Model Building
Seminar
Building successful market models can be one of the
most gratifying accomplishments a quantitative analyst can ever achieve. The precision
of thought and training in abstraction that one learns from quantitative
analysis are intellectual tools of profound importance. Training a computing
system to “behave” appropriately, within a non-stationary
market environment, can be as intellectually gratifying as it can be monetarily
rewarding.
To be sure, there are age-old principles which must
be observed. And, frequently, there are better
methods and many useful
techniques that can be employed in building successful models. The Seminar Syllabus
only indicates a range of topics that can be covered in a seminar at your
location. Seminars are tailored to client needs, usually lasting for three or
four days, and are taught with actual modeling software using real market data.
While the principles and techniques are applicable to intraday models, the class
examples are all accomplished using end-of-day data. The trading environment
that is contemplated is that of using data gathered near the market close to
effect trades on or before the close.
The purpose of the Seminar is to train entry and
intermediate level quantitative analysts in the methods and useful techniques
of model building with a view to implementing fully automated
trading systems. Market models should not be “black boxes.” To the
contrary, if a model’s components are not understandable, then it should be
viewed with suspicion. Generally, the daily computation of a market model
proceeds as follows:
1.
Various
raw
market time series data are used as inputs to the model building process.
2.
Mathematical
transformations are used to
pre-process those data.
3.
Mathematical
and/or statistical functions are applied to the pre-processed inputs, resulting
in statistical tables and/or a more well-behaved oscillator time series.
4.
Trading
signals (indicators) are generated from the more well-behaved time series.
5.
The
trading signal indicators are then combined with other indicators to produce a
more robust, multiple viewpoint trading model.
6.
Models
for different market instruments may then be evaluated to determine their
respective allocations within a portfolio.
With that overall structure in mind, the
construction process might include the following steps:
1.
Selection
of a model
type to accomplish the idea that is contemplated;
2.
Selection
of raw
input data appropriate to the type of model desired;
3.
Mining
of the time series data to determine the most useful mathematical / statistical
transformations to accomplish the purpose of the model;
4.
Formulation
/ Prototyping of the step by step process;
5.
Building
of the Training Procedure;
6.
Building
of the Operations Procedure;
7.
Testing
the model within a cross-validation context;
8.
Analyzing
the results with respect to:
·
Performance
·
Parameter
Sensitivity
·
Contribution
to Portfolio
9.
Integration
of the model into a portfolio.
The models constructed would usually be trained
periodically (weekly or monthly) and executed daily, for the purpose of generating
Long, Short, or Cover to Cash trading signals.
The overall nature of the Seminar is practical rather than theoretical. Higher level mathematics is not a pre-requisite. The Seminar proceeds with a minimum of theoretical mathematics. A good background in algebra is required, some elementary differential calculus or time series analysis would be a plus, and common sense is essential. The purpose is not to teach concepts related to the financial engineering of exotic derivative products, complex fixed income structures, mortgage backed securities, etc. And, while option data is sometimes used as indicator input data, the pricing of options is not a covered topic.
The Importance
of this Website to Your Business – As the markets become more volatile, you would do well to train your
quants to protect your portfolio against the ill effects of nonstationarity.
Exogenous Data Based Models – The good and bad characteristics of Exogenous Data. (Don’t
miss the interesting visualization of some SPX Index Option data.)
Visualization of Exogenous Data – In case you missed it above.
Quantitative Analysis Platform – A user-friendly modeling platform for improving
the productivity of quantitative analysts.
Overview – Advanced Automated trading Systems.
Consulting Services – Helping your quantitative analysts deliver a
better product for your clients.
Trading Model Building Services – Continuous and Discrete Models, using Price or
Exogenous Data.
Quantitative Analysis Training Seminars – Topics covered in typical training seminars.
Model Validation – A Catch-22 in the struggle between the Central
Limit Theorem and the “Law” of Requisite Variety.
Non Trend-Following, Non Technical Analysis Methods – The difference that a non-price market view can
make in your portfolio’s success.
© 1997-2004 Thomas W. Wright. All Rights Reserved