☰
›

Data Mismatch Kite and API 'Quote'

rejeesh31
rejeesh31 edited September 2020 in Python client
Hi, I've been fetching data using KiteConnect API 'Quote' mode and parsing to get (open/high/low) . This morning, I find the 'low' provided by the API and what shows in Kite chart to be significantly different. Please find screenshot attached where I'm printing ticks using a simple print(ticks) Please see highlighted low as per API as 22627 and low as per kite chart as 22642.5. How do you suggest I use this data since my algo is dependant on 15 min high/low?
Tagged:
  • rejeesh31
    rejeesh31
    As per the kite chart the open_low was broken twice, but my model did not trigger because of this mismatch
  • rakeshr
    rakeshr
    @rejeesh31
    You can go through this thread, which explains above query.
  • rejeesh31
    rejeesh31 edited September 2020
    @rakeshr @sujith Went through the thread, but what is the solution? For ORB strategies what numbers should I rely on? The chart or the API? Pls clarify

    why cant you just broadcast the same OHLC as you show in your own charts? is that difficult? I do not understand.
  • rejeesh31
    rejeesh31
    also, is the historical data API prone to this error? what if I use the OHLC from historical API instead of the 'quote' OHLC? pls respond
  • rakeshr
    rakeshr
    @rejeesh31
    what if I use the OHLC from historical API instead of the 'quote' OHLC?
    If you are just dependent on day's OHLC, then you can better use quote OHLC data.
  • rejeesh31
    rejeesh31 edited September 2020
    dude, are you mocking at me?! that is what we just found out to be wrong, isnt it? which one is accurate for intraday purposes? Quote OHLC or chart OHLC?
  • rejeesh31
    rejeesh31 edited September 2020

    @rakeshr @sujith Today, there was a delta of 66 points in BNF FUT between kite chart and websocket data. This is breaking my program. Can you suggest a solution? What is the best way to read OHLC of open 5 mins? how can I get it consistent with your charts?

    Please respond, and please share the link to open a case.

    Thanks
  • rakeshr
    rakeshr
    Today, there was a delta of 66 points in BNF FUT between kite chart and websocket data
    There can be a difference between Websocket Data/OHLC/Quote w.r.t Kite chart/Historical data for an open intra-day candle, because of reasons explained here.
    Can you suggest a solution? What is the best way to read OHLC of open 5 mins?
    As explained previously, the chart candles can miss a few ticks. So, better use Websocket/Quote data.
    Maybe you can request Quote data/fetch Websocket data exact at 09:19:59 AM for 5-minute candle, this fetches you exact OHL data.
  • rejeesh31
    rejeesh31 edited September 2020
    @rakeshr As explained previously Websocket/Quote data is what I'm using. and comparing with kite chart and finds it it be off. I do not have historical/candle data API subscription. Now are you saying that Kite charts are incorrect? 66 points is not few ticks by any measure.
  • rejeesh31
    rejeesh31
    by Websocket/Quote you are referring to ohlc parsed from ticks right? like this tick["ohlc"]["high"]
  • sujith
    sujith
    @rajeesh31,
    The Quote data comes from the exchange, so it will be the correct data always. We will ask the data team to take a look at the candle data.
  • rejeesh31
    rejeesh31
    great, that's all I'm asking @sujith . A straight answer to a simple question. Going in circles helps no one. So you are saying that the kite chart is inaccurate in this scenario.

    1. Should I create an issue ticket somewhere to track what your data team find on this?
    2. How does it mismatch anyway, since you and I are receiving the data from the same place at the same time and you are my data provider?! I'd be really keen to find out.
  • sujith
    sujith
    The historical data is built using a different stream which is the same as the Websocket API. It is off because it might have missed ticks. We have asked the data team to take a look at this.
    You can track progress on this thread itself.
Sign In or Register to comment.