|
PureBytes Links
Trading Reference Links
|
Be my guest with respect to posting my results on your Web site.
The delay cannot, of course, be negative. You cannot receive the data
before they are entered at the exchange! That is why I was careful to
define "lag" as simply the difference between the two time stamps - that on
the CME's data and that on the user's. I trust the CME's time, I do NOT
trust the user's.
Note that the user's time stamp can come either from the feed or from the
user's own system clock, depending upon the feed and software being used.
And if the stamp is from the user's clock, the result will depend upon
whether that clock is simply free-running (and perhaps greatly in error),
synchronized to the feed, or synchronized to a time server. (Incidentally,
I would disregard any data shown for the half-hour following the 16:00
normal closing time.)
Finally, note that the time from the feed does NOT originate from the
exchange (at least not for DBC feeds). It is attached by the data vendor at
the point of distribution and some results (those showing lags of 0 or -1)
suggest that they are "adjusting" the time in an attempt to approximate the
CME's time stamp.
Personally, I want to know the time when I receive the data and, if the
software provides that alternative, I would opt to create the time stamp
from my system's clock (which is periodically synched to a time server).
This is the only setup under which the reported lag will, in fact, be the
true delay.
Carroll
----- Original Message -----
From: Mike D <Mike@xxxxxxxxxxxxxxxx>
To: Carroll Slemaker <cslemaker1@xxxxxxxx>; <omega-list@xxxxxxxxxx>
Sent: Wednesday, August 11, 1999 7:17 PM
Subject: Re: Feed Performances Under Stress!
> great work Carol. a few questions.
>
> 1. how can lag be a negative number? price can not lead the CME???? so all
> feeds lag somewhat!
>
> 2. I have BMI cable. and I compare TICK counts with other buddies many
> times and I am always less then them? I have Tradestation and another
> program also!
>
> I heard eSignal was supposed to be one of the best.
>
> please keep us posted and this is great work and should be published. I'm
> making a S&P future section on my web site and would love toi add your
data
> and explinatoins.
|