CFDs are complex instruments and come with a high risk of losing money rapidly due to leverage. 66 % of retail investor accounts lose money when trading CFDs with this provider. You should consider whether you understand how CFDs work and whether you can afford to take the high risk of losing your money.

Automated Trading Programs and Data Sources

Hi,welcome to the community, I myself am kind of mid-long term trader, so I just use M1 data for test in MT4 tester, I personally use Alapri Pro-ECN demo account, and download histroy from their history center the M1 data from 1999 for most of majors, I think their data is very good and have longer history than other brokers, as they are a broker with longer history, and it is free.

For Tick data, I know there are some datasuite for sale, but I donot need it.

4 Likes

I am using Darwinex accounts and I download the standard hystory from metaquotes (m1) .
Tickdata makes the difference only for scalpers and I am an algo swing trader.
I suggest you to read topics in this cathegory:
https://community.darwinex.com/tags/algorithmic-trading
Your success with algo depends on your skill as developer and optimizer

3 Likes

I have used Alpari M1 data in the past in the same way that primeZor suggests, however I stopped when Alpari went bust. I hadn’t realised this was still an option? Since then I have downloaded M1 data from TickStory for free and then load into my Darwinex MT4 platform. Their primary data source is from Dukascopy. I can get 12-15 years worth of data for most pairs (I am also a swing trader and so tick data is an unnecessary overhead)

3 Likes

Alapri is only regulated in Russia now, but their data and software is still there and very good and professinal, better than a lot other brokers, I am doing a PAMM there and intend to keep some smaller clients, I suppose we should not have a problem with alapri.

1 Like

Asking what program and data is needed to be a profitable systematic trader is like asking what car and fuel you need to be a good driver. :wink:

My personal experience is that the web is a bad starting point, books are much better.

Hi guys, thanks for all the answers. Im going to try them out. Iv seen other testing programs like TradeBlox, is it worth it to buy a program for just testing? I think the data and its usage/interpretation is more important but does a software makes a difference?

Dont worry @CavaliereVerde. Im just purely asking the tools used. I am absorbing as much knowledge as possible from books, articles and net. All the while learning how to drive, I also want to prepare my car and what cars do the pros prefer with good experience on it. and What books would you suggest to start out? :grin:

1 Like
2 Likes

In general quality material is about futures and Tradestation but a good strategy and developing method works also with forex and mt4 .

Thank you! This will keep me busy for a while :grin:

1 Like

Since you are interested in testing franz0123, you might be interested a series of 4 articles I am writing on the major pitfalls that automated algo traders can fall in to when back testing trading systems. Especially over-optimization (over-fitting). I’ve only published parts 1 and 2 at the moment but part 3 is out at the beginning of next week. You can find part 1 here:

7 Likes

I use Metatrader 5 since backtesting is much better/faster. Using Admiral Markets data right now (in MT5 you just need to open Demo account to download data from the broker automatic), but since now you can import data into MT5 i will have a look soon for some other sources as well. But using MT5 is a little extra work, since most brokers still only use MT4, so you need to code it with mql4 as well, but with some includes and conditional compilations, we managed to find a solution to have one script which we can compile into MT4 and MT5.

3 Likes

Very good papers. Congrat!

My two cents contribution : what realy matter about stastistical significance is the number of events rather than the number of trades.

Il you have a strategy witch is pyramiding, you need to divide the number of trade by the average number of trade within a pyramide. See attached graphic where we have 30 trades but only 2 events.

Despite this basic case, the real number of events could be complex to evaluate.

3 Likes

Yes you need uncorrelated trades, like “representative trading decisions” used for the Experience investable attribute.
This is the reason why I test and trade just with one trade per signal.

4 Likes

Thank you so much. You are absolutely right. :slight_smile: I will make this clear in the article. Also, thank you @CavaliereVerde for your comment. It’s a difficult one to address/measure in an automated way, but it needs to be done!
Thanks both :slight_smile:

1 Like

Please consider backing up this request :blush:

Would appreciate to be able to backtest from native Dwx data outside MT4 to other higher-end trading platforms…

2 Likes

@CondorcetInvestment - I have now updated the articles based on your feedback. Thanks for the heads up. You’re a star :sparkles:

You’r welcome.

I am looking forward to read next papers.

@ignacio, is Darwinex recording tick or M1 data? this could be useful for doing walk forward tests

Thanks

Hi @Aimak,

We are indeed saving the data for a couple of months now. We still need to figure out how we will make them available in the future, we are of course open to suggestions.

Best,
Ignacio

8 Likes

Hey guys, I’m stuck at testing a strategy to a lot of markets/pairs (diversification) and having its cumulative overall performance. Mt4 is not making it easier either. Most of the software that offers this feature are all priced and Im not really sure about their accuracy. and manual testing is very tedious and time consuming. I need your recommendations guys :grinning: many thanks!

1 Like