PureBytes Links
Trading Reference Links
|
Dear Bill,
The following is an E-mail copy of a letter that I am sending to you for
the record by Certified mail:
Mr. Bill Cruz, President
Omega Research, Inc.
9200 Sunset Drive
Miami, Fl 33173
January 13, 1998
By Certified Mail/Return Receipt Requested
Subject: Bad tick editing
Dear Bill:
On several occasions in the past two years I have communicated to
you and your associates my concerns about TradeStation's ability
to save and display the real-time data in a reliable and usable way.
After TradeStation had been enabled with the capability to edit bad
ticks for symbols generating up to 33,000 ticks a day, two major
problems have remained:
(1) The inability to handle a seriously bad tick, received between
midnight and the start of a trading session. When TS receives a
tick like that, it alters the value of all subsequent ticks. What that
means is that a minor datafeed or reception problem, manifested
in a couple of extraneouos bits of data, may render Omega users'
data practically worthless. I do not want to contemplate what would
happen in case of a major datafeed problem.
As your own refresh data show, no one is immune to such data
destruction. Among the symbols I follow, I found eight such cases
over the past 30 days on your Web site. For instance, on Jan 5,
in addition to INTC and SPX, two more symbols were corrupted by
a bad initial tick -- Cigna (CI) and Weyerhauser (WY). Kellog's
data (K) were destroyed on Jan 2. Delta (DAL) and Burlington
Northern (BNI) were hit on Dec 31, and Cigna (CI) on Dec 11.
At present, the only way to restore the corrupted data is to do it
manually. However, that takes an enormous amount of time, so, for
practical purposes, it is not feasible.
(2) The inability edit bad ticks for stocks that generate more than
33,000 ticks a day. The consequences of such a tick occuring are
also severe -- an uneditable bad tick way out of range makes Omega
user's data useless. Also, each time editing of such tick is attempted,
the computer crashes, interrupting data reception and market
monitoring in real-time.
Now that Oracle was hit in this fashion in October, the question is
how often that will happen to other stocks.
As the trend in the growing volume of ticks indicates, many more
data collections will be destroyed. Look at what has happened to the
level of 13,000 ticks per day. A year ago that barrier seemed
unreachable. Now, however, the 13K days are commonplace. My
records show the following incidence of crossing the 13,000 tick
threshold by major stocks between May 5 and December 31:
AAPL - 3, AMAT - 18, ASND - 20, AMGN - 2, BA - 1,
COMS - 16, CPQ - 4, CSCO - 13, DELL - 32, INTC - 90,
MSFT - 14, NOVL - 1, NKE - 1, OXHP - 4, ORCL - 6,
SUNW - 5, WCOM - 3.
Given these trends, one can assume that the above stocks could be
crossing the uneditable 33K barrier with similar frequency within
a few months. As a result, they might become unsuable for stock
and option trading.
When this happens, will Omega continue to recommend its clients
to delete the days affected like it did recently in Oracle's case?
Would your company really have the gall to suggest to your clients
to delete 90 days of data for INTC -- or 32 days of data for DELL --
just because your products have an insufficient tick editing capability?
It should be obvious that without the full capability to edit bad ticks
and TS-corrupted data, Omega products such as TradeStation,
SuperCharts and OptionStation can not be considered functional for
stock and option traders. Using them may cause a loss of data,
render the existing data unusable for an extended period of time,
and generate erroneous signals.
I have heard that the editing problem would be resolved in a new
generation of Omega's products such as TradeStation 5. That,
certainly, is the good news. However, what about the products you
are selling today, such as the existing version 4? These products
may be unsuable tomorrow, not because of natural obsolesence,
but because of a known software defect. Therefore, the issue of
bad tick editing should be resolved now.
One possible solution, suggested by Mr. Rodney Grisham, is to
offer a simple utility that would allow Omega's clients to decompress
and edit the tick data externally. I am sure other alternatives are
available. For instance, you might decide to keep the omz
compression algorithm for the EL functions, while switching to a
publicly available compression method -- such as PKzip -- for the
data files.
At any rate, something has to be done. This is a serious problem
that will not go away. So, I urge you again: Please correct it so that
TradeStation -- a fine program for future traders -- can become
a fine program also for those who follow stocks and options.
I wish you and everyone else at Omega a healthy and prosperous
year ahead.
Sincerely yours,
Igor Uhrik
|