Question about using historical intraday futures data correctly, i.e., gaps between days...

Discussion in 'Data Sets and Feeds' started by Howard, May 2, 2016.

  1. Howard

    Howard

    Are you that easily offended? You must be very sensitive. :)

    Also, I think you're giving yourself too much credit.

    Using the index was already suggested by me in two posts before you suggested that. Nothing new there. Then, you're now back to saying how analyzing the market you're actually trading is the best bet, thus suggesting I ignore the index and focuse on the futures contract?

    Having followed (and traded) the ES market for 5+ years and having created a quantitative system, I'd like to consider myself elevated at least above the newbie stage. Intermediate, I hope.

    But hey. Thanks! :)
     
    #11     May 7, 2016
  2. Howard

    Howard

    I did a side-by-side chart comparison of the continuous non back adjusted futures contract versus the index for 2016.

    Observations:

    1. On a far majority of days, they both gap in the same direction, which is what matters, IMO.

    2. On a few days, the futures have a visible gap, while the index have a more flat open/smaller gap, but still in the same direction as the futures. Just not as pronounced.

    3. Only one day did I see a significant difference and that was on the 10th of March, rollover day for futures. Index is gap up while futures is gap down.

    If rollover day is the only day that will give me trouble, this could be as easy as being aware of that bias and weight this day/gap less than most other days, no?
     
    #12     May 14, 2016
  3. Howard

    Howard

    Hmmm...

    I just did another side-by-side comparison with index versus continuous BACK ADJUSTED contract.

    For 2016, these two match each other better than the continuous non back adjusted, including similar gap direction on rollover day.

    However, if I understood IQ Feed correctly (I did talk to them), the more you go back in time with the adjusted contract, the more prices are divorced from "reality", since the prices are smoothed too much.

    So, although they're not allowed to give any advice per se, they did suggest that the continuous non back adjusted may be better for historical analysis if you go a long way back.

    Currently, I don't seem to have index data going that far back such that I could make a comparison myself, although I suppose I could get some free EOD stuff at Yahoo or similar.
     
    #13     May 14, 2016