Futures bonding - finding seams

[Aleksey Vyazmikin]

In order to test the EA, we need to get rid of the seams on the futures glues, particularly on Si.

How do I know the dates of the stitches to eliminate them?

The idea is the expiry day - but where is the list of all expiry days? Maybe there is a way to find out the expiry date without having to collect information?

[Aleksey Vyazmikin]

Here's on Si compiled by http://www.moex.com/ru/contract.aspx?code=Si-12.17

Year/Quarter	3	6	9	12
2012	15.03.2012	15.06.2012	17.09.2012	17.12.2012
2013	15.03.2013	17.06.2013	16.09.2013	16.12.2013
2014	17.03.2014	16.06.2014	15.09.2014	15.12.2014
2015	16.03.2015	15.06.2015	15.09.2015	15.12.2015
2016	15.03.2016	15.06.2016	15.09.2016	15.12.2016
2017	16.03.2017	15.06.2017	21.09.2017	21.12.2017

What is the best way to exclude these dates?

[Aleksey Vyazmikin]

Implemented this way

void OnTick()
  {
//--Исключаем экспирацию по Si
   if(Symbol()=="Si Splice")
     {
      datetime  Open_timeExp=iTime(_Symbol,0,0);
      MqlDateTime strExp;
      TimeToStruct(Open_timeExp,strExp);
      strExp.hour=0;
      strExp.min=0;
      strExp.sec=0;
      for(int i=0;i<23; i++)
        {
         if(StructToTime(strExp)==StringToTime(ExpSi(i)))
           {
            BuyNow=false;
            SellNow=false;
            break;
           }
        }
     }
  }
//////
//+------------------------------------------------------------------+
//|Массив с датами экспирации опциона Si                             |
//+------------------------------------------------------------------+
string ExpSi(int i)
  {
   string Exp[24]=
     {
      "15.03.2012 0:00",
      "15.03.2013 0:00",
      "17.03.2014 0:00",
      "16.03.2015 0:00",
      "15.03.2016 0:00",
      "16.03.2017 0:00",
      "15.06.2012 0:00",
      "17.06.2013 0:00",
      "16.06.2014 0:00",
      "15.06.2015 0:00",
      "15.06.2016 0:00",
      "15.06.2017 0:00",
      "17.09.2012 0:00",
      "16.09.2013 0:00",
      "15.09.2014 0:00",
      "15.09.2015 0:00",
      "15.09.2016 0:00",
      "21.09.2017 0:00",
      "17.12.2012 0:00",
      "16.12.2013 0:00",
      "15.12.2014 0:00",
      "15.12.2015 0:00",
      "15.12.2016 0:00",
      "21.12.2017 0:00"
     };

   return (Exp[i] );
  }
//+------------------------------------------------------------------+

Is there a smarter way?

[pivomoe]

Aleksey Vyazmikin:

In order to test the EA, we need to get rid of the seams on the futures glues, particularly on Si.

How do I know the dates of the stitches to eliminate them?

The idea is the expiry day - but where is the list of all expiry days? Maybe there is a programmatic way to find this, so I do not need to gather information.

Why guess?

From the current date, you can get the name of the most liquid futures for the instrument and the next following one. And compare the bars of the glue and these two futures. That way you will understand where they make the transition and according to what algorithm.

[Aleksey Vyazmikin]

pivomoe:

Why guess?

From the current date, you can get the name of the most liquid futures on the instrument and the one following it. And compare the bars of the glue and these two futures. This way you will understand where they make the transition and according to what algorithm.

I do not understand your logic.

I'm not suggesting to guess - I collected the expiry dates on Si.

[pivomoe]

Aleksey Vyazmikin:

I don't understand your logic...

I am not suggesting guessing - I have collected the expiry dates on Si.

Where do you get the information that the gluing happens on the last day of the circulation? I understand that you think (or know) that the glue is made up of one futures and then the other. What if it does not? Where does this information come from?

[pivomoe]

By the way, there is a last day of circulation in the symbol properties. It can be accessed from the advisor.

[Viktor Vlasenko]

pivomoe:
By the way, in the properties of the symbol is the last day of circulation. This can be accessed from the EA.

there is, but it is useless for ruling out glue splices

For example the current Si-9.17 is 21.09.2017

and it is better to exclude not only the 15th, but also the 16th, imho

[Viktor Vlasenko]

Aleksey Vyazmikin:

Implemented this way

Maybe there is a more rational way?

I suppose you could just exclude the 14th to the 17th of the 3rd, 6th, 9th and 12th months

[Aleksey Vyazmikin]

pivomoe:
Where is the information that the gluing is on the last day? As far as I understand you think (or know) that up to a certain point in time the glue is composed of one futures and then another. What if it does not? Where does this information come from?

The stack consists of different futures. About the last day - an observation, including the chart shows it. But this is at Otkritie broker, others may be different.

[Aleksey Vyazmikin]

pivomoe:
By the way, there is a last day of circulation in the symbol properties. It can be accessed from the EA.

How to find old futures - they are not in the tools.