Modelling quality

New comment
 

Hi,

Just wanted to know where we can see the modelling quality and how we can improve it?

cheers,

Daniel

 

Backtesting and data:

- Backtesting/Optimization: big public thread.

- Historical alpari data and how to optimize/backtest the settings in EA: I am using this database; public thread with many links to get the data.

- backtesting with 99% modeling quality is here.

 

Hi newdigital,

Thank you very much for your answer. It's really nice :-))

 

Simple answer:

- download full M1 quotes (Alpari gives those away for free)

- import these quotes in the MT4. make sure you work offline (block all traffic using firewall) and delete all existing data for the desired pair. Otherwise its going to be corrupt.

- run the backtest using "all ticks model" on timeframes, higher than M1 (M5,M15,M30,H1,H4,D1,W1,MN)

This way you will get 90% modeling quality.

You can go for 99% but that only (my opinion, from this point on) makes results worse. 99% quality represents tick by tick data. Due to the fact every broker has different tick data, settings, obtained via optimizing with one broker's ticks will lead to completely different result on another broker's ticks.

Thats it! If you want more information, see links newdigital provided.

 
Shinigami:
Simple answer:

- download full M1 quotes (Alpari gives those away for free)

- import these quotes in the MT4. make sure you work offline (block all traffic using firewall) and delete all existing data for the desired pair. Otherwise its going to be corrupt.

- run the backtest using "all ticks model" on timeframes, higher than M1 (M5,M15,M30,H1,H4,D1,W1,MN)

This way you will get 90% modeling quality.

You can go for 99% but that only (my opinion, from this point on) makes results worse. 99% quality represents tick by tick data. Due to the fact every broker has different tick data, settings, obtained via optimizing with one broker's ticks will lead to completely different result on another broker's ticks.

Thats it! If you want more information, see links newdigital provided.

Do u get 90 % on 1m?

How do u get 90 % on 1m?

Regards

El cid

 

You may get data from this post.

25% is maximum for M1.

 

Never trust backtesting. Always use live forward testing. Backtesting is the backdoor to failure. It will onlyhelp you tweak your system to trade in the past, but not in the furture!

Regards

Yanni

Welcome to My Fx Report - Live Forex Trading

.

 

Hi,

Thank you very much to everybody for your answers :-))

Well I know that the best is forward testing to really evaluate a system, but we should be able to get a somewhat acurate set of parameters using backtesting, or am I wrong?

Cheers,

Daniel

 

funny thing is 3 different computers give 3 different sets of results on backtesting of same strategy and data

 

Yeah that's weird ,One other platform had the same problem(smarquant) when it was being developed and they fixed it finally ...

New comment
Reason: