New EA Auto Trading Portfolio (Free EA, No Expiration)

[QUOTE = "Jean-Leon, pós: 258.674, membro: 44967"] Hi autotradingportfolio,
Você pode obter todos os dados a partir de 2000, relativo TickStory (por exemplo).
Considera [/ CITAÇÕES]

I download my historical data from SQ tick downloader, it's the same company that does QUANT ANALYZER, the program is free.
it's much easier to download the historical

 

We hope steady thes years is very bad performance why

 

this is a good book, to get better results from your EA without there being a super optimization

 

any thing regarding the volume trading low in the years 2000-2006 and the last mouth OCT 2018 why loss please find the reason

 

Having a look at the numbers the "loss" is still small compared to the gains you would accumulate in the 10 years after. So having a proper money management in mind this whole drawdown period - although it is pretty long I have to admit - does not turn the strategy into a not working one.

That does not mean, that considering a further look into it in terms of further optimization would not make sense, but no matter how large your sample perios is, there will always be an out of sample period, that could lead to different results.

 

I do not think it is that different, he made a period of 11 years from 2007 to 2018 and the average of openings was 4000 trades per change and I took a period of 6 years from 2000 to 2006 and the average trades is of 2300 per exchange

 

in this period from 2007 to 2018 he did an optimization with thousands of back tests, and passed the best result, you do not see the negative results, this program does not allow adjustments in the parameters of the 20 strategies that are in their internal, we can help to improve!

What I did was to simply use the program in a period that had not been optimized.

and that was the result

 

In which way does this invalidate my statement? I am aware of the optimization process, do not worry. But it does not make any difference, as even with that bad out of sample period it remains an overall profitable strategy.

Besides that, taking timeframes into account too far into the past results in a reduction of the possibility for a meaningful comparison. 20 years ago algotrading was nearly non existent, volumes only a joke compared to nowadays and retail trading was as well non existent. Do you want to concentrate your trading logic on the behavior of quotes beeing that fundamentally different?

So an optimization on timeframes that far away should not be the main focus. Which leads back to the out of sample "problem" again..

 

Of course, the most significant risk in any trading strategy development is “over-optimization,” also referred to as “curve fitting.”
Over-optimization will result in trading strategies where the historical performance is vastly superior to subsequent live trading.

 

Thank you sir for back test but we can but separate paries doning well please

 

the separate back tests you can do alone, is to download the history and use the program ATP

 

I see.., thanks for the information.

 

Hi Heltoneli

Both SQ Tick Downloader and Tickstory only have data from 2003 both are dukascopy data, the only sources I know that have older data is forexite.com and histdata.com.

So how did you get the data from 2000?

Best regards.

 

To have a greater history, I am using Activtrades

 

Thanks for your information

 

Mr auto trading portfolio you must stady OCT 2018 when the first manth loos after 12 very bad signal I mean 2000-2006 come again please adviso

 

Dear autotradingportfo

Could you please make available to change the sl and tp? In that way we could help you in the optimisation. It's a very promising ea anyway, although it made me losses only so far. I think to change the tp/sl could help to avoid the losses.
Thanks for your hard work.

 

You guys still do not get it.

You are running the EA on a truly large account, that is why I am answering you in a serious way since you are doing it right (unlike most here).

Why should "changing" these static values change anything for good in the long run? The current setting is the one which produced the impressive results we can backtest until today.

So adapting blindly to short time market behaviour just because you had some "bad month" will most likely lead to new/worse results later on - where maybe the current setting will outperform again. It just do not makes sense. THAT would be overoptimization and curve fitting, which someone recently mentioned. Not the "building" of a setting based on a 12 year timeframe!
Losses are part of the game, we have to accept that. Even in the worst case scenario of having a drawdown in the upcoming six years (when having 2000-2006 in mind) you would be a wealthy man just two years later on (when having 2007-2009 in mind afterwards) INLCUDING the losses of the DD period.

The only problem is the person in front of the screen - impatient, greedy and emotional. Take care of the moneymanagement, which should be no isse having your account size and I highly guarantee you will be more than happy in 10 years from now with a risk as low you will hardly find in any other EA around here.

If you want a quicker way, use more risky grid/hedge/martingale approaches. Simple as that.

I am running the EA on a comparable account size and relax..

 

This was another test done with the ATP version v2.9, fixed size of 0.01, in a period of August 1 until today November 9, where an Over-optimized was not done, to know what the results would be.

I am doing these tests to be able to help with the development, not to criticize, and together we can find solutions for improvements, and to know what the true results of this EA will be.

 

Sure no problem.

So if you have historical data from 2000 to 2018, will you try to backtest my next EA?