Optimal f and Portfolio-Level Application: The Next Frontier
The Black Book of Day Trading Strategies
1,000 complete strategies · 31 chapters · Full trade plans
Introduction
While the application of Optimal f to a single trading system is a effective tool for maximizing geometric growth, the true potential of this methodology is realized when it is extended to a portfolio of multiple, uncorrelated trading systems. The portfolio-level application of Optimal f, often referred to as "Portfolio f," introduces a new layer of complexity and opportunity. It allows for the simultaneous optimization of position sizing across multiple systems, taking into account their individual performance characteristics and their correlations with one another. This article explores the conceptual framework of Portfolio f, the mathematical challenges of its implementation, and the profound implications it has for professional portfolio management.
From Single System to a Portfolio of Systems
The transition from a single-system Optimal f to a multi-system Portfolio f is not a simple matter of averaging the Optimal f values of the individual systems. To do so would be to ignore the important role of correlation in portfolio construction. The goal of Portfolio f is to find the optimal allocation of capital across multiple systems to maximize the geometric growth of the entire portfolio.
This requires a more sophisticated mathematical approach that considers the covariance matrix of the returns of the individual systems. The covariance matrix captures not only the volatility of each system but also the degree to which they move in relation to one another. A portfolio of uncorrelated or negatively correlated systems will have a lower overall volatility than a portfolio of highly correlated systems, and this will be reflected in the Portfolio f calculation.
The Mathematical Challenge of Portfolio f
The calculation of Portfolio f is a non-trivial mathematical problem that typically requires the use of specialized software or programming. The objective is to find the set of weights (f1, f2, ..., fn) for each system in the portfolio that maximizes the geometric mean of the portfolio returns. This can be expressed as:
Maximize: G(f1, f2, ..., fn) = ( (1 + R_p1) * (1 + R_p2) * ... * (1 + R_pm) )^(1/m) - 1
Maximize: G(f1, f2, ..., fn) = ( (1 + R_p1) * (1 + R_p2) * ... * (1 + R_pm) )^(1/m) - 1
Where:
- G is the geometric mean of the portfolio returns.
- fi is the fraction of capital allocated to the ith system.
- R_pj is the return of the portfolio in the jth period.
- m is the number of periods.
The portfolio return in each period is the weighted average of the returns of the individual systems:
R_pj = f1 * R_1j + f2 * R_2j + ... + fn * R_nj
R_pj = f1 * R_1j + f2 * R_2j + ... + fn * R_nj
Where:
- R_ij is the return of the ith system in the jth period.
This optimization problem is subject to the constraint that the sum of the weights must be equal to 1:
f1 + f2 + ... + fn = 1
f1 + f2 + ... + fn = 1
A Simplified Example: Two-System Portfolio
To illustrate the concept of Portfolio f, let's consider a simplified example of a portfolio with two uncorrelated trading systems, System A and System B. We have the following historical data for each system:
| Period | System A Return | System B Return |
|---|---|---|
| 1 | +10% | -5% |
| 2 | -5% | +12% |
| 3 | +15% | -8% |
| 4 | -8% | +18% |
| 5 | +20% | -10% |
We can now calculate the geometric mean of the portfolio for different allocations to System A and System B:
| Allocation (A/B) | Portfolio Geometric Mean |
|---|---|
| 100/0 | 5.2% |
| 80/20 | 6.1% |
| 60/40 | 6.8% |
| 50/50 | 7.1% |
| 40/60 | 7.2% |
| 20/80 | 6.9% |
| 0/100 | 5.8% |
As the table shows, the portfolio geometric mean is maximized at an allocation of 40% to System A and 60% to System B. This is the Portfolio f for this two-system portfolio. It is important to note that this optimal allocation is not simply a reflection of the individual performance of the systems. System A has a higher arithmetic mean return than System B, but the portfolio benefits from the diversification provided by the uncorrelated nature of the two systems.
Conclusion
The portfolio-level application of Optimal f represents the next frontier in sophisticated risk management. By extending the principles of geometric growth maximization to a multi-system framework, Portfolio f provides a effective tool for optimizing capital allocation and enhancing long-term portfolio performance. While the mathematical challenges of its implementation are significant, the conceptual framework of Portfolio f offers a clear and objective approach to portfolio construction. For professional traders and investment managers, the pursuit of Portfolio f is a journey toward a more robust and mathematically sound approach to the art and science of portfolio management.