Our algorithms detect current market regimes and adjust their behaviour accordingly. They learn from previous trades autonomously and continuously analyse how a gain or a loss could be better predicted going forward.
Our systems are developed to go long or short futures contracts throughout the day. The investment strategy includes a combination of mean reversion and momentum signals. We do not hold any position overnight. As a result, we capture intraday market trends with little correlation to underlying markets.
We are a fully systematic player. No orders are generated manually or in a discretionary way. Our system runs through our proprietary technology launched after three years of careful research and development.
We trade global liquid futures in the US, Europe, and Asia. That includes equity indices, bond indices, and foreign exchange futures.
Hard-coded constraints are implemented into the system and embedded into model optimisation. They include a wide range of non-flexible rules such as max allocation per trading model, per direction, per instruments. All positions are protected by target stop losses. The trading strategy also avoids overnight gaps by design.
We benefit from many more signals than longer term strategies which allows us to have significant statistical measurements. The systems run testing modules to control overfitting and accurately measure strategies’ robustness. These modules use machine learning regularisation and validation techniques as well as more traditional financial best practices, such as walk-forward analysis, data & parameter sensitivity analysis, Monte-Carlo simulation, out-of-sample validation, and underlying logic rationale.
Please contact us to learn more about Walnut.