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

Re: EL Question - Missing Data Condition



PureBytes Links

Trading Reference Links

I did not know that EL keeps the last value of data, forever.
Thanks for your insight into this endearing EL feature.
Time to go chart the Penn Central stock price :-).


----- Original Message ----- 
From: "Gary Fritz" <fritz@xxxxxxxx>
To: <omega-list@xxxxxxxxxx>
Sent: Tuesday, March 12, 2002 10:39 AM
Subject: Re: EL Question - Missing Data Condition


> > >I want to detect those bars on which Data2 is missing:
> > Try:
> >   if Time of data1 <> Time of data2 then ....
> 
> Nice, Bob.  Very clean solution.
> 
> ndtrader and others who run into situations like this:  the best way
> to find a solution (other than having somebody solve it for you on
> the Olist :-) is to poke around and see how things work in TS.  Try
> charting things using the Custom N Line indicators.  If you'd tried
> that, you would have seen that your guesses of "Close Data2 = 0" &etc
> wouldn't work.
> 
> For example, say you plot something like this in "Custom 2 Lines":
> 
>   Input1:  Close
>   Input2:  Close of data2
> 
> You would see that the Close of data2 doesn't go to zero when there's
> no bar in data2; it just repeats the previous value.  You might not
> have gotten the insight that Bob did, of using Time instead of price,
> but you would have avoided experimenting with a dead end.
> 
> Bob's solution works very nicely because Time (unlike price) is
> guaranteed to change on every bar where there's data.  It wouldn't be
> guaranteed to change if this was a *tick* chart, but you can't use
> multiple data streams with tick charts.  Furthermore, if there's data
> on both data streams, Time is guaranteed to be the same on both data1
> and data2 (assuming they use the same bar size).  If Time on data2
> falls behind Time on data1, you know there must be a gap in data2.
> 
> Now, just for completeness, what happens if there's no data on data1
> but there is on data2?  There you're out of luck.  Indicators and
> systems only get executed at the close of a bar on *data1*.  So if
> there's no bar on data1, your indicator code won't see it.  It won't
> get called until there IS a bar on data1.  So Bob's test won't do you
> any good in this case -- try the "Custom 2 lines" test again.  You'll
> see that Time on data1 & data2 is always the same (assuming you never
> have gaps in data2), because the indicator code doesn't get called
> during the gaps in data1.
> 
> Gary
> 
>