|
PureBytes Links
Trading Reference Links
|
We may be talking about different events -- I've only observed the following
cases:
1. First tick of the day is bad -- either zero or some value far from the
actual Open.
2. Some other tick during the day is bad.
The next question is whether or not the bad tick falls on the starting
boundary of one of the 60+ tick blocks. If it's the first tick of the day--
it does. If not, you have no way of knowing.
If it does **NOT** fall on the boundary, you can correct it and there is no
further problem. One problem here is that --other than the first tick of
the day-- you have no way of knowing, so it is best to correct it as opposed
to deleting it.
You say that the bad tick doesn't plot as bad -- in the case that I
explained-- it **DOES** plot as bad-- otherwise, how would you know that it
needed to be corrected?
It is important to keep in mind that there is a difference between the RT
data that is being displayed on your charts and data that is drawn from the
database. Your RT charts may be fine, but the data in your DB may be
corrupted-- due to a bad initial tick.
If the above scenarios don't address what you're seeing, I would need to
know more specifics.
There is one other case that I've seen that corrupts data, but it is even
more difficult to explain. It has to do with a value of an incoming data
stream exceeding 32767 -- this is the maximum value that can be represented
in two bytes of memory (or 16 bits). This occurs in the Advance/Decline
Indices almost daily. When the value exceeds 32767, the representation
turns negative and starts to descend, as opposed to ascending-- which is
what is actually happening in the real world. The 32767 may be represented
as a decimal value in TS (i.e 327.67), but it is broadcast as an integer.
If this example further confuses the issue, ignore it.
_____________________________________________-
At 07:39 PM 10/3/98 -0700, you wrote:
>bear with me ron. your explanation still doesn't jive with my
>observations.
>
>you're assuming that the tick in question is bad, period. i question
>that inference. i'm saying that you don't or can't know.
>
>question: why doesn't the "bad" tick plot on the tick chart as bad,
>but as a good or correct tick at the time it was received?
>
>TJ
>
>
>---Ron Augustine wrote:
>First of all the tick *DOESN'T* plot correctly on your chart--
>
>yes, it does on my charts. your comment above is the source of my
>confusion
>
>
|