|
PureBytes Links
Trading Reference Links
|
Dennis, what method would you suggest for comparison? I recently installed
DTN/Ensign realtime feed and have been running sampled comaprisons of the CME
Time and Sales file to the Ensign Time and Sales on SP8M. I pick one minute
periods at random and then match all T&S for that minute. Allowing for a one
second lag in time stamping, and backing out Ask, Bid, and Inserts included in
CME report, I'm getting 100% exact match. At this point I'm satisfied that I'm
getting everything, however I haven't any data to compare for fast market
conditions.
>From all appearances, comparing T&S data seems to be a reasonable approach to
verifying the completeness of data transmission.
Earl
-----Original Message-----
From: Dennis Holverstott <dennis@xxxxxxxxxx>
To: Omega List <omega-list@xxxxxxxxxx>
Date: Tuesday, May 05, 1998 2:15 PM
Subject: Re: Tick Counts - Study Proposal
>You can't directly compare CME tick counts with DBC/Signal. If a future
>ticks 25,50,50,75 the CME will call that 4 ticks. The CME will filter
>the 'redundant' 50 tick and send 25,50,75 to the data feed. TS will
>report 3 ticks. The NYFE, on the other hand, will send all 4 ticks to
>the data feed.
>
>Keep in mind how these prices get into the system... some clerk is just
>typing them into a terminal. Does it really matter if he hits 50 once or
>twice? He is just giving his best guess of the last price looking down
>into a chaotic pit. How many times he hits the button has *nothing* to
>do with the actual number and size of the transactions. And if things
>are really fast, he may not even know what price to report.
>
>The only fair comparison is among the various data vendors.
|