In theory, timing the fall and rise of financial markets is an impossible task. However, now more than ever, investment managers are looking to machine learning to optimize their investment strategies. By incorporating advanced analytics, new data sources, feature transformation, and customized quality metrics, Elder Research has established success identifying predictive signals and estimating the quality of an investment strategy.
Since 1995, Elder Research has worked with hedge funds and other financial institutions to evaluate, optimize, and create investment strategies with a measurable edge over the market.
Evaluate Investment Strategies
Because of the power of modern analytic techniques, it is often possible to find apparent predictive correlations in the market due to over-fit—where the complexity of a model overwhelms the data or, even more dangerously, from over-search—where so many possible relationships are examined that one can be found to work, but it is, in truth, only by chance. Wrestling with this serious problem, Dr. John Elder developed a powerful resampling-based method, called Target Shuffling to measure the probability that a result could have occurred by chance. It is far more accurate than t-tests, p-values, and other statistical metrics that don’t take into account the vast search performed by modern inductive modeling algorithms. With target shuffling, we can much more accurately measure the “edge” (or lack thereof) of a proposed investment strategy.
Further, to more accurately measure the quality of market timing, or style-switching strategies, Dr. Elder created the DAPY (days ahead per year) criterion. DAPY measures, in days of average-sized returns, the expected excess return for a timing strategy compared its benchmark fairly; that is, when exposed to the market the same percentage. The Sharpe ratio can be thought of as measuring the quality of a strategy’s returns; DAPY measures the quality of its timing edge. Together, they are much more useful than Sharpe alone. Most importantly, internal studies have shown DAPY to be better than Sharpe at predicting future performance.
Optimize Investment Strategies
Most investment strategies that we review are found to have serious problems. Discovering these flaws saves firms time and money. Our independent review helps clients invest with confidence before others discover and oversubscribe the talented emerging manager. Some strategies appear promising but can become even stronger with our assistance. In these intermediate cases, our skills in, and unique tools for, investment modeling and optimization can be engaged.
For example, John Elder’s PhD dissertation created a global optimization algorithm to solve problems in commercial and investment realms. The algorithm finds the global optimum value (within bounds) for the parameters of a strategy using as few probes (experiments) as possible. It was world’s best global optimization algorithm for over a decade.
Create Investment Strategies
We also design and create complete investment strategies. Our first investment strategy, in conjunction with the WestWind Foundation, performed extremely well for over a decade. For the nine years it was open to outside money it outperformed the S&P-500 index each of those years, with only two thirds of the risk. Its performance is illustrated in Figure 1.
Figure 1: Performance of our MVP 1 Hedge Fund vs. Leading Market Indices (log scale)
In its first years (prior to those shown in Figure 1 after it was open to outside investors) it achieved very strong returns with high volatility. Dr. John Elder recounts: “The sponsoring client was caught off guard by the volatility that attended the aggressive gains. It shocked me however that he seemed ready to call a halt to our huge success! Necessity being the mother of invention, I was able to devise the Target Shuffling technique and use it to clearly demonstrate that the system’s edge was very strong. The client then confidently increased his investment by an order of magnitude and invited outside investors. The fund took off and grew to capacity quickly. After a decade of success, with only minor updates to the core technology, 9/11 hit. Not long after, target shuffling revealed the market dynamics had changed more than we could accommodate, and the system’s edge was fading. So the fund was scaled back and later wound down. Astonishingly, every investor came out ahead, which is not the usual ending to hedge funds!”
Today, we are running proprietary long-short strategies that are so far meeting their performance targets and have weathered well the unusual markets of 2020. The strategies have been designed to have reasonable returns, low volatility, vast capacity, and very low correlation to other investment options, making them strong candidates for use in a large portfolio.
There are no guarantees in the world of investing, but we have found that expert practical application of advanced analytics can help investment organizations make the most informed and prepared choices possible. Our investment modeling consulting services can help you evaluate, optimize, and create investment strategies that deliver the highest return.