¿ªÔÆÌåÓý

ctrl + shift + ? for shortcuts
© 2025 Groups.io

Re: How to interpret "Count" & "Volume" in historical data?


 

Welcome to the group.

As far as I know there is no one standard for exactly what is counted in volumes, number of trades, and even OHLC prices. Just like I am often surprised what is being sold as "Bratwurst".

The largest chunk of difference comes from the fact that IBKR reports historical bars in "lots of 100" and you get much closer to the number reported on other sites after multiplying the bar lot volume with 100.

We don't work with bar data or historical data that much but I happened to have some IBKR historical bar data for IBM in January 2020. For 2020-01-24, for example, IBKR reports for exchange=SMART:

  • 37,922 lots -> 3,792,200 shares during regular trading hours only
  • 42,571 lots -> 4,257,100 shares during regular and extended trading hours
Yahoo finance and NYSE report about 5.58 million shares for that day (however they define day). Here some thoughts on where the difference comes from:
  • IBM is being traded on over 20 exchanges and SMART covers only some of them. Would be interesting to iterate over all valid exchanges and add up all results. That may get you pretty close to what others report. For example for exchange=NYSE, IBKR reports 10,960 lots -> 1,096,000 shares (rth=false) for that day.
  • It could also be that IBKR only counts "reportable" trades and does not count "odd lots". Just like the difference between the real time ticks RtVolume and RtTradeVolume. Not counting odd lots would seriously reduce the number of trades and could account for the gap.
Hope this points you in the right direction. Let us know in case you learn more

´³¨¹°ù²µ±ð²Ô

Join [email protected] to automatically receive all group messages.