How do I know if the data I received is accurate ?

Discussion in 'Data Sets and Feeds' started by Iwilldoit, Aug 5, 2015.

  1. Hi there,

    So I have a collection of various data for various instruments from various providers; esignal, market replay, big mikes forum, colleagues etc.

    I am relying on this data for manual backtesting (non programming backtesting on excel) .

    How can I confirm that the data I have is accurate? I am importing it into Ninjatrader.

    Thank You

    IWDI
     
  2. qxr1011

    qxr1011

    it does not matter how accurate is the data

    if method works it works with any data
     
    Baron likes this.
  3. you sure?
     
  4. Doesn't Ninjatrader have reliable data for back testing? -- I don't use NinjaTrader, so don't really know. But the effort you are about to embark on (Manual back testing in EXCEL) sounds painful and labor (and fat-finger) intensive. TOS has fairly reliable data, but is not really meant for back-testing {Been there hit all the brick walls}. I have heard that TradeStation has best back testing capabilities, however, I have no firsthand knowledge of it, especially if options are involved.
     
    Iwilldoit likes this.

  5. Not everyone knows how to program. I graduated from a business program, and have no experience coding at all.

    Yes i did take some Intro to Programming classes from Lynda, but learning how to program is a career within itself. I can code some console scripts, but the complexity of the models I'm dealing with....you have to be a great programmer.
     
  6. qxr1011

    qxr1011

    i am
     
  7. This is the dumbest post ever on elite trader. Data accuracy has EVERYTHING to do with EVERYTHING when it comes to testing methods. Surely you are jesting--- right???
     
  8. wrbtrader

    wrbtrader

    Accurate data depends on the trading instrument and the data provider. Many low tier data providers (e.g. eSignal) have inaccuracies by a few ticks here and there in comparison to top tier data providers (e.g. Bloomberg, Thomson, CQG).

    There's also the issue of your ISP with lost data packets. For example, I have two different ISP (primary and backup). I would notice a few ticks difference via one ISP in comparison to the other ISP using the same data provider...usually occurring in extremely volatile data surges.

    Some exchanges (reason why I said it depends on the trading instrument) allows purchase of their data directly without the need to use a eSignal or Bloomberg. I've never done such but I would imagine it would be the most accurate data possible considering you're getting it directly from the exchange without a middle man (e.g. eSignal, Bloomberg) getting it first and then sending it to you.

    Just remember this, low tier data providers are cheaper than top tier data providers...reason why most retail traders use low tier data sources...its a cost issue. In contrast, institutions and professional traders use top tier data providers...they really want the best data available or they themselves get it directly from the exchange.

    You can tell the difference in how the low tier providers advertise the data versus the top tier providers. Mainly the top tier providers spend more time advertising the accuracy of their data in comparison to the low tier providers.

    With that said, years back, I used Bloomberg (top tier) and TradeStation (low tier) when I traded in an office with other traders...I would see differences in the data a few ticks here and there each trading day but not enough for me to be concerned.

    The really problematic data providers are the ones that does something strange with the bandwidth and tick count in their data...I believe IB does this or did such in the past...lots of complaints about it here at ET in the past. Something to be concerned about if such is your only source of data to make trade decisions especially if you're using small time frames (3min charts and less).

    There's also the issue that some data providers start (open) and end (close) there data times differently than the exchange itself...they do that for cost savings reasons. For example...years back the data provider QCharts (now owned by the parent of eSignal) had its closing regular session for the Emini futures as 4pm est...an issue that lasted for many years even though folks complained constantly about such. I've heard they now use the exchange regular session close. Another example, eSignal in the past cut off early the Hang Seng futures quotes...they essentially gave it a new regular session closing time. eSignal no longer does that. Therefore, make sure you know the real opening and closing times of your trading instrument and compare it with the opening and closing times of your data provider...if its different...you have a big problem.

    P.S. There's some old videos on Youtube where folks put different data providers side by side for comparison of the DOME (times & sales, speed of the quotes) of trades they did. One guy did a 50 contract Emini ES futures trade and he recorded how it showed up in the times & sales. His broker showed it as 50, one data provider showed it as 45, 5 and another data provider showed it as 25, 5, 20. If you're a DOME trader...this could be a problem.
     
    Last edited: Aug 6, 2015
    marketsurfer likes this.
  9. So here I am at 3am in the morning, after having **finally** completed the bloody daunting task of formatting Sierra text data >>> OpenSourceOffice spreadsheet >>> Ninja text data.

    Now I am not confident that the data for this ESU15 contract is accurate:


    Can someone simply explain how I can verify that the prices match the times?

    So the price at 17h15 should match the closing price on the daily chart?

    I am assuming Sierra exports all data in UTC timezone?




    @pak @marketsurfer @wrbtrader

    @kiwi_trader @MarketOwl @k p @i960 @
     
    #10     Aug 12, 2015