Download MetaTrader 5

Back-testing results and a question.

To add comments, please log in or register
Luciano Ola
1007
Luciano Ola  

I've been running back-tests some instruments and I decided to test my EA against volatile instruments, I started with what I thought was volatile e.g XAUUSD then I moved on to instruments I had never traded on before, these instruments turned out to be very volatile but the volatility differed and my EA handled each instrument with different levels of efficiency.

The first is Nikkei tested over a period of nine years, using the D1 timeframe.

The second is Hang Seng tested over a period of 5 years, using the D1 timeframe.

qjol
3243
qjol  
a pretty sharp drop in the middle (hang seng)
Luciano Ola
1007
Luciano Ola  
My question is concerning a principle I have always had about back tests (if it can happen on one instrument, it can happen on any of them) now I have had back tested on over 10 instruments using Do over a 5 to 9 year time period and I had really issues trading fuse and Hang Seng. Hang Seng, why I used a 9 year to pushed my EA too the limit (my EA survived) and the turn over (20%) was quite small imo, ftse ripped my EA to pieces but in all honest I was much more aggressive with ftse than I should have. I only have one extern so it's not an issue of optimisation (which I dislike using).
Luciano Ola
1007
Luciano Ola  
qjol:
a pretty sharp drop in the middle (hang seng)
the I had no intention of trading hang Seng, I just wanted to test my EAS survivability against it after you said it was very volatile, initially I did not think so, it was after that dip that I made my apology to you.
Luciano Ola
1007
Luciano Ola  
Am I correct in thinking that what happened on Hang seng can happen on any instrument? Hang seng was quite peculiar I find that my EA traded it at some sort of disadvantage, just from when I was watching the trades happen and monitoring growth, it's a feeling I can't shake.
qjol
3243
qjol  
i'm wondering how a backtest on the us indices on a specific single date looks
the date i'm talking about is may 6th 2010 (the flash crash day)
Luciano Ola
1007
Luciano Ola  
qjol:
i'm wondering how a backtest on the us indices on a specific single date looks
the date i'm talking about is may 6th 2010 (the flash crash day)
well this nickel is from 2005 to 2014, hang seng is from 2009 to 2014.
Luciano Ola
1007
Luciano Ola  
MetaNt:
well this nickel is from 2005 to 2014, hang seng is from 2009 to 2014.
not nickel nikkei sorry
Luciano Ola
1007
Luciano Ola  
qjol:
i'm wondering how a backtest on the us indices on a specific single date looks
the date i'm talking about is may 6th 2010 (the flash crash day)
not sure if I could find any US instruments during my back testing and I've turned off the testing computer now.
qjol
3243
qjol  
i'm talking about the S&P Nasdaq & dow Jones a backtest date: 06/05/2010 (may the 6th 2010)
Luciano Ola
1007
Luciano Ola  
qjol:
i'm talking about the S&P Nasdaq & dow Jones a backtest date: 06/05/2010 (may the 6th 2010)
Oh right, I tried s and p and the liquidity didn't seem high (I was only testing the EAs resilience so high liquidity was a must) thus I curtailed the backtest.
123
To add comments, please log in or register