If you were concerned you had brain damage, what would you trust for a diagnosis: an experienced clinician or a blind rule based on statistics? The temptations to favor a seasoned expert are obvious, but you’d be wise to resist. When given access to the same patient testing data, simple, rule-based systems derived from statistical relationships are far better at diagnosing mental impairment than seasoned clinicians.
This outperformance isn’t limited to clinical diagnoses. Rule-based decision making is relied upon throughout medicine and the hard sciences, and finance has followed suit. Building their trading models using dozens—if not hundreds or thousands—of statistical relationships, quants are the rule-based purists of the financial world. While an entirely rule-based decision-making system may seem ideal, there is a limit to what can be systematized with current technology. It is here that the quantamental enters the picture. Moving beyond the limitations of strictly rule-based systems, quantamentals embrace superior machine-human hybrid strategies that may be the future of finance.
Quantamentals and the Limits of Systemization
Simply put, a statistical decision making process is a formal mapping between information inputs (e.g. Apple beats earnings and revenue estimates) and decisions (e.g. buy Apple). The job of a quant, roughly speaking, is to assemble a trading model based on a carefully calibrated library of such input-to-decision rules.
Creating trading rules based on information-dense areas—company financials, for example—isn’t all that difficult: the input data is clearly defined and frequently measured, and the relationship between company performance and stock price is straightforward. Trading rules of this nature are low risk…but also low reward. Therefore, to build alpha-generating systems, more than standard rules are required. Quants are aware of these limitations and have tried to create rules for information outside of the straightforward, but the process of systemization in this realm becomes increasingly difficult and often prohibitively expensive. In short, there is currently a threshold to what can be realistically systematized.
The quantamental approach is a potentially powerful answer to the quants’ systemization dilemma. Striking a balance between human and machine systems, quantamentals rely on both rule-based decision making and expert human judgment to execute trades. By leveraging the objectivity and accuracy of machine systems, the rule-based component of the quantamental approach provides a reliable, formal framework for trading decisions. These systems, as we discussed earlier, have undeniable advantages over human judgement in certain scenarios. But where such formalization is difficult or impossible, quantamentals have recourse to human judgement, allowing them to adapt to situations and act on information that is beyond the capacity of systemization.
Interestingly, this “cyborg” approach is considered to be the pinnacle of chess, as hybrid, human-machine teams can outmatch any opponent—human or machine. And what is true for chess is also true for finance: human and machine decision-making systems each have unique limitations but are stronger together than either is individually. Doomsday robot takeovers may make great headlines, but the reality is far more nuanced. Technology is simply not ready to take humans out of the investment process.
The real revolution taking finance by storm isn’t robots. It’s cyborgs. It isn’t quants. It’s quantamentals.
Prattle automates investment research by quantifying market-moving language. Prattle provides central bank and equities analytics that give clarity to investors struggling with a flood of unstructured data and information. Prattle was founded by experts in textual analytics and economic forecasting and is backed by top-tier Wall Street and Silicon Valley investors including GCM Grosvenor and NEA. Prattle produces its data using Portend, a proprietary data science software platform. For more information, visit www.prattle.co.