Backtesting/Optimization - page 3

kalahariduin
48
kalahariduin  

Tick data

Thanks Nicholishen I got it. I will have a play and see how it goes.

Nicholishen
1016
Nicholishen  

MT4 Strategy Tester - M1 EurUsd Tick Data

The files include a .hst and a .fxt that contain over one year of live tick data collected by broker.

see install instructions - https://www.mql5.com/en/forum/173508

Cheers

Files:
eurusd1.rar  2271 kb
eurusd1_0.rar  9993 kb
Nicholishen
1016
Nicholishen  

Backtesting/Optimization

The files include a .hst and a .fxt that contain over one year of live tick data collected by broker.

see install instructions - https://www.mql5.com/en/forum/173508

Cheers

Files:
eurusd1.rar  2271 kb
eurusd1_0.rar  9993 kb
Mariusz Gumowski
10
Mariusz Gumowski  

Hi,

thanks for the data, you did a great job! Unfortunately the .hst file is corrupted, header is messed up. I can see you've changed copyright field, but you should do this without shifting data, in some kind of "overwrite" mode. Can you post unmodifed .hst file?

Thanks,

CodeMaster

Mariusz Gumowski
10
Mariusz Gumowski  

Hi,

thanks for the data, you did a great job! Unfortunately the .hst file is corrupted, header is messed up. I can see you've changed copyright field, but you should do this without shifting data, in some kind of "overwrite" mode. Can you post unmodifed .hst file?

Thanks,

CodeMaster

greg
4
greg  

this is a M1 bar data, not real tick data

the backtester works with simulated not real ticks

the only real quotes are open, close, high and low

that is why you always get 25% quality on M1 backtesting

if you want real ticks then log them dont use backtester

greg
4
greg  

this is a M1 bar data, not real tick data

the backtester works with simulated not real ticks

the only real quotes are open, close, high and low

that is why you always get 25% quality on M1 backtesting

if you want real ticks then log them dont use backtester

Nicholishen
1016
Nicholishen  
greg:
this is a M1 bar data, not real tick data

the backtester works with simulated not real ticks

the only real quotes are open, close, high and low

that is why you always get 25% quality on M1 backtesting

if you want real ticks then log them dont use backtester

According to the broker from which I obtained this data, the fxt file is what MT creates from the hst data. They are saying that the fxt file emulates the server. The broker states that the fxt which they included is not emulation, but actual tick data recorded from their server. I am just going on what they are telling me. I hope it helps. I have always been able to get 90% or better. It is possible that the file was corrupted from compression. The original file that they sent was 99Mb

Nicholishen
1016
Nicholishen  
greg:
this is a M1 bar data, not real tick data

the backtester works with simulated not real ticks

the only real quotes are open, close, high and low

that is why you always get 25% quality on M1 backtesting

if you want real ticks then log them dont use backtester

According to the broker from which I obtained this data, the fxt file is what MT creates from the hst data. They are saying that the fxt file emulates the server. The broker states that the fxt which they included is not emulation, but actual tick data recorded from their server. I am just going on what they are telling me. I hope it helps. I have always been able to get 90% or better. It is possible that the file was corrupted from compression. The original file that they sent was 99Mb

[Deleted]  

Importing and converting history data for strategy testing - 1HR bug?

Hi,

I have followed CodersGuru's tutorial on loading historical price data from Alpari and have gotten this process to work for all time periods EXCEPT the 60 minute time. I get a total of 573969 records for the 1M file, 132860 records for the 5M , 44723 records for 15M, 22175 records for 30M, but only 536 records for the 1H (60M) timeframe. Given that 573969 divided by 60 = 9566.15 I'm mystified why I get far less bars than I should.

In addition, the earliest date listed for the 1H timeframe in the History Center is 2006.03.06 whereas for all the other timeframes it is 2004.06.16 - nearly 2 years' more data!!

Has anyone else encountered this issue? Seems like a bug to me. Given that I need the 1H time period in order to backtest my strategy, I really need to find a solution.

Thanks,

Ian