• don
  • Licensed Member
  • Offline
  • Registered: 2012-05-15
  • Posts: 6

Topic: Strategy Generator vs MT4 Backtesting

Hi ALL,

I am new to FSB and I see that this has a lot of potential to traders, especially for a newbie like me. I hope someone can help me clarify something. I'd like to know how Strategy Generator was designed to do its work. I've been reading a few pages on it and I have tried to use it myself and it's easy to use. I was amazed how fast it was able to produce its results.

In MT4, I use two models of backtesting. I first backtest using "Open Prices Only" and then compare it to "Every Tick". Obviously, using the first method is a much faster way of doing backtest and the second method would be a much slower way of backtesting.

Using "Open Prices Only" is not a really effective means of getting reliable data according to my personal newbie experience since the results are way off compared to the "Every Tick" model - which according to MT4, is the most precise method to do a backtest.

The reason I use the "Open Prices Only" method is because it is easier to optimize a lot of combinations that way.  Then I use the "Every Tick" method on the best result from optimization to see how it really performs.

So with that explained, I am still amazed how FSB calculates everything so fast in just a few seconds. The question is, does Strategy Generator use "Open Prices Only" or "Every Tick" for its backtesting?

My goal is to be able to backtest as accurately as possible with the least testing time possible for the longest time period possible.

All your thoughts are most welcome. I'd love to learn from professionals like you. Thanks so much in advance!

  • footon
  • footon
  • ◄≡≡≡►
  • Offline
  • Registered: 2009-08-17
  • Posts: 2,924

Re: Strategy Generator vs MT4 Backtesting

Hey,

I suggest you start from the beginning:
http://forexsb.com/wiki/fsb/start
http://forexsb.com/wiki/fsb/manual/inte … on_methods

  • don
  • Licensed Member
  • Offline
  • Registered: 2012-05-15
  • Posts: 6

Re: Strategy Generator vs MT4 Backtesting

Hi Footon,

Thanks for your reply. I have read those references prior posting but I still don't understand which style of backtesting FSB uses compared to MT4. From what I understand, FSB only uses entry and exit points - and that it doesn't use Every Tick. Is this correct?

  • From: Australia
  • Registered: 2011-07-01
  • Posts: 431

Re: Strategy Generator vs MT4 Backtesting

I think you should re-read the links that footon posted.

  • footon
  • footon
  • ◄≡≡≡►
  • Offline
  • Registered: 2009-08-17
  • Posts: 2,924

Re: Strategy Generator vs MT4 Backtesting

don wrote:

I have read those references prior posting

Are you serious? It says all here:

Following a fixed route, Forex Strategy Builder goes round each bar in order to execute or cancel all orders sent by the strategy. This route is determined by the four base points - Open Price, High Price, Low Price and Close Price. In most cases, this information is sufficient for a correct calculation of the historical test. If the strategy sends close orders, however, and they are within the limits of the current bar, it is not always possible to determine for sure what the order of order execution is.

Those bars where we cannot tell for sure what the order of order execution are called Ambiguous Bars.

Forex Strategy Builder provides various interpolation methods for calculating the back test. Each one of these methods interpolates the ambiguous bars in a specific way. If the strategy contains a large number of ambiguous bars, the end balance of your virtual bank account can vary significantly with the different methods.

A few more comments - I understand you come from MT and that defines your experience and sets a benchmark of sorts, but it is better to separate FSB from MT, they are not really in the same league. Testing on MT results in unverified calculations, do you expect reliable outcome from "Open Prices Only" or whatever it is called? And what do you mean by "style "Every Tick""? Is it the custom import of tick data because default MT doesn't have real tick data testing, it "simulates" ticks. How it does the latter is anyone's guess.

Have you developed or backtested trading systems/strategies manually? If you have, it should be fairly easy to figure out how FSB works if you can't understand articles, manual, tutorials. It is not a matter of style, it is the matter of reliability in terms of data interpolation.

6 Reply by dabbler (edited by dabbler 2012-05-19 09:35:27)

  • Registered: 2012-05-18
  • Posts: 51

Re: Strategy Generator vs MT4 Backtesting

don wrote:

So with that explained, I am still amazed how FSB calculates everything so fast in just a few seconds. The question is, does Strategy Generator use "Open Prices Only" or "Every Tick" for its backtesting?

My goal is to be able to backtest as accurately as possible with the least testing time possible for the longest time period possible.

FSB is certainly amazing in its speed compared to MT4, and by the looks of it is is more accurate too! Your goal of getting the best accuracy possible is laudable. To this end you might want to consider ticking the Testing - Automatic Scan option.

If the strategy uses close Stop Loss and Take Profit or other rules that make several deals inside a bar, the backtest may show Ambiguous Bars. In that case, you can reduce the number of ambiguous bars and to make the test more reliable by switching on the Testing → Automatic Scan option. It will force Forex Strategy Builder to use all available intrabar data for better bar interpolation. See more information about scanning in the Intrabar Scanner article. Anyway, you can set a maximum number of the ambiguous bars limit for the generated strategies.

from http://forexsb.com/wiki/fsb/manual/generator

When MT4 strategy tester is used in every tick mode it uses data from all available lower timeframes. Ticking the FSB Automatic Scan box also uses the lower timeframe data.

  • don
  • Licensed Member
  • Offline
  • Registered: 2012-05-15
  • Posts: 6

Re: Strategy Generator vs MT4 Backtesting

Thanks so much for clarifying footon and dabbler. I really had my head into MT4 to the point I expected FSB to act the same way when I shouldn't. I really like FSB so much because of all the options and accuracy it gives. And yes, it's way faster than MT4.

I do have one more question to ask if I can ask it here or if I should post it as a new topic? I keep encountering this error - always:

Scanner
Data for: EURUSD 15 Minutes - Maximum days off: 45
The data is probably incomplete and the scanning may not be reliable! You can try also "Cut Off Bad Data".

I got my data from FXDD Metatrader. I downloaded FXDD's  m1 data (hst format) from http://global.fxdd.com/ru/mt1m-data.html and loaded it into MT4. I then exported it to CSV format which I then loaded into FSB. I loaded m1, m5, m15, m30, h1 data only.

Is this data reliable? Did I do it right? Is there a better way to do it?

  • footon
  • footon
  • ◄≡≡≡►
  • Offline
  • Registered: 2009-08-17
  • Posts: 2,924

Re: Strategy Generator vs MT4 Backtesting

don wrote:

Thanks so much for clarifying footon and dabbler. I really had my head into MT4 to the point I expected FSB to act the same way when I shouldn't. I really like FSB so much because of all the options and accuracy it gives. And yes, it's way faster than MT4.

I do have one more question to ask if I can ask it here or if I should post it as a new topic? I keep encountering this error - always:

Scanner
Data for: EURUSD 15 Minutes - Maximum days off: 45
The data is probably incomplete and the scanning may not be reliable! You can try also "Cut Off Bad Data".

I got my data from FXDD Metatrader. I downloaded FXDD's  m1 data (hst format) from http://global.fxdd.com/ru/mt1m-data.html and loaded it into MT4. I then exported it to CSV format which I then loaded into FSB. I loaded m1, m5, m15, m30, h1 data only.

Is this data reliable? Did I do it right? Is there a better way to do it?

That data is crap, as far as I understand, it is still Metaquotes data:

Attention: The loaded data can differ from historical data stored on the trade server.

Everything you download through History Center is from MQ, it is notorious for gaps and blanks. My advice is to use broker's data, so load as much bars in MT as you can, use arrows, Home/End/Page Up/Page Down, then Save As -> FSB data directory. There might still be errors though, but definitely less than in MQ data.

Read and watch more here:
http://forexsb.com/wiki/fsb/metatrader_data
http://forexsb.com/wiki/fsb/forex_data_from_mt_video

  • From: Brazil
  • Registered: 2012-04-12
  • Posts: 15

Re: Strategy Generator vs MT4 Backtesting

I believe that it is of extreme importance a greater quality in modeling tests. As for example in MT4, with a 99% modeling, already here in FSB my biggest modeling was Scanned MQ 57.72%. Downloaded the historical data by TrueFX and still can't get a better modeling.

Unless the Scanned MQ, differs from MT4 modeling, wanting to believe that the first is better than the second.