[ While reading the following, ponder the word "different" and
the impact in diversifying your portfolio indicated by the red titles. ]
Not Technical Analysis
Eighteen years ago I decided to approach the markets from a different viewpoint. Now, my automated trading systems are not in any way related to technical analysis (no stochastics, no moving averages, no Elliott waves, no trend lines, no bands, no MACDs, no cycle analysis, no candlesticks, no entropy, no Gann angles, etc.). I do not dislike technical analysis, but that is what everyone else is doing; I just want to be different. With hundreds of thousands of computers analyzing price data, I believe that any opportunities to be exploited due to price moves have been arbitraged away. Consequently, I process the data that is the most difficult to obtain, and in an extraordinary way. (My work does not embrace "chaos theory" -- it is not THAT different.)
Nonlinear
My approach is nonlinear. I tried linear methods, but get better results with nonlinear methods.
Nonstationary Market Assumptions
My approach is based on non stationary market assumptions. That is, I believe that in addition to price moving around as it reacts to various forces, the very underpinning structure of the market is also moving around. If the structure itself is moving around, then the usefulness of any parameterization must have a half-life. If the usefulness of a parameter set is decaying, then the parameters must be made to adapt.
Not a "Black Box"
Consequently, what your quantitative analysts build must not be a "black box." Only an understandable "process" can address the nonstationary market assumption. An indicator cannot. A model can, but only if properly constructed. System maintenance must be transparent enough such that parameter sets can be tracked as they migrate in parameter space.
Not Based on Price Data—Consequently, Not Trend Following
The very idea that prices react to news and other market forces demands that price be on the effect end of a causality spectrum. Furthermore, most logicians since Aristotle have believed that an effect cannot be its own cause. Price cannot be its own cause, though the herd-boundedness and self-fulfilling prophesy arguments have some merit. Price cannot be at both ends of a causality spectrum. If price is acting as a recursive force at the extreme effect end of the spectrum, it is merely a small force dominated by news, corporate events, and many other market forces. I only use price to measure performance and to construct objective functions. Consequently, without using price, it is difficult for my technology to be trend following - isn't it? (Candidly, some of my clients have asked me to build price-based models for specific purposes; I have built them when my clients have requested them.) On the other hand, I do believe that econometric and fundamental data is causal to price. However, my systems do not use a fundamental approach, since data on the extreme causal end of the spectrum is difficult to obtain in daily bars. I would use econometric data if I could get GNP, Retail Trade, Trade Deficit, etc. data in daily bars.
It follows that, in terms of causality, one should look for data as far from price as is possible. As one moves farther away from price, one must shift to data surrogates in order to obtain daily values. For example, daily inflation related econometric data does not exist. But one could use CRB (or perhaps gold in times past) data as a surrogate. Likewise, sentiment, sector funds, sector indices, and other data may act as surrogates for leading economic indicators, etc.
Different Input Data for Indicators
My buy/sell signals are a function of sentiment, volatility, inter-market, monetary, option, basis, volume, open interest, and other data which is exogenous to price. For example, I believe that the option players are the smartest players in the markets. Consequently, I use the sentiment expressed in the option matrix data to trade the underlying securities. While YHOO will have an Open, High, Low, Close, and Volume values at the end of day, the option matrix may contain hundreds of active contracts, each with call and put premiums, volume, and open interest.
Not the Usual Moving Average Based Technology
All of the foregoing non-price data are used, together with proprietary transformations (sample graphics), to create the independent variables used to synthesize a dependent variable (objective function) using statistical pattern recognition (not chart patterns). The dependent variable (objective function) oscillates about zero, and is constructed to trade nearly perfectly. It does not try to trade the noise in price, so it misses a few trades by design. The synthesis is done with proper respect for the number of degrees of complexity involved. I do not "violate" the central limit theorem or the "law" of requisite variety (from cybernetics).
Not a Back-Tested Curve Fit
I use one of three kinds of cross-validation and am very conscious of statistical validity issues. I avoid over training, curve fitting, linear regression, efficient market and random walk hypotheses, and other known trading system abuses.
Since 1986, I have restricted my studies to common stocks traded on U.S. exchanges (NYSE, AMEX, NASDAQ), fixed income instruments, currencies, mutual funds, and futures contracts on the S&P 500, US Treasury Bonds, Crude Oil, Natural Gas, Heating Oil, Euro (Deutschemark, Swiss Franc), Japanese Yen, British Pound, Gold, and Silver. I have a healthy respect for options and, though using option data, my technology does not embrace option strategies, option pricing, or "extreme" derivative products.
(Probably) Not Correlated to Your Portfolio
My goal has been to help smooth out portfolio equity curves by contributing non correlated equity curves which lower the co-variance of the portfolio. Risk Management involves at least that, but much more. I diversify at the raw data level (different inputs for a single indicator), at the indicator level (different transformations for sub-indicators), at the model level (multiple indicators trading at multiple trading frequencies combined into a model), at the instrument level (multiple market instruments in a portfolio), and at the portfolio level (varying allocation techniques). My equity models trade at the swing cycle (every 2 to 7 days). My futures trading frequency is about 7,000 round turns / $1,000,000 under management. I believe the principles outlined in BARRA's paper (now available in book form) on "The Fundamental Law of Active Management," by Richard Grinold (1989). I am not trying to identify and kill a Bull or a Bear; I prefer killing a hundred Muskrats. I use many of my proprietary mathematical transformations for indicators and the Sharpe Ratio as a risk-adjusted measure of return.
I have aimed at different targets—that many people don't even see—and hit
them.
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.
