At the Crossroads of Trading and Technology
Financial market behavior has taken on renewed vigor in recent years. HFT trading firms has been underperforming as speed is capped by the speed of light, machine learning is challenged with overfitting and the intrinsic characteristic of the financial market makes sequence prediction inefficient. In addition, finance is one of the only fields which is not collaborative. In fact, investors or quants have no incentive in collaborating. Or, at least not until we create one.
As opposed to conventional machine learning methods, at Vestun, we are adopting a much more thoughtful approach using multivariate modelling enabled by neural network. We are not simply systemizing a trading strategy. We create the entire job of an analyst. We extract, store and analyze large amount of data from various sources using neural network, big data storage facilities and distributed computing architecture to exploit market inefficiencies. Beyond this, we are challenging the problem of overfitting by using encrypted data to enable market participants (mainly data scientists) collaborate and create a metamodel more accurate than any individual model operating on its own.
We are not only competing within the 3.5 trillion alternative investment industry but we are also actively covering the digital asset market which is currently among the fastest growing asset class. Unlick conventional machine learning-based quant firms, we do not aim to predict future sequences based on the past or rely on speed as HFT firms does. Our approach focuses on translating market psychology, emotions, reacting to cognitive biases which derives from them and ultimately benefiting from unique investible insights.