NF Swing Trading using Fractal AFL

oldtrader

Well-Known Member
#11
The results look ok for a mechanical trading system. On lowest year of 782 points and assuming we trade 75 Nifty ( one contract) conservatively for Rs 1,50,000 account, it still gives lowest annual return of 39 % which is not bad.

Doing swing trade on 7 min timeframe probably is increasing the number of trades..ideally in a month there should not be more than 5-7 trades.

Any idea of the total number of winning /losing trades, average winner/average loser for some years ( if you have worked it out ) . I am hoping that the profits are net of brokerage and all charges.The losing months is a concern as points lost are more..but for a totally mechanical system we have to take this in our stride.

The above are general comments without knowing the exact details of the system....so some of them may not be relevant to the system.

Finally as the system is making money, dont try to fine tune it with too many tweaks else it will loose its robustness.

Smart_trade

Smart_trade Sir, Thank you very much for taking time to express your opinion about the system. There are 2 variables in this AFL , I had run optimization on this AFL using Amibroker. Got around 40000 rows of data. I had selected the parameters which were giving good number of points for less number of trades. Testing period was 108 months, trades were 1115 and gross points were 18380 (I didn't consider brokerage and slippage). So monthly average comes to 170 NF points with 10.3 trades. I am in the process of calculating the maximum continous loss, which in other words will give an idea of the max. drawdown. Will try to post the result shortly.
 

mechtrader

Well-Known Member
#12
Smart_trade Sir, Thank you very much for taking time to express your opinion about the system. There are 2 variables in this AFL , I had run optimization on this AFL using Amibroker. Got around 40000 rows of data. I had selected the parameters which were giving good number of points for less number of trades. Testing period was 108 months, trades were 1115 and gross points were 18380 (I didn't consider brokerage and slippage). So monthly average comes to 170 NF points with 10.3 trades. I am in the process of calculating the maximum continous loss, which in other words will give an idea of the max. drawdown. Will try to post the result shortly.
While picking up the values after an optimization, you must keep few points in mind.

1) The values in the vicinity of those "2 variables" also give almost same results.
2) The values picked are not the best values but belong to a group of best values.
3) The picked up values give consistent results on other similar asset classes as well.
4) The picked up values show consistency in other time frames as well.

If all the above points are valid, you are good to go wit those values.. otherwise it is just a curve fit.

MT
 

manishchan

Well-Known Member
#13
Here is the latest snapshot of the AFL as of 19-02-2016. Later on I will be posting the trades executed in this year i.e., January and February 2016.

Posting a summary of trades from August 2015 to date. Trader friends please compare this with your known trading systems and let us discuss the positive and negative points of this system.



Looking at the chart, I can tell you that your PnL is still inflated. I think you are looking at the automatically mearged chart. Usually at contract expiry, when trade gets carried over from last contract to the new one.. the premium difference will get counted in your favor which doesnt happen in realtime. You better do the backtesting manually.
 

oldtrader

Well-Known Member
#14
While picking up the values after an optimization, you must keep few points in mind.

1) The values in the vicinity of those "2 variables" also give almost same results.
2) The values picked are not the best values but belong to a group of best values.
3) The picked up values give consistent results on other similar asset classes as well.
4) The picked up values show consistency in other time frames as well.

If all the above points are valid, you are good to go wit those values.. otherwise it is just a curve fit.

MT

I will go through the points mentioned by you and recheck the optimization results. Alternatively , I would like to upload the excel file of the same, if it will help.
 

oldtrader

Well-Known Member
#15
Looking at the chart, I can tell you that your PnL is still inflated. I think you are looking at the automatically mearged chart. Usually at contract expiry, when trade gets carried over from last contract to the new one.. the premium difference will get counted in your favor which doesnt happen in realtime. You better do the backtesting manually.
To go thru a live example of how the rollover is effecting the trades, I will be posting the results of August and September 2015, expiry trades, today evening.
 

oldtrader

Well-Known Member
#16
While picking up the values after an optimization, you must keep few points in mind.

1) The values in the vicinity of those "2 variables" also give almost same results.
2) The values picked are not the best values but belong to a group of best values.
3) The picked up values give consistent results on other similar asset classes as well.
4) The picked up values show consistency in other time frames as well.

If all the above points are valid, you are good to go wit those values.. otherwise it is just a curve fit.

MT
I will go through the points mentioned by you and recheck the optimization results. Alternatively , I would like to upload the excel file of the same, if it will help.
Dear MT Sir,

Please go through the attached XL file of the optimization results and help in selecting the best . "Acceleration" and "Max. Acceleration " are the variables , other columns are self explanatory.

Regards

P.S. : The file size is around 15MB. It may take sometime to download, sorry for the inconvenience.

http://www.filedropper.com/7minfortrdrji
 

Similar threads