Digital ACSTrend - page 33

 
John Last:
Well,

There are problems with the SSA end-pointed when you try to use for EA. But it is beyond the problem of repainting. It is something different. And unfortunately it is beyond my technical competence to fix that, even some very competent guys tried and did not make it work until now.

However I am not sure if it is practical to use the SSA ep into an expert because you could not do extensive testings (very heavy) but you can adapt very well to the local conditions.

yep - running a simulator is painful. really painful.

running backtests will make you cry with it, then when you look back and try to correct where it went wrong - even with EP what it shows on the chart is not what the EA reads.

it is only through sheer chance i got anything working im afraid. there is no method to success with ssa.

 

SSA ep and EA

Meanwhile do not attempt to do EA based on SSA ep.

However it is good for manual trading. And on this thread my objective was not to develop an EA, or to ask someone to do that. However thank you for the time and effort.

However a Brain Trend expert was developed with the double smoothed jurik of Mladen. This algorithm is reliable for EA. The logic was exactly Stop and Reverse from signal to signal.

I will not mention the daily where the profit ratio was arround 8 (for the last year), but you can achieve that with every MA based expert. It is consistent for the 30 m frame. The percetage of the profitable trades is arround 30 (most of the time, in this example it is higher), which is low but almost OK for trend following system. The problem is that most of the loosing trades were at at least 30 pips profit before they become loosing trades.

The good thing is that result on the 30 m was achieved with the default settings. And I achieve a plateau of good solutions showing stable results, so the logic looks robust enough. However it was tested on the EURUSD only. A quick check showed that it does not work as well on the GBP/USD.

So meanwhile. Those BT are great for entries but the exit is still troublesome for me. So manual trading can help with this in order to appreciate in situ to hold or to exit.

Files:
30m.rar  27 kb
daily.rar  10 kb
 

Hi

Well

The results are interesting

This is the beginning w/ this kind of indicators

I think we will go ahead soon

Rgds

 

Well, this is an interesting development. The key thing is making it constant.

Ovbiously, much harder said than done.

 
So manual trading can help with this in order to appreciate in situ to hold or to exit.

Hi John Last,

yes I agree.

because you posted backtesting results with modelling quality: n/a (not available). It may be good to see some manual trading results for at least few days (or for 1 day). In this case - this development will move forward in more speed way.

Just my personal experience.

 
John Last:
Meanwhile do not attempt to do EA based on SSA ep.

However it is good for manual trading. And on this thread my objective was not to develop an EA, or to ask someone to do that. However thank you for the time and effort.

However a Brain Trend expert was developed with the double smoothed jurik of Mladen. This algorithm is reliable for EA. The logic was exactly Stop and Reverse from signal to signal.

I will not mention the daily where the profit ratio was arround 8 (for the last year), but you can achieve that with every MA based expert. It is consistent for the 30 m frame. The percetage of the profitable trades is arround 30 (most of the time, in this example it is higher), which is low but almost OK for trend following system. The problem is that most of the loosing trades were at at least 30 pips profit before they become loosing trades.

The good thing is that result on the 30 m was achieved with the default settings. And I achieve a plateau of good solutions showing stable results, so the logic looks robust enough. However it was tested on the EURUSD only. A quick check showed that it does not work as well on the GBP/USD.

So meanwhile. Those BT are great for entries but the exit is still troublesome for me. So manual trading can help with this in order to appreciate in situ to hold or to exit.

cool. will look at these tonite.

if it has profit ratio of 8 for daily then i will just use that & if there is an MA based EA that achieves the same can you direct me to it? i already have such things as the universal MA ea...

 

Hi,

Yes I use only open prices. As the signal is definitive only when the bar is closed, that is why I am not interested in any intra-bar dynamics. However we still work on our prototype.

The interesting part of the story was that on the back test all the combinations were profitable with very similar equity curve: length from 30 to 100 with step of 10. That makes me think that, this may be interesting for further developments. Unfortunately I am not an owner of the code, so I cannot reveal it. We just discuss the principle and the basic so everybody can do it.

 
John Last:
Hi,

Yes I use only open prices. As the signal is definitive only when the bar is closed, that is why I am not interested in any intra-bar dynamics. However we still work on our prototype.

The interesting part of the story was that on the back test all the combinations were profitable with very similar equity curve: length from 30 to 100 with step of 10. That makes me think that, this may be interesting for further developments. Unfortunately I am not an owner of the code, so I cannot reveal it. We just discuss the principle and the basic so everybody can do it.

john, have you got some ok data? Tick data - 99% modeling quality with Metatrader 4 | Birt's EA review

do you use the optimization in mt4? i hear mt5 is much better for this - i must investigate.

so far nothing profitable with jurik - will go up the tf's.... ssa is still playing havoc and am looking for alternatives in each creation - ie normal bands instead of ssa bands...

 

Specific

Please be aware that those are only instruments and you are fully responsible for your actions.

If we make general statements, this or that does not work that leads to nowhere. Please let all be specific what exactly did not work.

The testing is something very specific issue, thanks for the information and the links.

Here are the results from 2007 with length 70.

I add the results with length 50 too.

This strategy looks robust, but is far, far away from money machine.

From the test I learned that daily frame is good and 30 frame is good for entry.

1h looses money on almost all possible combinations

4h looses money on all combinations

Files:
 
John Last:
Please be aware that those are only instruments and you are fully responsible for your actions.

If we make general statements, this or that does not work that leads to nowhere. Please let all be specific what exactly did not work.

The testing is something very specific issue, thanks for the information and the links.

Here are the results from 2007 with length 70.

I add the results with length 50 too.

This strategy looks robust, but is far, far away from money machine.

From the test I learned that daily frame is good and 30 frame is good for entry.

1h looses money on almost all possible combinations

4h looses money on all combinations

hold up, these are still general statements on data we dont know is reliable. These are backtests.

The whole thread is mostly ssa and ive just checked my code - its totally random (to me) code for the stuff that works - the ea is not doing what i programmed. Exactly's mean sweet FA.

the more measured you get the further away you really are.

its either a daily with 30 min entry on a mtf combination or not - so what is it?

40 pages of exactness?