8 CHAPTER ◆ 1 Introduction
3. Customer. All projects must have a strong seed capital provider support.
4. Budget/Cost. Experience and good historical data from similar projects reduce
this type of project risk.
5. Schedule. Trading/investment system development teams must be part of devel-
oping and modifying the project schedules; management must not impose them
externally.
6. Project Content. Existence of documentation of trading system specification
and design will ensure that knowledge will not be lost.
7. Performance. Complete performance testing of all modules and interfaces is
crucial. Inadequate testing contributes to project risk.
8. Project Management. Managers need experience and understanding of trading/
investment system project management processes.
9. Development Process. New trading/investment system development processes
should be focused on quality assurance and analyses of alternative processes.- Development Environment. The lack of adequate physical environment and
devel opment tools contributes to project risk. - Staff. Clearly, having an experienced and proven product team reduces project
risk. - Maintenance. This category of project risk includes maintenance and vendor issues,
and software bugs after the trading/investment system has been implemented.^15
As David Hulett points out, “ project managers who assess the risk that the project will
overrun its cost estimate or schedule, or will fail to meet performance objectives or speci-
fications often improve their likelihood of a successful project. ”^161.7. Buy versus Build
If you work in the front office, you may say, “ we buy all of our indicators and tools from
third-party vendors; they take care of all that development process and testing stuff for
us. ” But even if this is the case, ask yourself a simple question, “ If a third-party piece
of software contained an error—say, it calculates the price of options incorrectly—how
much would knowing that be worth? ” You should benchmark off-the-shelf components
in order to isolate the performance of individual pieces of a trading/investment system.
Benchmarking will expose their strengths and weaknesses, maybe even errors and prob-
lems in the underlying assumptions used to create them.
Here is an analogy. Suppose you want to race stock cars. The layman with little or
no understanding of engineering might test drive several muscle cars at the local dealer
and purchase one with good performance. Now, no one would be under the impression
that such a car and a driver could compete with professional racing teams. Stock cars are
in fact highly customized. Professional racing teams have engineers who understand the
workings of race cars inside and out. They know what components to buy and what to
machine themselves and how to put the pieces together to maximize performance. Part A
may be used on hot days, while part B may be used on cold ones. Automotive engineers
even use computer simulations that test different gears and shift points on the track for
faster speeds. Without this depth of understanding, though, you cannot even begin to