Unique Edge

Our investment strategies are fundamentally based on the theoretical framework of the Fractal Market Hypothesis, as described by Benoit Mandelbrot, in order to describe the way financial markets behave. The Fractal Market Hypothesis is an alternative investment theory to the widely utilized Efficient Market Hypothesis, which fails however to explain why markets experience bubbles and crashes.

The Fractal Theory, developed by the Polish-born French and American mathematician Benoit Mandelbrot, is a powerful and revolutionary mathematical apparatus describing both social and natural processes. It has been applied successfully in various disciplines with great success: from medicine to describe the lungs formation to complex computer-assisted image processing. Based on this theoretical framework, the Fractal Market Hypothesis has been formed, which dictates that financial markets follow cyclical and repeatable patterns depending on the presence or absence of available liquidity.

The Fractal Market Hypothesis regards each market fragmented in different investor groups, with each group having a different investment horizon. According to this, investors share similar risk and reward levels – once an adjustment is made for the scale of the investment horizon – and these similar ratios explain why the frequency distributions of returns look the same across different investment horizons. In this context, markets exist to provide a stable, liquid environment for trading and they remain stable when many investors participate and have many different investment horizons.

However, the Fractal Market Hypothesis seeks to explain investors’ behavior in all, even volatile, market conditions, something the popular Efficient Market Hypothesis fails to do. According to it, during stable economic times information does not affect investment horizons and market prices. There are various numbers of long-term investors who balance the numbers of short-term investors, ensuring securities can easily be traded without dramatically impacting valuations.

Albeit, this assumption changes dramatically during volatile market conditions, caused by any kind of external shock, when liquidity becomes less available and non-uniform. Suddenly, all investors trend towards short-term horizons, reacting rapidly to price movements and incoming information. This shift causes markets to become even less liquid and more inefficient, triggering crashes and crises but still following cyclical and repeatable patterns that can be identified.

Our investment strategies implement a systematic trading process on a diversified basket of major financial instruments, aiming to achieve exceptional risk-adjusted returns. Our strategies consist of two different modules that work in tandem to identify the opportunities with the highest statistical likelihood of success and adjust the positions’ risk parameters to maximize the expected return:

› Pattern Recognition module

Our alpha discovery models consider price as the only input: no fundamental factors are taken into consideration, thereby removing all subjectivity and human bias. Our pattern recognition algorithms scan all the price charts in our instruments’ universe – across three different timeframes – identifying and listing all the recognizable, repeatable patterns.

We also investigate the synchronization of the identified patterns across the three different timeframes, in line with the principles of the Fractal Market Hypothesis, as a common phase on two or more different timeframes improves the statistical likelihood of accurately predicting its outcome. Thereafter, at each point in time on a tick-by-tick basis, every instrument in our universe is uniquely described by its current pattern and its timeframes’ synchronization and a unique identifier called “Synchronization ID” is assigned to it. Finally, our Machine Learning modules are applied to define the optimal trade management rules that the system will follow for each specific “Synchronization ID” when an execution signal is triggered.

Our strategy could be simplistically described as a statistical exercise, in which we identify the pattern formations that offer a statistical edge and place a bet in a standardized matter in order to achieve a positive return. However, the true uniqueness of our process lays in the use of Machine Learning techniques in screening the most profitable “Synchronization IDs”. This allows us to dig deep and identify not only the setups with the most favorable statistical characteristics but also determine their respective optimal trade management parameters to capture the maximum possible expected return.

› Market Sentiment module

The Fractal Market Hypothesis considers market liquidity as the missing link between the mainstream Efficient Market Hypothesis and its blatant failure to predict and explain the rather frequent market bubbles and crashes. As such, in order to protect the overall performance from sudden liquidity-based shocks, our strategies employ a Market Sentiment module to monitor whether the market bias appears one-sided and hence, poses a risk from a price availability standpoint.

We make use of Bloomberg’s API to create a calendar of upcoming events, both financial and otherwise, and every event is ranked according to its expected market volatility. Thereafter. we monitor the public chatter to gauge the prevailing market sentiment tracking all relevant tweets to each event. Our strategies regard the prevailing market sentiment as a contrarian indicator and they apply the relevant risk filters on every opportunity on an affected instrument, either reducing trade size by 50% or restricting trades towards a specific Long or Short direction.

Furthermore, the systematic nature of our research process and the strict risk management rules applied to each opportunity, year in and year out, allow our strategies to remain invested in the market during times of uncertainty and/or increased volatility. Our strategies have delivered consistent results made possible by the three pillars of our risk management modules:

• Religious use of stop-loss: an initially fixed and then progressive trailing stop loss is attached to every opportunity, ensuring that there is no “open risk”. Moreover, each stop level is derived directly from the range and real-time volatility of each instrument, resulting in the appropriate stop loss distance depending on the current market conditions.
• Letting profits run to their maximum extent: this is an essential part of our strategies’ money management rules. Their performance comes not only from a number of small and medium-sized wins but is also amplified to a great extent by capturing a handful of large, “home run” moves.
• Conservative fractional exposure per trade: all the trade opportunities are standardized to include an initially fixed stop loss and take profit threshold calibrated at 1% maximum loss and 1.5% maximum profit compared to the hypothetical portfolio balance.

The standardized stop loss and take profit approach prohibits any strategy relevance to leverage. Therefore, leverage only relates to margin. Regardless of what happens in the world and the markets, our strategies’ profit-to-risk profile always remains the same.

Finally, a key principle of our risk management approach is that, regardless of the strategy’s performance, they never compound their gains. It is our decision to miss out on some extra profits by not compounding gains, but we prefer to exchange them for a smoother equity curve and significantly smaller drawdowns.