Teach a Man to Fish – "Rise of Machines" version

[ag_fx]

Pre-Market levels 25-Jun-20

 

[ag_fx]

Pre-Market levels 25-Jun-20

View attachment 43119

EOD Update - 25-Jun-20

Tough day for any trend following system. Zig -Zag markets are bane for a trend follower. The system lost a bit too.

 

[prashant.joshi]

Good morning everyone. The NIFTY and in general the market has decoupled from fundamentals. The asset prices are moving completely decoupled from intrinsic values. Please trade with tight stop-losses and discipline. Interesting, volatile times ahead .

 

[prashant.joshi]

6 Month chart for Ruchi Soya. What kind of voodoo is this

 

[ag_fx]

EOD Update 26-Jun-20

Again a sleepy low volatility day. System did well to avoid losses and made minor wins.

 

[vats]

Sir, how is the result calculated here ?

 

[ag_fx]

Sir, how is the result calculated here ?

For each stock, 1% risk is assumed from Entry to SL. Result (Profit/Loss) is the %age made based on the exit price.

Example: BPCL Trade on 26 Jun 20. Entry Price: 380.4, SL= 373.05. Thus, risk = 7.35. That is, every 7.35 rs 1% will be made or lost.

Exit = 395.6.
Profit=Exit-Entry = 394.6-380.4 = 14.2.
If 7.35=1%, then 14.2 = 1.93%

 

[vats]

Thanks a lot for your detailed reply, Sir. Much appreciated.

 

[hupadhya]

Hi
How far we should go back with historical intraday data for back testing a intraday 5 minute timeframe strategy? Whether is it enough to back test with 6 months to 1year of 5 minute bars or we should look back for more than a year of historical bars to test for any intraday systems?

Thanks

 

[prashant.joshi]

Hi
How far we should go back with historical intraday data for back testing a intraday 5 minute timeframe strategy? Whether is it enough to back test with 6 months to 1year of 5 minute bars or we should look back for more than a year of historical bars to test for any intraday systems?

Thanks

In my opinion, the longer the period of backtesting, the better will be the result. However having said that, there will be some caveats:

1) Most algo trading systems performance will fluctuate over longer periods, so no algo gives consistent result across all periods
2) A potential approach is to have a meta model selection algorithm, that decides which of your designed algo strategies to run in a given period