Crowdsourced Investment Research Through Tournaments

Abstract

Traditionally, the development of investment strategies has required domain-specific knowledge and access to restricted datasets. Hence, the majority of data scientists do not research investment opportunities because they lack one or both of these requirements. In this article, the authors discuss the merits of tournaments as a crowdsourcing paradigm for investment research. Tournaments can overcome these research barriers, hence enabling the wide population of data scientists to contribute to the development of investment strategies.

TOPICS: Simulations, big data/machine learning

Key Findings

• • There are several barriers to the development of investment strategies: domain-specific knowledge barrier, budgetary constraints and confidentiality restrictions, inability to monetize the value of data, and backtest overfitting.

• • Tournaments offer a solution for overcoming the barriers associated with developing investment strategies.

• • The modern investment process the authors suggest allows data scientists without an investment background to contribute forecasts to a systematic asset manager.

• • This does not mean that tournaments obviate the need for financial expertise. On the contrary, deep financial expertise is required to cast the problem crowdsourced through the tournament.