Re: Clock

[Ruth Ruby <RuthRuby@xxxxxxx>]

Thanks for your expertise in checking out this whole situation about tic
by tic price reporting-,my question to you is-does it Actually affect
YOUR trading situation?
It does mine.

Good Trading,
             Ruth
Carroll Slemaker wrote:
> 
> Some people seem to believe that the time associated with a real-time
> price update is absolute - that is, that ALL people receiving this
> update from ALL vendors will see the same "time stamp" attached to the
> price.  They believe that all such recipients will see the same set of
> price updates included in any particular "bar".
> 
> This is not true.  After considerable investigation I finally unearthed
> the following:  the CME (and, presumably, other exchanges as well) does
> NOT transmit a time-stamp with their data.  They claim that their goal
> is to transmit the data within a couple of seconds of its entry into
> their system and, therefore, a time-stamp is unnecessary.  It turns out
> that the time-stamp is added by the data vendor.  Furthermore, for DBC
> Signal at least, the time is added at the individual broadcast sites
> with the result that Signal data collected at different sites around the
> country (or world) will often have different time-stamps.
> 
> I have confirmed this by two comparisons:  I have compared my collected
> data for an S&P contract with a time-&-sales file downloaded from the
> CME (which contains entry times to the second) and found that the times
> on my data differ from the CME's.  Even more important, I have compared
> my collected data with that downloaded from Omega, who collected their
> Signal data using the same distribution medium (FM radio at the time I
> made this comparison) as I did, but located in Florida instead of
> Southern California.  Guess what?  My data compared exactly with theirs,
> tick for tick, except that the times differed.
> 
> The changeover from one minute to the next occurred at slightly
> different places in the two tick streams, sometimes different by as many
> as four or five ticks.  The significance of this is that one-minute or
> five-minute bars which I produce might begin and end with different
> ticks as compared with similar bars which you produce.  This could
> result in quite different system results, trend lines, and so on.
> Likewise for historical data obtained from a data vendor or other
> source.
> 
> So, if you attach special significance to the first or last tick of any
> particular bar, give careful consideration to the preceding.
> 
> Carroll Slemaker