EA Backtesting Journal

FxCoder
59
FxCoder  

Latest Result:

Total Backtesting : 17

Winner so far : M1 v0.1

Total net profit : 41,579.81

Profitable so far : 6

Hello everyone!

Is EA really profitable? you may find the answer here.

You can join me in this quest. Post your result here.

Or if you have slow PC, but you want to backtest an EA, ask here.

(My Notebook is suck! it will take up to 4 hours to backtest while my PC will do it for max. 10 mins)

Special Note to all reader:

1. Please note that this result is for 7 years trading!! What's your expectation after 7 years?

2. Don't use EA if you don't know what it means!

Preparation:

1. Make sure you've already download all available data from History Center.

2. Make sure you didn't post wrong pair or wrong time frame, some EA has "built in" code which will make this EA work for just one pair or just one time frame.

3. Also some author make his EA for just one pair or just one time frame without "built in" code. Make sure you're testing the right pair and the right time frame suggested by the author.

To make it "fair", use this Setup for all backtesting:

Start Date : 2001.01.01

End Date : 2007.12.31

Lots : 10.000

Model : Every tick (the most precise method...)

EA that gain highest "Total net profit" will be the winner.

Format for the result will be:

No. EA_Name, Broker, Pair, Total_Net_Profit

Note: Note

1. MACD Sample, North Finance, EURJPY, H1, -5816.46

2. Moving Average, North Finance, EURJPY, H1, -6079.05

3. simplefx, North Finance, EURUSD, D1, 4546.28

Note: Author said that this EA best attached to the daily D1 chart time frame.

4. simplefx, North Finance, EURUSD, H4, -76.78

Note: Just curious, if this EA can do better so i test it using H4.

5. simplefx, North Finance, EURUSD, H1, -629.76

Note: Test using H4

6. gpfTCPivotLimit, North Finance, EURUSD, H1, 2339.86

7. Multi Lot Scalper, NF, EURUSD, H4, -9799.05

Note: good result at first, until some point it's go straight to the bottom. (No stop loss maybe)

Mod: I don't know if this thread is in the right place. Please move it if it isn't.

That's all for today

Sergey Golubev
Moderator
122430
Sergey Golubev  

If you are talking about backtesting so it is the winner. And it is trading system.

To say seriously backtesting is not showing with real situation sometimes. For example, if we do not know: it was coded on open bar or close bar, or it was coded in high/low of the close bar, if it is used some indicator with open/high/low settings ... in those cases backtesting is useless. If i is open source code so we can see it inside the code. But if it is just ex4 file so that's nothing.

Just forwars testing only for couple of years may be ...

Or learning the system (every Ea is based on some trading system) may help.

Sorry for opinion.

Sergey Golubev
Moderator
122430
Sergey Golubev  

And the other winner: MACD_Sample with MM (money management https://www.mql5.com/en/forum/173055/page29

First EA (from previous post) was EA based on momentum indicator (very simple one).

Backtesting of course.

FxCoder
59
FxCoder  

Hello newdigital,

Don't be sorry. in fact thank you for your opinion. All opinion are welcome

My Purpose is to share my backtesting here, so the others won't have to backtest it again.

I know there are many things to consider before develop an EA, which is may get very complicated.

Pair ,Time Frame, Type Technical Indicator, Different signal based on one technical indicator (for ex. MACD Crossovers, MACD center line crossovers, Divergence, Overbought), Setting value for Technical Indicator, Pivot Support & Resistance (Pivot, Fibo, Camarilla inc. its timeframe), Market hours, News Time, MM, Different Combination on Take Profit, TS, SL, Hedging / Locking method, Pattern Recognition, entering time & exiting time.

My target is:

1. Backtest as many EA that i found.

2. Meanwhile, i will develop my own EA which only has 1 technical indicator as a signal, and off course backtest it

3. From target 2 above, i think all indicator will be failed. i will pick two of them which can stand longer, develop and back test it.

4. if target 3 failed, Pick 3 indicator, develop and backtest it (3 indicator is max for now)

5. Add other setting (MM, Pivot, Market hour, etc)

6. Adding other indicator if needed.

btw i'm a programmer who want to start trading so this thread will be my journal since i'm lack of discipline. I hope this thread will keep me in the right path.

Nice to know you.

FxCoder
59
FxCoder  

I decide to upload Strategy Tester Report from now.

Ichimoku5, NF, GBPUSD, D1, -8684.10

Files:
FxCoder
59
FxCoder  

^AC-CenterLine, NF, EURUSD, D1, -847.64

Note: Bill Williams Accelerator Oscillator

Uh.. is there a way to upload several picture an attachment below text in one post?

I want to make it like this:

Text

Picture

Attachment

Text

Picture

Attachment..

so I don't have to submit alot of post like this.

I stop posting for today, til i get a way to do that.

Have to learn how to post first.. duh..

FxCoder
59
FxCoder  

^AC-CenterLine-ReverseSignal, NF, EURUSD, D1, -972.44

Note: Bill Williams Accelerator Oscillator

FxCoder
59
FxCoder  

^AC-ChangeColor, NF, EURUSD, D1, -3175.00

Note: Bill Williams AC, entry position at second candle after histogram change its color

FxCoder
59
FxCoder  

^FC-Bill-AwesomeAccelerator, NF, EURUSD, D1, 3299.85

Note: Joining 2 indicator, Bill Williams ^AC-ChangeColor with Bill Williams ^AO-CenterLine as filter. Only enter Long position if Awesome Oscillator above Center Line & only enter Short position if Awesome Oscillator below Center Line.

I backtest several EA from others thread. got alot of error on opening, closing or modifying order.

But reply from other people are sooo good, i wonder if they really test it???

If you can't send order or modifying order, how come it's profitable?? how come you use such EA???

Please be carefull before using other people EA. backtest it to make sure it doesn't have error code.

And if you don't know how to read a program in this EA, kindly ask other people who understand it.

For developer, please debug your code before you submit it. It would be silly if a good trading system failed because of an error on the code.

It's just an advice, gtg to work, will continue later.

Sergey Golubev
Moderator
122430
Sergey Golubev  

Did you try to optimize the settings to find the profitable one?

FxCoder
59
FxCoder  

No, I still use basic indicator signal for this EA.

I think the good result come from this indicator below.

^AO-CenterLine, NF, EURUSD, D1, 1678.48

Note: Bill Williams Awesome Oscillator, Center Line crossing signal.