stevehopwoodforex.com
https://www.stevehopwoodforex.com/phpBB3/
Print view

My first EA, interpreting results
https://www.stevehopwoodforex.com/phpBB3/viewtopic.php?t=2749		 Page 1 of 3
Author:  aSushiRoll [ Tue Jul 23, 2013 2:39 am ]
Post subject:  My first EA, interpreting results
Hey guys,

First: I know about backtesting, it's not reliable etc. however I do want to see how my program performs and so I let my indi show what trades it would have taken..

I just coded my first EA.. well it's not really an EA yet, it's still in indicator form, but it already shows all trades that it would have taken in the past 5 months and calculated profit&loss (without spread, so we subtract about 700 pips in the end from each of the lines shown below).

Below are the results for each symbol for the 30 min & 1 hr charts.

My question is this:
When looking at this.. most lines are pretty much > 0, even accounting for spread (I will add that later in the calculations), which means they run a profit. Just a few are under or around 0 (BE).

How should I proceed from here?
1) I could just create an EA from this, which runs ALL symbols therefor averaging out the positives with the negatives, I should get a pretty straigth line up (I'll add calculation & chart tomorrow, I'm tired as it's 04:36 AM now)
2) I could see on which pairs it performs best and have the EA run on those. Except.. a backtest is in no way an indicator of how it will perform in the future on those pairs.... correct?

Just wondering how you guys ussualy deal with this. I remember reading in various threads about EA's that you should only run those on XXX or YYY pairs, because it performs better on those. Isn't that using "backtest" data to predict the future performance? Even if you did a live forward test and you see which pairs are bad for that EA.. at that moment the results are already "old" and you still don't know how it will perform in the future on those pairs.

P.S. the scales whow pips on the vertical axis and the horizontal axis is just the number of trades (30 min chart about 5000 bars, so 5 months, 1 HR chart about 10 months)

Sushi
Author:  kwanann [ Tue Jul 23, 2013 3:21 am ]
Post subject:  Re: My first EA, interpretting results
So you want to convert this to an ea that does trading? Or what you looking to achieve
Author:  phil_trade [ Tue Jul 23, 2013 6:12 am ]
Post subject:  Re: My first EA, interpretting results
Hi

Your strategy seems to return good result ! Congrats !

You need to confirm with a good EA (without bugs, quick Open/Close trade, dealing with high spread, poor liquidity period etc) and a very good Money Management.

Problem : Empty4 can't backtest multi-pairs. So difficult to appreciate risk/money management. That's why I have learn Delphi and write my EA first with forextester, before rewriting them for Windev/Api/Empty4.

If you want to only use Empty4, I suggest you to write it for mono pairs, and check if backtest return good result as your indicator. Then select the best pairs, run EA for each pairs one month Live, then rerun backtest on the same month, to confirm. You will not with this technic appreciate if multi pairs money management is OK but you could approach it, by increasing single pairs Lot, to simulate multi-pairs.

Could you explain or publish your code ? a good strategy is always interesting to learn :)

cheers

Philippe
Author:  kwanann [ Tue Jul 23, 2013 6:38 am ]
Post subject:  Re: My first EA, interpretting results
Phil,

Can direct me to where i can find out more info about windev/api/Empty4?

does this work only with delphi or most high level languages will be able to work with this?
Author:  phil_trade [ Tue Jul 23, 2013 7:05 am ]
Post subject:  Re: My first EA, interpretting results
kwanann wrote:Phil,

Can direct me to where i can find out more info about windev/api/Empty4?

does this work only with delphi or most high level languages will be able to work with this?
Delphi or C# is for ForexTester http://www.forextester.com/

API for Empty4 is a .NET interface for all high level language : http://www.mt4api.net/

Windev is a high SQL level language : http://www.pcsoft.fr/

Philippe
Author:  kwanann [ Tue Jul 23, 2013 7:13 am ]
Post subject:  Re: My first EA, interpretting results
Ahhh interesting, didnt know forextester allows C# integration. Thanks for the info Philippe

K
Author:  phil_trade [ Tue Jul 23, 2013 7:24 am ]
Post subject:  Re: My first EA, interpretting results
kwanann wrote:Ahhh interesting, didnt know forextester allows C# integration. Thanks for the info Philippe

K
Yes, I'm using Embarcadero Delphi XE2, but I know some users using Visual studio C#
Author:  aSushiRoll [ Tue Jul 23, 2013 9:12 am ]
Post subject:  Re: My first EA, interpretting results
So you're saying choose the best pairs..

But isn't that the entire reason why steve doesn't want us to post backtest results? Because you can't assume that the way a certain pair behaved in the past will happen in the future.. ?
Author:  mjws00 [ Tue Jul 23, 2013 9:39 am ]
Post subject:  Re: My first EA, interpretting results
The problems with backtesting mostly relate to Empty4 and some of it's weaknesses. It isn't that the logic, or practice itself is flawed. All strategies have to make assumptions about the future based on what has happened in the past. The problem comes in where people have bad data, poor modeling, and no idea that their results are garbage. It is possible to be quite rigorous with Empty4 testing.

Forward testing is great, but you it is nice to be able to crank out a basic model and do some testing and basic optimization before committing that much time. Especially if you are working on something reasonably complicated or innovative.

Good luck.

Mike
Author:  aSushiRoll [ Tue Jul 23, 2013 9:40 am ]
Post subject:  Re: My first EA, interpretting results
Wow, sleeping for a few hours increases your productivity by about 600% :) Anyway I added a 3 pip spread per trade and added all the pairs together, as if I had a basket of all pairs concurrently (I didn't use the actual time of the trades to do this, just added them by trade#, so it's an aproximation and at any point in time equity could be down more than shown, if there's only 1 bad pair trading at that time). Over the long run these are the results however.

The curve looks promising, however like I said and I will state it again (just a few lines later lol) that this equity curve would NOT be so, only when all pairs have open trades concurrently. If there is at a certain time only 1 bad pair trading we could go into DD.
All times are UTC Page 1 of 3

Powered by phpBB® Forum Software © phpBB Limited