Backtesting Issue... - page 2

New comment
 
Fernando Carreiro #:

You seem to be misinformed! When MetaTrader 5 is connected to your broker's trading account, you get broker historical data and tick data, not data from MetaQuotes. I have been using MT4 and MT5 for a long time on several brokers and can guarantee that the tick data is different between Metaquotes and brokers and also between different account types, such as between Demo and between Live, and between Cent, Standard, Pro and ECN types of accounts. By the way, night spread depends on the type of account you are using. ECN accounts have much better spread than Standard or Pro accounts.

Obviously, depending on the broker, some keep plenty of history data and other keep less. If a broker does not keep record of tick data far back in time, then simply connect to another broker or account type. So far, I have only needed tick data from Dukascopy and with TickStory when testing on MT4. On MT5 I have never needed to do that.

Also, there is also a difference between "Every tick" and "Every tick based on real ticks". The former is virtual ticks and the later are real ticks.

That's how I was thinking till recently, but... Ok, let me copy paste here the response from IC Markets:

The reason for the low-quality test is due to the fact that backtesting requires tick data something which we do not offer. There are third-party services that offer tick data from different providers which you will be able to purchase for back-testing purposes. We do not store the tick data in the server as we offer a big range of instruments and that will not be feasible as it will compromise the correct functionality of the server,( we tried in the past).  We can only provide the tick data for a specific time in which a trade may require investigation( around 10seconds of data)."

"Please note when you are downloading the data from MT5 that data is downloaded from the Metaquotes server itself, not us. When testing on “ Every tick based on real ticks”, and requesting the tick data from Metatrader directly, there is a scope for missing data. If there are numerous data mismatches, your % will be reduced. Please note we do not upload all the tick data on the server, this is not from a few weeks ago, you may be able to get a better % in the past, however, this is due to fewer mismatches on your filter search. If you search for one-minute data and perform the data on every tick you will not get 100%

Weird..

Also it's weird, because the backtesting is indeed very similar to live results. That shouldn't be possible if the ticks were simply generated by the terminal.

 
Alexandru Casian #: That's how I was thinking till recently, but... Ok, let me copy paste here the response from IC Markets: Weird.. Also it's weird, because the backtesting is indeed very similar to live results. That shouldn't be possible if the ticks were simply generated by the terminal.
They are referring to MT4, not MT5! Who ever answered you does not seem to know the difference!
 
Fernando Carreiro #:
They are referring to MT4, not MT5! Who ever answered you does not seem to know the difference!

I believe you're right. Although, I mentioned it's MT5 when I asked them. It's time to bother them again with the same question xD.

 
Alexandru Casian #: I believe you're right. Although, I mentioned it's MT5 when I asked them. It's time to bother them again with the same question xD.

Don't bother with their support staff. Just open up the symbols window and look at the tick data yourself in the ticks tab. You can also see how far back it goes. I don't use IC Markets so can't say for myself.

 
Fernando Carreiro #:

Don't bother with their support staff. Just open up the symbols window and look at the tick data yourself in the ticks tab. You can also see how far back it goes. I don't use IC Markets so can't say for myself.

Thanks for idea! It's indeed faster this way.
My question to my broker would be - why they suddenly deleted ~5 years of tick data, which made backtesting scalping EAs almost impossible for their clients. But, the answer won't change too much anything (._.)

 
Alexandru Casian #: Thanks for idea! It's indeed faster this way. My question to my broker would be - why they suddenly deleted ~5 years of tick data, which made backtesting scalping EAs almost impossible for their clients. But, the answer won't change too much anything (._.)

I opened a Demo account with them, and did a test on EURUSD and I was able to get tick data as far back as 2011.12.30 for it. It was however VERY SLOW. The network speed is somewhat lacking.

Which symbols were you back-testing that lost 5 years of tick data?

 
Fernando Carreiro #:

I opened a Demo account with them, and did a test on EURUSD and I was able to get tick data as far back as 2011.12.30 for it. It was however VERY SLOW. The network speed is somewhat lacking.

Which symbols were you back-testing that lost 5 years of tick data?

GBPUSD,EURUSD,EURCHF,EURAUD,USDCHF,AUDUSD
Weird, for me on Real, earliest tick on EURUSD is 2021.06.29

 
Alexandru Casian #: GBPUSD,EURUSD,EURCHF,EURAUD,USDCHF,AUDUSD. Weird, for me on Real, earliest tick on EURUSD is 2021.06.29

To double check, what is already available, look in the "<data folder>\Bases\<trade server>\ticks\EURUSD" and see how far back it goes.


Then, in Tick Tab request the data in smaller batches (as the downloads on that broker are VERY SLOW and they may fail and stop). So, keep repeating in smaller batches going back further until it does not download any more.


 
Fernando Carreiro #:

To double check, what is already available, look in the "<data folder>\Bases\<trade server>\ticks\EURUSD" and see how far back it goes.

Then, in Tick Tab request the data in smaller batches (as the downloads on that broker are VERY SLOW and they may fail and stop). So, keep repeating in smaller batches going back further until it does not download any more.

Ok, so, I installed a new MT5 application, and made tests with 3 MT5 servers:
ICMarketsSC-Demo (demo). Earliest ticks: 2011.12.19    The download speed for the demo server indeed is slow, 10-100 times slower than for "live" servers.
ICMarketsSC-MT5-2 (live).  Earliest ticks: 2021.06.29
ICMarketsSC-MT5 (live).  Earliest ticks: 2022.04.08
These being said, it seems that demo account ticks are very similar to real account ticks. I never thought to backtest using ticks from a demo account, thinking that ticks would be different. So, thanks again for helping!

ICMarketsSC-MT5-2


ICMarketsSC-MT5


ICMarketsSC-Demo

 
Alexandru Casian #: Ok, so, I installed a new MT5 application, and made tests with 3 MT5 servers:

ICMarketsSC-Demo (demo). Earliest ticks: 2011.12.19    The download speed for the demo server indeed is slow, 10-100 times slower than for "live" servers.
ICMarketsSC-MT5-2 (live).  Earliest ticks: 2021.06.29
ICMarketsSC-MT5 (live).  Earliest ticks: 2022.04.08

You don't have to install a new MT5 for each account for back-testing or collecting tick data. Just log onto a different account and download the tick-data or run the back-tests. Each trade server's data is kept in a different directory.

12345
New comment