the last one month of the Dax on an hour chart (not that this is going to prove much) and the system was certainly profitable without setting an exit target other than as dictated by the first occurrence of any the following parameters: a clear change in direction of the 6-hour EMA; a clear change in direction of a standard MACD; or an obviously dangerous candlestick (such as "bearish engulfing" in a long position, or three birds shooting-up, or three ducks sinking or whatever the damn things are called).

Specifically (and here are the two observations that might actually be helpful rather than merely apologetic or explicatory):-

1. There were a couple of days with very big profits when the trade could remain open all day, even with trailing a stop-loss ("trend days" missed by a specific target closure);

2. There were a few days on which a "stop-and-reversal" would have been very profitable and, I think, no days at all on which it would have been expensive.

It seems clear to me that further research may be worthwhile, and at the very least, there's something in your original theory!

I'd suggest allowing a maximum of one stop-and-reversal if the first trade fails and there's a subsequent breakout in the opposite direction from the first hour's range.

I wish I had better backtesting facilities!

 

[cd173]

Hi Roberto

Thanks for the update. There's food for thought in your exit strategy as the target was ending trades early on good trending days! I'll probably run a test using your method, so might need to PM you for more specific details if that's ok?
If you need some DAX back data, I've got a couple of months, so PM me if required.

I'm looking into a similar strategy for the ES Mini with a 3.30pm break, so I'll see how it works out and post the results.

I'd be interested to know if anyone else has anything to add??

Cheers
Chris

 

[Roberto]

might need to PM you for more specific details if that's ok?

Certainly, with pleasure. Look forward to hearing from you, and sorry not to have been of more use!

 

[dod]

I traded a Dax breakout system. I used a breakout of the first 40 minutes, with a 5 point stop. I had various ways of exiting, usually at a target of 15 points, bringing my stop to breakeven once I was 5 points up, and to 5 once I was 10 points up. Other times, I simply took a 5 point target for a 5 point risk. I also used the break of a EMA to exit. With hindsight, the best results were obtained from risking 5 points with a target of 15 and not moving the stop at all; but I found I couldn't do that! I only took the breakout trade if ADX was above 30. For a while, I also took the opposite view when ADX was below 20, trading against the breakout suceeding, with the same 5 point risk/reward.

This was highly successful for several months, before starting to tail off earlier this year, when I packed it in. I haven't checked recently.

I chose the forty minute timescale because it seemed to me that so often the Dax would take off shortly after 8.40.

 

[weyhey]

I found with break systems like you mention that you need the market to be having wide ranging days. If you trade these intra day systems when ranges are low you will not get the big winners you need for this type of strategy to work.

From my testing they are much more profitable when the Average True Range of an indicie is extreme. For example the 10 day rolling Average True Range of the FTSE at the moment is 45 points, from my testing a break out system based off the opening or first x hours range works better when the ATR is above 105. This way you will have days when you will capture big winners to offset you numerous small losses.

Hope this helps.

 

[dod]

That's an interesting observation. Presumably, it would apply to stocks as much as indices. A while back, I set my software to work out what I call a 'percentage range', so I can compare like to like. After all, a big range on a high priced share can compare unfavourably to a modest seeming range on a lower priced one. I base this on ATR divided by price multiplied by 100.

When was the FTSE range last above 105?

 

[weyhey]

Yes you can apply it to any market.
I used a percentage of ATR, so in the case I mentioned when the FTSE ATR was above 105, I would buy/sell on a breakout of 40% or more of the ATR value. Anything less than 40% was just noise. Before the open I would get the ATR for the previous 10 days, say 109, I would calculate 40% of 109, 43.6 I would the get the open of the FTSE add/subtract 43.6 and these would be my stop in points
My stop would be one point above/below the open depending if i was short/long. And I would only risk 1%.
I would have a trailing stop and exit on the close. Another exit would be to exit on the next days open if profitable.

But as I said it only really worked when ranges were wide.
You would need to test each market/stock as to what is the best time to trade such a system.
The only time I would trade such a sytem all the timn is if I had REALLY deep pockets because you can go through months of not being profitable. I bet people who trade this system have had huge drawdowns lately.

The last time the FTSE ATR was above 105 was in March 03, then only for 2 or 3 months.

 

[dod]

In your example, was your trailing stop also 43.6?

How did this work out in realtime?

It would be nice if 40% was generic, rather than specific to the FTSE. Or maybe it applies across the board.

It's a neat looking calculation, anyway. I'll definitely look into it.

What was the level of the FTSE during those two months in 2003 when its ATR was above 105? I ask because i want to work out the 'percentage range'.

 

[weyhey]

The stop I actually used was a percentage of the 40% of ATR, I used 60% of the value of 43.6, so my stop would have been 26.6 away from my entry level (rounded to the nearest 0.5).

As I said in real time it was awful when markets had small daily ranges and excellent when markets had wide ranging days.

The FTSE was around 3800 -4000 in March 03

 

[dod]

That ATR of 105 when the FTSE was 3900ish makes a 'percentage range' of 2.69 compared to an ATR now of 45 with the FTSE at 4700 which comes out as 0.95.

This suggests that any underlying market that has an 'percentage range' of 2.7 or higher might be a good candidate for such an approach

Anyway, it's all food for thought, so thanks.

(Later)

Thinking about what you said, I realise I haven't necessarily understood it. Do you mean you

1. Added 40% of the ATR to the open, and you also subtracted 40% from the open, and you then used the outer edges of this '80% band' as entry points.
2. Or you added/subtracted 20% of the ADR to both sides of the open, resulting in a '40% band', whose edges you traded from.
3. Or you had a 'moving 40% band' that shifted as new highs or lows were made and that signalled a trade when price moved out of the 40% range?

 

[LION63]

CD173 and Roberto,

I presume that you are still testing and/or trading your DAX theory, how has it panned out so far considering you have had a month to try it in real time?