CrowdCent is on a mission to enable decision-making pipelines with humans at the edge and automation at the center. Any data or model generating community with sufficient size and longevity is inherently valuable; in the long-run, we believe a community’s userbase should be compensated for its data/models commensurate to the value that can be extracted from those data/models.
The decision making process and organizational design of the discretionary hedge fund are under attack. A large portion of the industry relies on broken, pre-internet strategies that are too slow—most mispricings are quickly identified and corrected by the wide swath of factor mining algorithms that can make quicker and less biased decisions than can a portfolio manager. As a result, “quant” funds have outperformed discretionary funds in recent years. Numerous discretionary funds are throwing money at big data and machine learning in an effort to stay competitive, but this pursuit creates a deluge of new information that will serve more to inundate the portfolio manager than to inform investment decisions.
To solve the problem, we can use a portfolio manager’s opinion not as the ultimatum, but as a feature in a broader dataset. To replicate this situation, we will turn to exclusive online investment forums where users post lengthy “pitches” to buy or short a stock. Through the combination of traditional valuation techniques, natural language processing, and the wisdom of the crowd, we can incorporate fundamental investment data and opinion into a systematic investment decision pipeline. In this pipeline, an NLP algorithm combs through community-submitted investment pitches, extracts topics, and evaluates each user’s write-up for sentiment, objectivity, and key-word weightings. However, in the face of new facts, positions need to be re-evaluated. Constant calibration comes in the form of community feedback and discussion. Each time a user comments or rates an investment idea, the algorithm incorporates the feedback into the decision making process, re-evaluating the probability of success for a given investment idea.
This solution should effectively operationalize a portfolio manger’s opinion, removing any cognitive biases he or she might have while also providing a clear picture of which information is important to the investment decision making process. If a portfolio manager makes the final decision, we lack this key information of seeing why a certain decision may have been made. In crowdsourcing investment recommendations, we pull from a larger net of ideas and ensure a larger diversity of thought. With all of these slight tweaks to the typical discretionary investment decision, we should be able to amplify returns, on average.