Futures bonding - finding seams - page 4

New comment
 
pivomoe:

You can't trade on glues :)

So it's not about trading, it's about testing - you can't do it purely technically, not because it's complete heresy.
  
//+------------------------------------------------------------------+
//| Проверяет, не попадает ли дата в период экспирации на splice-ах  |
//+------------------------------------------------------------------+
bool IsSpliceExpirationPeriod(const string _symbol)
 {
   MqlDateTime now;
   TimeCurrent(now); 
   if(_symbol=="BR Splice" && now.day>29 && now.day<2) return true;
   if(_symbol!="BR Splice" && StringFind(_symbol,"Splice")>=0 
               && (now.mon==3||now.mon==6||now.mon==9||now.mon==12) 
               && now.day>=14 && now.day<=16)
          return true;
   //---
   return false;
 }
 
vito333:

Interesting code, but what if the gluing was on the 15th and17.06.2013 as per Si?

 
Aleksey Vyazmikin:

Interesting code, but what if the gluing was on the 15th and17.06.2013 as per Si?


the vast majority of price gaps will be removed from testing

and the rest should not affect the result much

as a last resort, extend the period to the 17th

as far as I have watched - the main problems with Si Splice fall within the period

 
vito333:

the vast majority of price gaps will be removed from the test

and the rest should not affect the result too much

at the very least, extend the period to the 17th

as long as I looked - the main problems with Si Splice fall into the period


Forum on trading, automated trading systems and strategy testing

Futures Splice - search for seams

Aleksey Vyazmikin, 2017.07.26 12:17

According to the chart there are such gluing days

15.03.2013
17.06.2013
16.09.2013
16.12.2013
17.03.2014
16.06.2014
15.09.2014
15.12.2014
16.03.2015
15.06.2015
15.09.2015
15.12.2015
15.03.2016
15.06.2016
15.09.2016
15.12.2016
16.03.2017
15.06.2017

And for other futures other than brand and B, have you looked at bonding times?

 
Aleksey Vyazmikin:


And for other futures, other than brand and B, did you look at bonding time?

I don't really care, the system should be stable, even such gaps

But it is on Si, even in MT5, even in Finam data, that the most noticeable gaps requiring correction are observed

 
vito333:

I don't bother, the system must be stable, even with such gaps

But it is on the Si, even in MT5, even in Finam data, that the most noticeable gaps requiring correction are observed


I also took a look at Sberbank and the data mostly show 15 and 16 gaps.

My Expert Advisor, on the contrary, was overestimating the profit - I didn't believe it and got worried about these glues...

 
Aleksey Vyazmikin:

I looked at Sber - the data is also mostly 15 and 16 glued together.

My Expert Advisor, on the contrary, was overestimating profits - I didn't believe it and wondered about these glues...


Well, thanks to you I've created the code above and will do my own testing

I didn't get around to it

 

Why not just consider a gap? An extra gap every three months won't hurt anything, imho.

 
vito333:

Well, thanks to you, I made up the code above and I'll put it in, let the testing be more accurate

I never got around to it.


All for the best :)

1234567
New comment