|
PureBytes Links
Trading Reference Links
|
Wondered if sale recording might not be part of the explanation as most realtime
software packages include time and sales reporting which would be impossible
without recording each sale rather than reporting price changes. I've performed
limited comparisons of identical 15 minute time slices of the daily CME "Volume
By Tick" and "Time And Sales" files (http://www.cme.com/market/prices/ftp/) and
have concluded that they are not derived from the same data stream. In
particular, the Volume By Tick file includes numerous bad prices throughout each
day such that bars constructed from the data include obviously bad high/low
prices. The CME does build seperate files for globex and non-globex trading. I'm
writing a program to construct/reconstruct intraday files from the CME data and,
based upon my examination of both files and the earlier availability of T&S
files, have elected to use the T&S data. I'm still trying to learn more about
the files but the CME has yet to respond to inquiries.
Earl
-----Original Message-----
From: Randy Murphree <murpran@xxxxxxxxxxxx>
To: Carroll Slemaker <cslemaker1@xxxxxxxx>
Cc: Omega User Group <omega-list@xxxxxxxxxx>; eadamy@xxxxxxxxxx
<eadamy@xxxxxxxxxx>
Date: Monday, May 04, 1998 8:50 PM
Subject: Re: CME Tick Data
>Carroll:
>A quick perusal of my tick file for today's SP8M reveals exactly what you
>describe. In one instance the price remained the same for 6 consecutive
>ticks, then 1 tick up, then returned to the original price for 6 more
>ticks. I have discovered, however, that although I only chart the daytime
>SP, ticks are collected for the night market as well. Thus the high tick
>count.
>Randy
>----------
>From: Carroll Slemaker <cslemaker1@xxxxxxxx>
>To: Earl Adamy <eadamy@xxxxxxxxxx>; murpran@xxxxxxxxxxxx
>Subject: Re: CME Tick Data
>Date: Monday, May 04, 1998 7:26 PM
>
>Hi Earl -
>
>I'm using the "volume by tick" file because it is a "tick" file (a
>report of price changes) whereas the time-and-sales file is, presumably,
>a list of all individual sales and would, therefore, contain groups of
>consecutive sales occurring at the same price.
>
>Incidentally, I notice periods of time on my GLOBEX chart for the S&P
>where the plot is horizontal - inspection of the data window shows
>groups of consecutive same-price "ticks". I wonder if BMI perhaps
>transmits all sales (for the day session) rather than just price
>changes. I got a mail message from Randy Murphree, using
>satellite-delivered BMI, who reported that his tick counts for 4/29 and
>4/30 were higher, even, than the CME figures! (5983 and 5530
>respectively, compared with CME figures of 5664 and 5470.)
>
>Regards,
>
>Carroll
|