I tried to create concatenated futures contract for US 10 Yr bond futures in eSignal and compared it to Bloomberg.
In Bloomberg, the setting I used was 22 days before expiration. Bloomberg automatically includes holidays such as Sat / Sun within the 22 days, but if the day of the rollover is a holiday such as Sat / Sun, it will rollover the previous trading day. The adjustment was using difference method, because that's the default adjustment allowed by eSignal.
Most of the time, the two charts appear to be matching.
Here's the charts for Bloomberg (TY1 Comdty) and eSignal (%QTY 1!;22EH;B) where there is a huge gap in the eSignal chart on 11/21/2011 and no gap in the Bloomberg chart for the same date (circled area):
I tried to uncheck the "Include Holidays in roll date calculation" option for continous contracts and the above gap on 11/21/2011 got fixed and the chart (i.e. %QTY 1!;22E;B) once again resembled Bloomberg chart on 11/21/2011, except that now other gaps in the data started occurring such as one on 08/01/2005, see the two graphs below - one from Bloomberg and the other from eSignal:
Neither "%QTY 1!;22E;B" nor "%QTY 1!;22EH;B" are able to give me a correct plot of continuous contracts on eSignal.
Is there anything else that I can do?
Thanks.
In Bloomberg, the setting I used was 22 days before expiration. Bloomberg automatically includes holidays such as Sat / Sun within the 22 days, but if the day of the rollover is a holiday such as Sat / Sun, it will rollover the previous trading day. The adjustment was using difference method, because that's the default adjustment allowed by eSignal.
Most of the time, the two charts appear to be matching.
Here's the charts for Bloomberg (TY1 Comdty) and eSignal (%QTY 1!;22EH;B) where there is a huge gap in the eSignal chart on 11/21/2011 and no gap in the Bloomberg chart for the same date (circled area):
I tried to uncheck the "Include Holidays in roll date calculation" option for continous contracts and the above gap on 11/21/2011 got fixed and the chart (i.e. %QTY 1!;22E;B) once again resembled Bloomberg chart on 11/21/2011, except that now other gaps in the data started occurring such as one on 08/01/2005, see the two graphs below - one from Bloomberg and the other from eSignal:
Neither "%QTY 1!;22E;B" nor "%QTY 1!;22EH;B" are able to give me a correct plot of continuous contracts on eSignal.
Is there anything else that I can do?
Thanks.