Our adaptive portfolio system is continuously self improving and learning. It considers all conditions it has seen to date, across all time windows, and optimises for the conditions that we are experiencing.
This ability to dynamically adapt as markets change is our edge.
The Strategy Decay Problem
Traditional portfolio theory is based on combining a number of uncorrelated, low Sharpe components together in order to optimise the return potential for a given level of risk, at a single point in time.
However, this approach does not consider that correlations (and therefore risk and return profiles) will change over time. Using our out-of-sample forward testing method, we have proved that this approach is prone to rapid decline in performance over time. We call this decay.
Historically, this decay happened relatively slowly, and it was still possible to generate interesting returns for a period of time. But the rate of decay is rapidly increasing.
Where a portfolio constructed in 2007 would not start to delay until around 3 years later, decay now begins in as little as 6-12 months from implementation.
Adapting to these rapid changes is challenging for traditional CTAs, who typically focus on combining low Sharpe components, rather than identifying the specific components for the current conditions.
Our solution is an adaptive approach to portfolio construction, using genetic algorithms and neural networks to increase our speed of change and remove human bias.
Our Portfolio Builder is central to our portfolio selection and prediction system. This bespoke technology enables us to continually adapt to changing market conditions, continuously self-learning and adapting and protecting against decay.
Our portfolio builder reviews all the opportunities that are available to it at a point in time, and uses sophisticated filtering techniques in order to identify the trades with the highest probability of future returns. It then uses genetics to combine these into a portfolio.
Over time, the system changes and adapts based on what it learnt from the previous period, applying this knowledge to the next forward period prediction methodology in order to continually improve.
The chart below shows how, by continually optimising our portfolios, we are able to continually improve our return curve over time. This demonstrates how our investment proposition continues to generate alpha over an extended period, protecting against decay.
This ability to continually generate alpha and dynamically adapt as conditions change is our edge.