At Focus
Strategy, Methods, Analytic

Trading strategies

Options allow for historically robust, product-specific, alternative risk premia to be monetized in various ways and to generate, through their non-linear pricing, alternative risk and reward profiles. Panathea’s Option strategies are designed and optimized to take advantage of the Option characteristics favorable in each market phase. They differ from conventional Option strategies, which are typically geared towards a specific characteristic or a risk premium.

Panathea’s Futures strategies are driven by a quantitative, directional allocation on different time scales. The time horizons of the various components range from a few minutes to several days. Panathea is constantly developing new strategies and is working to refine existing strategies and methods. All strategies are in proprietary use by Panathea and are tradable on popular institutional platforms.

Quantitative methods of the strategies

The quantitative methods that drive Panathea’s trading strategies range from rule-based models to methods of stochastic optimization to state-of-the-art machine learning techniques.

The objectives of the approaches vary from prognosis-free portfolio optimization – such as the determination of optimal allocation times of Option strategies through stochastic control methods – to the calibration of process parameters through various deep learning methods in the context of prognosis-driven optimization, to guiding directional Futures strategies.

In applying machine learning methods, Panathea is strictly problem-oriented. It is ensured that specific characteristics of a problem can be intuitively represented by the respective architecture. This generally requires a significant adaptation of the models from their form used in other areas of application – Panathea does not use off-the-shelf models.

Analytics

In the analysis of trading strategies, the real revenue drivers of a strategy as well as their robustness as a source of income are derived from the data. Risk behavior is analyzed and a peer-group assignment based solely on data is carried out independently of the proclaimed or advertised peer-classifications. The analysis uses mathematical-statistical methods, which differ from conventional indicators. The analysis often results in segment allocations and ranking lists, which differ significantly from qualitative analysis.