EA scores 98% but 1 trade blows it all

Hello,

I coded an EA using this website: http://sufx.core.t3-ism.net/ExpertAd...index.html#Var

I ran the strategy tester and the strange thing is that the EA produces almost only winners (it is a scalping strategy). Strange thing is that 1 trade blows the account, also when i put a stoploss in the EA from 100 or 200 pips. My guess is that this trade is the only loser which is not closed for unknown reasons. Please take a look at my report, maybe one of you can see the solution. Of course i am willing to share this strategie but only on PM basis.
tnx !

by the way, this is my first EA, I am a total newbie at programming

It looks like that it only made sell-scalping, without stoploss, and when price go to the low and start to go up it inevitably blow the account

Thats the trade off. No stops or spanish stops (HUGE stops) will kill you every time the market turns. You have to decide when a trade has gone bad and LET IT GO!! - Hold and Hope is not a strategy. It will always blow up (just like martingales) when the market turns against you. Its easy to make a 90+% EA that makes money as long as the market is going your way but its a heartbreaker when 1 trade blows your account or reduces your capital so much that it'll take a ridiculous number of wins to get you back to breakeven.

That said 99% looks pretty good - Test some realistic StopLoss values 10 to 200 step 2 in the optimizer and see if it remains profitable.

I agree but this EA seems not to execute the stoploss. when I turn it on and run the stategy tester, the statement still says usestoploss=false. therefore i guess there is a programming error somewhere

it seems that the problem was caused by a data mismatch. I converted the M1 data to H4 and now I have the result I was expecting. I think it is time to put this EA on a forward test...........

FIRST!! run the test with 90% accuracy - If the test isnt 90% it means NOTHING!!!

90% Setup Info - http://alansforexblog.com/2008/02/27...pert-advisors/
FREE Data - http://www.forextester.com/datasources.html

thanks fulltime, i will do that and post the outcome. The modelling quality N/A was bugging me also. The historical data is/was causing all this, I am pretty sure about that.

Second I will put the EA on a demo account on my VPS to see what happens live.

Old saying "problem shared is half solved already". I suggest that you share it and see how it comes out (Refined or Unnecessary).

Respects.

D8d3

I have followed the steps to get 90% modelling quality. I added several stoplosses and each time the account gets blown. So it goes from holy grail to holy shit.

But at least I know now that the modelling quality is ok, and I know how to build a (simple) EA so I can keep testing strategy ideas. Thanks for the help guys and I will post a new thread if I have something interesting.

HAHA that made me laugh.

Toldja!!! - NEVER EVER trust a backtest thats less that 90% - I've fallen into that hole before!

Trading without a SL gets that kinda result, TP at 100 and some trades take days to come around, others will trend against you never get there and empty your account.

If you long and you keep buying another Lot every 100pips down, recalc your average and set the TP to average +100 that'll generally work better.

BUT I think that 2.0000 to 1.5000 GU sell off woulda still killed you.