Since the advent of electronic trading the ways firms evaluate, integrate, and interpret information has radically changed. Investors may have always relied on new housing data, employment reports and federal reserve guidance but the voracity and depth of information processing has seen incredible development. Global markets mean that there are numerous new indicators coming out on a round the clock cycle, and the sheer volume of information can be overwhelming. How does one develop, validate, and apply a statistical model when by the time the modeling process has been completed new information has arrived that has rendered the analysis outdated?
Rather than rely on purely qualitative or purely quantitative methods, most analysts employ both to some degree. Traditional statistical techniques such as Bayesian methods to incorporate these ideas are often slow, sensitive, and difficult to estimate. To make decisions, one must combine knowledge of statistics with knowledge about the nature of a given market to make instantaneous, real-time decisions.