Should you ignore trades with size less than 100 in your data feed?

Discussion in 'Data Sets and Feeds' started by Luke_4466, Oct 15, 2021.

  1. Ah, ok. I'm referring to data directly from the exchanges.
     
    #11     Oct 23, 2021
  2. You can see in the data example i posted at first post, that the trades include what exchange it is from. But these real time tick data are not, i am assuming, all the data from exchange, but rather reported trades from the exchange.
    I honestly dont know what type of trades are in that included or not, but Finage and IB both report exactly same trades from same exchanges.
    I do wonder tho, if the data you mentioned are something that should be included to get better idea of what is happening with the price of the stock. That would mean including data that all IB, TW and Finage ignore essentially. In this regard i dont know whats the best approach.
     
    #12     Oct 23, 2021
  3. The data you get from IB is probably fine for most purposes. I use their API for trading but I get data from polygon.io since IB is so cumbersome and limited when it comes to getting data. The raw trade data from the exchanges (via polygon.io) is a little too raw to be practical due to all the trade conditions that must be sorted through. Polygon does provide a call that will return 1-min bars that have been filtered for you based on trade condition flags on the raw trades.

    Here is an explanation of trade conditions and how to use the trade data. This chart isn't enough though. You nee this 200+ page document that explains more, which I have a link to somewhere.

    https://polygon.io/glossary/us/stocks/conditions-indicators
     
    #13     Oct 23, 2021

  4. Interesting, i will have a look. Do you think this might be related only to longer bars, like 1-min? There is nothing like this when it comes down to real time tick data, you just get every trade from all exchanges. I think it might be that with longer term bars they exclude some trades based on some additional conditions but with real time tick data, you just get everything in real time with no filters.
    I tested Polygon for real time tick data and they had issue with trades being reported at wrong times. Ie real time tick data would have slightly wrong timestamps ( compared to IB, TW and Finage ). But their 1-min bars might be totaly fine and this issue can be addressed there, dont know, but that was reason why i didnt end up using Polygon.
     
    #14     Oct 23, 2021
  5. The trades from polygon are correct and are exactly as they come from the exchanges. These aren't "issues". It's just that what is coming from the exchanges is more complex and complete than most people realize. Trades are not "being reported at wrong times" by the exchanges. Sometimes data is reported late, corrected, etc. This is what the trade condition field is all about. You'll have to deal with this when using real-time data from the exchanges regardless of whether it is polygon or not.
     
    #15     Oct 23, 2021
    Polygon.io and qlai like this.
  6. Well, i dont agree with this statement, but if it works for you, then thats all that matters i think.
     
    #16     Oct 23, 2021
  7. Polygon.io

    Polygon.io Sponsor

    I'd be happy to clear up any issues you had with our service. @FreeGoldRush is correct about the exchanges' data.

    We do not alter the trades or quotes that come through in any way, we just rebroadcast the messages as is. This means that these Late Reported, Cancelled, or Corrected trades will come through representing a different time. If you could please elaborate on what you saw, there is likely an explanation.
     
    #17     Oct 25, 2021
    beginner66 likes this.