• From: Egypt- Saudi Arabia
  • Registered: 2009-12-12
  • Posts: 968

Topic: A New idea for backtesting If cab be Done!!!!

Hello Popov

the idea is to add function to allow human intervention

Ex: after we find a strategy in the automatic way through strategy builder we can shift to the Demo mode this function will use the same price data again but  enter bar by bar and with each new bar if the setup is found we have the two options one is default which is accept the trade and the other can be done buy us to not accept the trade and continue to add the bars until we finish , the results will be a smaller number of trades than the normal number .
So what is the advantages:
1-some times the strategy ( any strategy) gives a contraindicating signals when the market consolidates so  this option ( Demo mode ) will help us to visually filter the bad signal , then we compare the demo mode findings and the regular FSB findings
2- this gives us the real feelings as we trade and take a decisions if to take the trade or not

Is that Modification Can be done

Thanx again for your great efforts and the hidden team behind that peace of art

Regards

  • Popov
  • Popov
  • Lead Developer
  • Offline
  • From: Bulgaria
  • Registered: 2006-11-30
  • Posts: 7,885

Re: A New idea for backtesting If cab be Done!!!!

This is not the way FSB works.  It sounds like additional tool or separate program. Something like Strategy Player where the user can confirm or cancel a signal.
I think there is such a program - Forex Tester. I haven't tested it.

Probably you can try Open Forex Platform. It's freeware and open source.

Popov's Website

  • From: Egypt- Saudi Arabia
  • Registered: 2009-12-12
  • Posts: 968

Re: A New idea for backtesting If cab be Done!!!!

Hello and thanx for that site it looks good but i will keep focusing here

Another Idea for the backtesting ( if done i think it will be a fantastic addition )
Idea is: why we use all the history data for backtesting and optmization , why not we use for example 80% of data in backtesting and the other 20% recent data to apply our strategy on it ( it looks like the forward testing)

Is that option may validate any strategy results ( just thinking loudly smile )!!!!

Regards

  • Popov
  • Popov
  • Lead Developer
  • Offline
  • From: Bulgaria
  • Registered: 2006-11-30
  • Posts: 7,885

Re: A New idea for backtesting If cab be Done!!!!

Idea is: why we use all the history data for backtesting and optmization , why not we use for example 80% of data in backtesting and the other 20% recent data to apply our strategy on it ( it looks like the forward testing)

There are several options to do this:

1. You can set the desired testing period in "Market" ->"Data horizon" dialog.
2. You can temporarily exclude bars for entry by using the "Data Bars Filter" indicator:

[Opening Logic Condition]
Data Bars Filter
     Do not use the newest bars
     Newest bars  -  1000
     Oldest bars  -  0

3. You can use Out of Sample testing in the Generator and Optimizer.

Popov's Website

  • From: Egypt- Saudi Arabia
  • Registered: 2009-12-12
  • Posts: 968

Re: A New idea for backtesting If cab be Done!!!!

Thanx popov
But regarding OOS option in generator and optimizer how we use it in the proper way and regarding to (30 ) what is the meaning and what if this variable changed to 10 or to 50 !!!!!

so can you enlight us what the uses of that option ?

regards

  • Popov
  • Popov
  • Lead Developer
  • Offline
  • From: Bulgaria
  • Registered: 2006-11-30
  • Posts: 7,885

Re: A New idea for backtesting If cab be Done!!!!

But regarding OOS option in generator and optimizer how we use it in the proper way and regarding to (30 ) what is the meaning and what if this variable changed to 10 or to 50 !!!!!

30 is the percent of data bars not being tested. You can see the vertical line with date note. The Generator or Optimizer take the balance  from that date. The following bars are calculated also but they are not criterion for evaluation.

Popov's Website