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

Re: Corrupt data



PureBytes Links

Trading Reference Links

According to list member Bob Scott, this is caused by the price 
movement exceeding the 15-bit (plus/minus) limit from the reference
tick.  If the price scale is 1/1000, a movement greater than 32.768 
will cause the problem.  (This causes problems with the Net stocks.)
If 1/100, greater than 327.68.  On a fine day like today (Dow +420), 
you need 1/10 price scale. 

At his suggestion (six months ago) I changed my price
scale to 1/100 for stocks instead of 1/1000.  No more problems.
And I changed the INDU to 1/10.    Do the changes off-line or weird
things may happen.

don


> Subject:           Re: Corrupt data
>    Date:           Wed, 15 Mar 2000 18:15:33 -0800
>    From:           "BobR" <bobrabcd@xxxxxxxxxxxxx>
>      To:           <omega-list@xxxxxxxxxx>
>
>Yes, I am seeing the same thing.  I am also seeing the open stuck at the
>open for February 28th on the daily data but not the intraday data, 9854.66.
>Each day I have to manually change it.
>
>BR>
>
----- Original Message -----
From: "Gary Fritz" <fritz@xxxxxxxx>
To: <omega-list@xxxxxxxxxx>
Sent: Wednesday, March 15, 2000 5:59 PM
Subject: Corrupt data


> Has anyone else (on BMI, TS4 build27) been seeing corrupt data lately?
>
> I've been seeing a lot of glitches in INDU -- places where the price
> suddenly drops (or jumps?  not sure if it's done that) by a few
> hundred points, stays offset for a while, then pops back up to the
> right price.
>
> Today I didn't see any problems in INDU.  I have a 1min chart that
> plotted all day with no problems, and it still looks fine.  But if I
> create a *new* 1min INDU, it's messed up.  At 3:11ET it drops from
> 10135 to 9481, bounces between those levels until 3:15, then stays
> down in the 9400-9500 range until 3:49.  During that time the price
> waveform looks right, except it's 600+pts low.  At 3:49 it popped
> back up & down a few times, then resumed at the 10136 level, and
> continued properly for the final 10 minutes of the session.
>
> Since it plotted correctly in realtime, apparently I received the
> data correctly.  But somehow it got stored wrong.  So I'm guessing
> this is NOT a BMI problem (bad ticks etc), but some kind of problem
> on my end.  Any idea what causes this, and how to fix/avoid it?
>
> Thanks,
> Gary
>
>