Great EA in backtest! - page 121

New comment
 

backtest vs. live

Anyone know why when always backtesting rezult it's amazing and in trading live the rezults it's so losses. How can I interpret the rezult of the backtest for future live trading? what is the point who must analize for considering the rezult of backtesting is trusted?

 
 

Where to begin?

I have the new Build 200 of metatrader 4 now. I find that little squeaker noise the backtester makes amusing the first three or four times but now it's just plain annoying. I wish I could turn it off. Anyone know how?

The backtester does appear to be running differently. I'm not sure it's a good thing. It is what it is, we'll all just have to learn to adapt to this new tester. The CT still tests very strong in the backtester, astronomical returns until you go live with it.

My answer to the 'can backtesters be trusted' question...

yes but only to a point, they can be trusted to represent to some degree what changes to a code do relative to each other. They are not the best representation of what actual trading performance of a code will generate. So you need to understand and accept what it is that the tool is useful for and not expect it to do something that presently it just can't do. Don't expect it to show you how an EA with trade live. However it is good to show what changes to an EA will do compared to before the changes. Does that make sense?

I can see one reason for the divergence between backtest and live trading. The pip spread is varied by the broker in live trading and the backtester doesn't do this. As far as I know the backtester has no way of adding this broker side manipulation to it's testing. There is also variance in getting fills to actual orders. That's what they always tell me. Whatever the reason these or others the fact is that it doesn't trade the same live and in backtests so the backtests are only rough estimates of actual performance. the only way to get actual performance data is to run it live. I am doing that with the ppf version using maxlots=.01 just to do win/loss pip counts of how it trades live. It's also showing me how active it is. Yesterday it didn't take any positions all day until midnight ibfx time and then it's taken it's 5th position of the day already today. two wins two losses and now it's in trade 5. the two wins matched the two losses exactly with the changes I added. It totally broke even. -7,-11,+9,+9 right now it's down -5 in the position it's in but it's not closed yet who knows if it will win or not...the odds look like 50/50. It's not going to make any substantial monetary difference to my account but over the next few days it should allow it to make an accounting of itself which will tell me if it's enough to be useable or not. If it is I'll increase the lots and if not, well...who knows.

Templar it's good to have someone else working on improving this. I don't feel so alone now. I REALLY wish you had made these additions to the ppf version!!! instead of the plain old 1.93 euro. I think it's missing the mark to assume that the ppf version is not better because it generates lesser rewards in the backtester. The goal of the version ppf is not to generate more rewards in the backtester but to generate more rewards in live account. Where would you rather see the profits, in backtesting or in live trading? If it can do better live who cares if it doesn't backtest as well. I would like to see the upgrades you made added to the ppf version.

I suggest you look into the string concatenate functions and or simple stack multiple variables on the same line you want to print to file. For example:

study this code snippet and see if you could't do something similar....

//+------------------------------------------------------------------+

//| Conclusion of errors with the entrance into the market |

//+------------------------------------------------------------------+

int PrintErrorValues()

{

Print("ErrorValues:Symbol=", Symbol(),",Lots=",Lots, ",Bid=", Bid, ",Ask=", Ask,

",SlipPage=", SlipPage, "StopLoss=",StopLoss,",TakeProfit=", TakeProfit, "tp * point= ",(TakeProfit*Point));

return (0);

} [/PHP]

One more quick question. Does anyone know a broker who supports metatrader 4 who welcomes scalper ea's that may only hold their positions for one minute?

for example:

[PHP]31 2006.01.12 14:30 s/l 2 0.80 1.2062 1.2062 1.1016 -36.80 798.65

32 2006.01.12 14:30 sell 3 0.70 1.2095 1.2220 1.1095

33 2006.01.12 14:31 modify 3 0.70 1.2095 1.2075 1.1095

34 2006.01.12 14:31 s/l 3 0.70 1.207531 1.2075 1.1095 13.78 812.43

35 2006.01.12 14:31 sell 4 0.70 1.2100 1.2230 1.1100

36 2006.01.12 14:32 modify 4 0.70 1.2100 1.2075 1.1100

37 2006.01.12 14:32 s/l 4 0.70 1.207531 1.2075 1.1100 17.28 829.71

38 2006.01.12 14:32 sell 5 0.80 1.2103 1.2236 1.1103

39 2006.01.12 14:33 modify 5 0.80 1.2103 1.2075 1.1103

40 2006.01.12 14:33 s/l 5 0.80 1.207531 1.2075 1.1103 22.15 851.86

41 2006.01.12 14:33 sell 6 0.80 1.2107 1.2244 1.1107

42 2006.01.12 14:34 modify 6 0.80 1.2107 1.2075 1.1107

43 2006.01.12 14:34 s/l 6 0.80 1.207531 1.2075 1.1107 25.35 877.21

44 2006.01.12 14:34 sell 7 0.80 1.2111 1.2252 1.1111

45 2006.01.12 14:35 modify 7 0.80 1.2111 1.2075 1.1111

46 2006.01.12 14:35 s/l 7 0.80 1.207531 1.2075 1.1111 28.55 905.76

47 2006.01.12 14:35 sell 8 0.80 1.2115 1.2260 1.1115

48 2006.01.12 14:36 modify 8 0.80 1.2115 1.2075 1.1115

49 2006.01.12 14:36 s/l 8 0.80 1.207531 1.2075 1.1115 31.75 937.51

50 2006.01.12 14:36 sell 9 0.90 1.2114 1.2258 1.1114

51 2006.01.12 14:37 modify 9 0.90 1.2114 1.2075 1.1114

52 2006.01.12 14:37 s/l 9 0.90 1.207531 1.2075 1.1114 34.82 972.33

53 2006.01.12 14:37 sell 10 0.90 1.2121 1.2272 1.1121

54 2006.01.12 14:38 modify 10 0.90 1.2121 1.2075 1.1121

55 2006.01.12 14:38 s/l 10 0.90 1.207531 1.2075 1.1121 41.12 1013.45

56 2006.01.12 14:38 sell 11 0.90 1.2110 1.2250 1.1110

57 2006.01.12 14:39 modify 11 0.90 1.2110 1.2075 1.1110

58 2006.01.12 14:39 s/l 11 0.90 1.207531 1.2075 1.1110 31.22 1044.67

59 2006.01.12 14:39 sell 12 0.90 1.2105 1.2240 1.1105

60 2006.01.12 14:40 modify 12 0.90 1.2105 1.2075 1.1105

61 2006.01.12 14:40 s/l 12 0.90 1.207531 1.2075 1.1105 26.72 1071.39

62 2006.01.12 14:40 sell 13 1.00 1.2112 1.2254 1.1112

63 2006.01.12 14:41 modify 13 1.00 1.2112 1.2075 1.1112

64 2006.01.12 14:41 s/l 13 1.00 1.207531 1.2075 1.1112 36.69 1108.08

65 2006.01.12 14:41 sell 14 1.00 1.2113 1.2256 1.1113

66 2006.01.12 14:42 modify 14 1.00 1.2113 1.2075 1.1113

67 2006.01.12 14:42 s/l 14 1.00 1.207531 1.2075 1.1113 37.69 1145.77

68 2006.01.12 14:42 sell 15 1.00 1.2114 1.2258 1.1114

69 2006.01.12 14:43 modify 15 1.00 1.2114 1.2075 1.1114

70 2006.01.12 14:43 s/l 15 1.00 1.207531 1.2075 1.1114 38.69 1184.46

71 2006.01.12 14:43 sell 16 1.10 1.2116 1.2262 1.1116

72 2006.01.12 14:44 modify 16 1.10 1.2116 1.2075 1.1116

73 2006.01.12 14:44 s/l 16 1.10 1.207531 1.2075 1.1116 44.75 1229.21

74 2006.01.12 14:44 sell 17 1.10 1.2113 1.2256 1.1113

75 2006.01.12 14:45 modify 17 1.10 1.2113 1.2075 1.1113

76 2006.01.12 14:45 s/l 17 1.10 1.207531 1.2075 1.1113 41.45 1270.66

77 2006.01.12 14:45 sell 18 1.20 1.2111 1.2252 1.1111

78 2006.01.12 14:46 modify 18 1.20 1.2111 1.2075 1.1111

79 2006.01.12 14:46 s/l 18 1.20 1.207531 1.2075 1.1111 42.82 1313.48

80 2006.01.12 14:46 sell 19 1.20 1.2108 1.2246 1.1108

81 2006.01.12 14:47 modify 19 1.20 1.2108 1.2075 1.1108

82 2006.01.12 14:47 s/l 19 1.20 1.207531 1.2075 1.1108 39.22 1352.70

83 2006.01.12 14:47 sell 20 1.20 1.2111 1.2252 1.1111

84 2006.01.12 14:48 modify 20 1.20 1.2111 1.2075 1.1111

85 2006.01.12 14:48 s/l 20 1.20 1.207531 1.2075 1.1111 42.82 1395.52

86 2006.01.12 14:48 sell 21 1.30 1.2113 1.2256 1.1113

87 2006.01.12 14:49 modify 21 1.30 1.2113 1.2075 1.1113

88 2006.01.12 14:49 s/l 21 1.30 1.207531 1.2075 1.1113 48.99 1444.51

89 2006.01.12 14:49 sell 22 1.30 1.2111 1.2252 1.1111

90 2006.01.12 14:49 modify 22 1.30 1.2111 1.2075 1.1111

91 2006.01.12 14:59 modify 22 1.30 1.2111 1.2066 1.1111

92 2006.01.12 15:00 modify 22 1.30 1.2111 1.2066 1.1111

93 2006.01.12 15:00 s/l 22 1.30 1.206625 1.2066 1.1111 58.18 1502.69
 

sorry double post

 

i test Cyberia Trader 1.93 Euro and lose -25.67

2006.11.15 12:33 sell 1.51 eurusdm 1.2793 1.2810 1.2696 2006.11.15 15:11 1.2810 0.00 0.00 0.00 -25.67

111111 NeuroCluster-testing-AI-LS1[sl]

 

I'm having computer issues today...it keeps repeating my posts.

So moneymaxs what did you learn? what is the point of your test? Was it worth $25? to make one trade? what have you concluded? Personally I'm only allowing maxlots = .01 until I can see more live results of it's performance. my results so far...

Closed Transactions:

Ticket Open Time Type Lots Item Price S / L T / P Close Time Price Commission Taxes Swap Profit

12197952 2006.11.15 15:05 sell 0.01 eurusdm 1.2803 1.2820 1.2706 2006.11.15 19:05 1.2820 0.00 0.00 0.00 -0.17

12192173 2006.11.15 12:33 sell 0.01 eurusdm 1.2793 1.2802 1.2696 2006.11.15 13:35 1.2784 0.00 0.00 0.00 0.09

12190431 2006.11.15 11:18 sell 0.01 eurusdm 1.2793 1.2801 1.2696 2006.11.15 11:32 1.2784 0.00 0.00 0.00 0.09

12180357 2006.11.15 07:25 sell 0.01 eurusdm 1.2818 1.2829 1.2721 2006.11.15 08:10 1.2829 0.00 0.00 0.00 -0.11

12179942 2006.11.15 07:10 sell 0.01 eurusdm 1.2814 1.2821 1.2717 2006.11.15 07:25 1.2821 0.00 0.00 0.00 -0.07

0.00 0.00 0.00 -0.17

Closed P/L: -0.17

Open Trades:

Ticket Open Time Type Lots Item Price S / L T / P Price Commission Taxes Swap Profit

12203587 2006.11.15 19:07 sell 0.01 eurusdm 1.2819 1.2837 1.2722 1.2822 0.00 0.00 0.00 -0.03

0.00 0.00 0.00 -0.03

Floating P/L: -0.03

[/php]

It's interesting that my platform opened a position the same time as the one you posted

[php]

2006.11.15 12:33 sell 1.51 eurusdm 1.2793 1.2810 1.2696 2006.11.15 15:11 1.2810 0.00 0.00 0.00 -25.67

111111 NeuroCluster-testing-AI-LS1[sl]

and mine won 9 pips. You see that's what I wanted from the ppf version. By posting your loss from the same time with the 1.93 version that shows me that the 193.ppf version has made something better I think. Could also indicate different broker feeds too. that's very interesting.

 

An interesting point on testing in demo/live account that I find really important is to have it running for a long period, like a month or more, and see pips balance you got...

I am with 1.93 (not the ppf one) in a demo account with FXDD, and got already 2 losses of 12 and 14 pips; I am not having any conclusion after the 2nd or 3rd week, because I trust that CT 1.93 can maintain the 75% of wins.

But of course that the data feed could interfere on the CT v1.93 results;

anyways, I think that it is wise to wait for a couple of weeks until see the average wins of this version, because it has some small drawdown in the beginning sometimes...

 

What time frame you are using? Notice with M1 no any trades on both version.

 
fibo:
What time frame you are using? Notice with M1 no any trades on both version.

H1 is default for last releases of this EA.

See the journal if you are getting some error

 
templar:
H1 is default for last releases of this EA. See the journal if you are getting some error

are V1.93 and v.193ppf based on the open version or Pro version

1...114115116117118119120121122123124125126127128...150
New comment