NF Swing Trading using Fractal AFL

oldtrader

Well-Known Member
#71
Thanks dear Oldtraderji.
In your system how to keep stop loss? & someone after booking profit by keeping some points or certain percentage for target,he wants to enter again in trade between crossovers how can he enter? And pl. guide us about how you personally trade to achieve max.optimization.
Thanks & regards.
This is a Stop and Reverse system --SAR---. For example , If you had entered the buy entered on 11th april with 1 lot at 7630. You would be placing a SL-M at 7802.1 , as at 11 am today, with 2 lots. This value will be changing as the day progresses. Once the SL-M order for 2 lots gets executed , 1 lot to close the previous position, another lot to open a fresh short position. Pls see the attached chart for clarity.

 

Vipul_84

Well-Known Member
#73
Dear Oldtraderji,
Please excuse me for this post.

I severely doubt OPTIMIZATION for backtesting! Optimization simply gives BEST FIT curve for the look back period. As you are putting so much of your energy for backtesting, please consider following.

Yes, I agree that entire logic of TA is that price action is repeating. But the problem with OPTIMIZATION is that, it wants to say "Price Action is Repeating AS IT IS!" That is the big problem. So till backtesting we get good results, but then horror starts.

My suggestion:
1. Optimize for 2006 and with those parameters chech results for further years
2. Repeat step 1 for all years. Like Optimize for only 2007 data and then check results for all years ....

We will see HUGE VARIATIONS!
 

oldtrader

Well-Known Member
#76
Dear Oldtraderji,
Please excuse me for this post.

I severely doubt OPTIMIZATION for backtesting! Optimization simply gives BEST FIT curve for the look back period. As you are putting so much of your energy for backtesting, please consider following.

Yes, I agree that entire logic of TA is that price action is repeating. But the problem with OPTIMIZATION is that, it wants to say "Price Action is Repeating AS IT IS!" That is the big problem. So till backtesting we get good results, but then horror starts.

My suggestion:
1. Optimize for 2006 and with those parameters chech results for further years
2. Repeat step 1 for all years. Like Optimize for only 2007 data and then check results for all years ....

We will see HUGE VARIATIONS!

Very good idea, will try to do it in spare time. Nowadays not finding much time for R&D. Office project is in final stage, so need more time for office work.
 

Similar threads