[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Why bad realtime results? The plot thickens . . .



PureBytes Links

Trading Reference Links

"Hango Winspy" <hwinva@xxxxxxxxxxx> writes:
>At the end of Globex session around 9:15AM, the last bar usually
>don't have full 45 tick counts. It doesn't matter how many tick
>counts in the previous bar, at 9:30, it starts a fresh new bar in
>the real-time charting anyway. Later on, if you reload the same
>chart, it will fill up to the full 45 tick counts on that last bar
>before it moves on to plot the next bar around 9:30. Therefore, if
>the original unfill tick counts near the mid range of 45, say 20 to
>25, your relaod charts will be dramatic differernt at every turning
>point for the rest of the day. 

It will also start a new bar when the first real-time tick arrives,
whenever you load a chart, short-changing the last bar on the chart,
formed from historical data, to start that new bar, formed from
real-time data.  The result is, you can't simulate real-time tick-bar
charting with TS from historical data. 

IMO, it's correct to start a new bar at the beginning of a new
session; I just wish the charting was clever enough to do this when
loading historical data.  After all, it has all the info it needs to
act this way. 

Short-changing a bar in mid-session is simply a bug.  The charting
should keep track of the tick count in the last historical bar, and
if we're within a session, add real-time ticks to it. 

Jim