EAs tested

[finimej]

My requirement for a good EA, relative and max draw down shall less than 18%, open floating position shall be less than 18% equity drawdown, and profits factor > 1.5.

If it is a scalper, then one loser shall only wiped out 2 winners. that's means stoploss = 10 pips, takeprofits = 5 pips. if takeprofits =3 pips, then hardstoploss = 6 pips. and profits factor > 1.5.

 

FAPturbo and magardroid do not satisfy me, because one loser wiped out 10 winners. Yourlucky or Thunder4WD has one losser that wiped out 7 winners. can not pass my standard also. Robominer, it has floating position loss more than 20%, does not pass my standard either.

 

EAs I have backtested and do not satify this standard,

Robominder

Neural Scalper EURUSD M5

Magic Machine

FAPturbo

Magadroid

Thunder4WD (all versions, include march 2009 version)

YourLucky (all versions, include march 2009 version)

Euro Blaster

Intelectual Pro

Bogie-nn

Ilan

Stomper

Xterme

 

 

 

If we are not able to find a such good EA, then maybe we can tell what EA do not work, so we do not waste time on test all the EAs.

 

http://ckowyong.com/blog/statements/

here is one guy who has tested some EA.

[APopovA]

Re: EAs tested

 

It is a meaningless statement without settings. You can easily change any of the EA's setting so that one looser will wipe out less then 1 winner, but you will end up having less winners etc.

Enterra Fx Star 3.1 is nothing else than nicely optimized FAP as far as I know, and it makes all the difference-- optimization.

 

I have another suggestion, thou, close to yours. Sometimes I can see a very good EA that is lacking something little, some feature that could improve its performance. I'll post some examples:

1.FAP has a room to improve exit strategy. Maybe instead of simply resetting TP at certain time, it would be better to add trail at a negative profit value growing each new time increment. So often I've seen after a few hours a loosing position climbs back almost to break even just to slide down and hit the stop after...

2. FAP could be using 2 or even more trading time periods instead of one.

3. FAP cold be trading a second position as Aldha, while the first is waiting to exit in minus. No reason to waste good trading opportunities simply because we got a bad position hanging.

 

I know that you guys have a lot of talented programmers among you(William comes to mind for example). So maybe instead of simply trying and throwing away potentially good EA's we could together improve them to our liking?

It all starts with positive critics and ideas. What do you think about it guys?

[finimej]

Re: EAs tested

 

There is EA whos logic is wrong, then no settings will save it. Like Magic machine, the money management and order management code is same as rubbish. also it has a good logic, it failed.

 

Why all the scalper EA failed? Because it failed on trend session.

Why all the trend following EA failed? Because it failed on range session.

 

How to detect it is trend in a baby phase? No EA on the market implemented technique to do this.

 

Thunder4wd and Your_lucky recent version is the only EA that tried to do this.

1) time stoploss, the longer time to getting to profits, the less pips to takeprofits.

It tried to use the time stoploss, however, it has bug in the code, so it is not succeed in what it intend to do.

2) indicator damiani

It has also a function called SideWay() and tried to determin the sideway, however, the Damiani multiframe does not give the EA what EA what. Because if you trade on the current M5 bar, and use M30 or H1 do determin the value, then you have to wait until current H1 finished and get the right indicator value, otherwise the MTF is repainting for your little M5 bar.

3) Waddah Attar explosion indicator

It this time uses the WadaAttah indicator, it is the correctway to go.

4) I-trend <0.0017, it is also good way to use it.

 

But the EA failed on the condition, when it has very low volatility and gradually falling or has falling bias, it keeps add buy position, and all turned to big losser.

 

What thunder4wd and your-Lucky lack is the hedge function when things go wrong. Because EURGBP always range trade back anyhow, hedge will allow the lost get under controll, and get the both hedged position come out with profits.

 

Thunder4wd and your_lucky lack also a good stoploss and trail. It is many time, we see it above break even, and not reach to target yet, then it allow the position to get down under the break even, and turned into a losser.

 

All EA get losses, the first I will check it how it handle the stoploss and trail.

 

There are good stoploss and trail code. no problem.

The only thing that we need is: a good winning logic, and how to dected the trend and range session, and avoid / or catch news time.

[APopovA]

Re: EAs tested

 

Now you're talking!!

I do not think that it is possible to predict or even recognize a baby trend before committing to transaction, but we could start from adding the trailing to FAP. Why I keep suggesting FAP: it took a lot from rdb's and nobody can argue that it is good enough Lucky's ripoff to take as a basis, after all it truly multiplied by 10 my account in 2 months.

They took the same logic as in Thunder(4 time periods versus Thunder's 5) and simply would move the TP 4 times gradually down after 4 periods of time. What I would do is to add a separate trailing point trigger for each of these periods. This way we will not see again what you described : when it would reach breakeven and then hit a stop loss.

void NoiseFilter() {
  int li_12;
  double ld_16;
  double ld_24;
  int li_32;
  double ld_36;
  int li_unused_0 = MarketInfo(g_symbol_928, MODE_SPREAD);
  int li_4 = OrdersTotal() - 1;
  for (int l_pos_8 = li_4; l_pos_8 >= 0; l_pos_8--) {
     if (!OrderSelect(l_pos_8, SELECT_BY_POS, MODE_TRADES)) {
        if (WriteDebugLog) Print("NoiseFilter: OrderSelect() error = ", GetLastError());
     } else {
        if (OrderMagicNumber() == g_magic_832) {
           if (OrderType() <= OP_SELL) {
              if (OrderSymbol() == g_symbol_928) {
                 li_12 = g_datetime_964 - OrderOpenTime();
                 if (li_12 > gi_1044) {
                    ld_16 = NormalizeDouble(OrderClosePrice(), Digits);
                    ld_24 = NormalizeDouble(OrderOpenPrice(), Digits);
                    li_32 = 0;
                    if (OrderType() == OP_BUY) ld_36 = ld_16 - ld_24;
                    else ld_36 = ld_24 - ld_16;
                    if (li_12 < gi_1048 && ld_36 >= gd_1096) li_32 = 1;
                    else {
                       if (li_12 > gi_1048 && li_12 < gi_1052 && ld_36 >= gd_1104) li_32 = 2;
                       else {
                          if (li_12 > gi_1052 && li_12 < gi_1056 && ld_36 >= gd_1112) li_32 = 3;
                          else {
                             if (li_12 > gi_1056 && li_12 < gi_1060 && ld_36 >= gd_1120) li_32 = 4;
                             else
                                if (li_12 > gi_1060 && ld_36 >= gd_1128) li_32 = 5;
                          }
                       }
                    }
                    if (li_32 != 0) CloseOrder(OrderTicket(), OrderLots(), OrderType(), g_slippage_876);
                 }
              }
           }
        }
     }
  }
}

 

That is how they handle the matter in FAP, let's try and add trailing stops to that?

[Abdulisback]

Re: EAs tested

 

My requirement for a good EA, relative and max draw down shall less than 18%, open floating position shall be less than 18% equity drawdown, and profits factor > 1.5.

If it is a scalper, then one loser shall only wiped out 2 winners. that's means stoploss = 10 pips, takeprofits = 5 pips. if takeprofits =3 pips, then hardstoploss = 6 pips. and profits factor > 1.5.

 

FAPturbo and magardroid do not satisfy me, because one loser wiped out 10 winners. Yourlucky or Thunder4WD has one losser that wiped out 7 winners. can not pass my standard also. Robominer, it has floating position loss more than 20%, does not pass my standard either.

 

EAs I have backtested and do not satify this standard,

Robominder

Neural Scalper EURUSD M5

Magic Machine

FAPturbo

Magadroid

Thunder4WD (all versions, include march 2009 version)

YourLucky (all versions, include march 2009 version)

Euro Blaster

Intelectual Pro

Bogie-nn

Ilan

Stomper

Xterme

 

 

 

If we are not able to find a such good EA, then maybe we can tell what EA do not work, so we do not waste time on test all the EAs.

 

http://ckowyong.com/blog/statements/

here is one guy who has tested some EA.

 

You are talking my Language. I sometimes see people gathering all over an EA, then it turns out to have 20 pips Profit, and 500 pips stoploss.

 

This means I am better off flipping a coin and closing my eyes to decide (Heads for buy, Tales for sell :P )

 

Only the unexperienced would fall for such dumb scam BS

[pultimo]

Re: EAs tested

 

My requirement for a good EA, relative and max draw down shall less than 18%, open floating position shall be less than 18% equity drawdown, and profits factor > 1.5.

Ehy..!! You are lookin' for my EA !!

:-)

 

I have tried and coded several eas like this but there is always a price to pay: small losses, small risk--->small gains (if all goes right....) you can't circumvent this.

And..sure, if you'll find an EA with such characteristics and big gains, I'm really true interested.. :-)

:peace:

[finimej]

Re: EAs tested

 

I have posted an EA that I think has very good trailstop and takeprofits, is

http://www.indo-investasi.com/viewtopic.php?f=6&t=2701