PureBytes Links
Trading Reference Links
|
>[Bob says]
>It would be interesting to know what you could not do in TradeStation or
>QuantStudio that prompted you to undertake such a major effort.
I can't say much about TradeStation vs. Uptick. Lots of people are
complaining about TS restricting them in many ways and them not being
able to do anything about the shortcomings.
The focus should be on the 1/4" hole as opposed to the 1/4" drill and I
do agree that it doesn't matter what tool you use to drill that hole so
long as you are an expert with that tool. People who are adept at trading
with TradeStation will likely not want to use anything else but there
should be tools for the rest of us.
I'm not looking to directly compete with TradeStation as that would be
futile. I'm looking to occupy the extensible and focused trading toolkits
niche. I'm looking to compete with QuantStudio and this is the product
that I know best.
I think that QS is a good idea badly executed. The main reason that
prompted the Uptick effort is my frustration with the owner and main
developer of QS. Most of the wrongs are business-related but there are
many technical ones as well. I would like to focus 100% on Uptick. My
savings will last a couple of months and I'll whip something up in that
time and then start to look for donations and sponsorships. I'd rather do
that than pitch to people something that they will likely be frustrated
with. So I will bring the skeletons out of the closet.
I have been selling QS for almost a year now. That's scores of users of
the full-featured product (there used to be a light version) and numerous
source code licensees. That's also many tens of thousands of dollars in
revenue and profits. That is a forum full of customer requests,
suggestions, complaints, etc. I have tried to righten the wrongs and
tried to convince Anton that certain approaches are better than others. I
tried to get involved technically as well. I realized that the stumbling
block is not the product but the owner behind it.
For insight into what QuantStudio was browse the years 2001 and 2002 at
http://groups.yahoo.com/group/smartquant . Most of the problems brought
up then are still not resolved today. Notice Joe Robe bringing issues up
all the time until giving up in despair and going off to build a toolkit
for his own use.
To see where QuantStudio is today watch replies to this thread http://
www.smartquant.com/forums/viewtopic.php?t=715 . These guys are probably
furthest away with QS. And they have a kick-ass C# programmer (they said
it) working for them. There have not been any QS success stories ever
since I started selling QS or since 2001 that it more or less started.
There will not be any success stories for a while either.
People are buying the QS source code and either integrating features into
their own systems or replacing the parts that are not working. The source
code sells for a pitiful 5K right now. If you have anything to do with
.NET and trading you should be buying now while the price is low. Contact
me if you want to buy. I'll reinvest the proceeds into developing Uptick.
Since I started a year ago documentation has not been touched, the
development team has not expanded beyond Anton and another guy that I
brought in. I have not seen any of the proceeds reinvested into the
business. The foundation of the product has not been improved. Instead of
implementing major customer requests or posting clear examples of how
systems should be coded (500 lines of code to do a simple system do not
count!) Anton chose to add advanced portfolio features or sugarcoat the
problems in nifty graphics. Let me start with the technical issues...
The data server is the foundation of any trading toolkit. There is no
data server in QS. It's foundation is still a set of Access tables albeit
those can be used on SQL server or Oracle now. The structure of those
tables does not withstand scrutiny. There are no indexes, etc.
Performance is not even on the horizon. The database layer in the code is
horrendous and full of holes. Objects do not refer to each other with
foreign keys. If something refers to instrument A then it will have the
string "A" as the reference. Forget about referential integrity.
I had a guy look into performance issues recently and he was able to
improve performance of storing trades in the portfolio many times-fold
just by moving the sorting of trades (in the code not in the database!)
to the moment of trade retrieval instead of storage.
There is no notion of a data server per se. If you want to capture
quotes, traders, etc. then you must whip one a server of your own or make
sure that your application stores trades and quotes while it's running.
You actually CANNOT use a separate data server right now because access
from two apps to the same "warehouse" blows up QuantStudio.
Bar compression issues have not been resolved as of yet. There's no easy
way of writing code that will run in backtesting and real trading without
any modifications. Anton did not support my scalper initiative. His take
on it was that there's no need for it since an use QS at it is and I'm
bringing in nothing new.
Yes, you can write assemble a car from parts each time you want to drive
but who does it these days? Hell, lets assume that you want to boost your
car's performance. You'll still want an easy way to do it. I know I will.
I'll go buy engine chips to replace the standard ones, I'll buy
performance parts that plug in. I won't go and design an engine of my
own. I don't think writing 500 lines of code for a simple moving average
cross-over that works in backtesting and realtime is what I want to
entertain myself with.
The whole architecture of QS is not geared towards usability of
performance. The code is full of holes and crashes frequently. There are
no unit tests to speak of. Even a year after QS started selling
officially Anton still treats it as an open source project of sorts.
Notice that I say of sorts. He's gathering development suggestions from
people, patches that people fixing the source code submit, etc. Except
you are not getting the code free. You are paying hundreds of dollars for
the privilege of getting frustrated.
There is no documentation. There is no documentation. There is no
documentation. I can go on but do I need to? I have figured that to fix
up QS I will need to write documentation, do technical support, patch
holes in QS, add features that customers want but that Anton is not
adding and then sell QuantStudio. All the while struggling with Anton
over the direction that the product should be taking as well as other
business issues. I figured that I'll have much better luck letting market
forces deal with Anton and using those same forces to my advantage with
Uptick.
I haven't seen a project like this before. But I believe in achieving
something here. Even against TradeStation. I particularly like the
following quote from this article http://news.com.com/2100-7344_3-
5206671.html?tag=nefd.hed :
"How do you go up against the titan? The barriers to entry are high, the
cost of entry is high," Barrenechea said. "The only way to compete
against that kind of existing technology is via a grass-roots movement.
If you have got a million contributors to doing something semi-
orchestrated, you can revolt, and Linux is a perfect example of that."
Cheers, Joel
--
http://www.wagerlabs.com
Home of the Uptick trading engine
|