Computational Finance: Beyond the "Invisible Hand" to a Science of Market Dynamics

TL;DR

The global economy remains fragile because our financial models are archaic. This paper argues that finance must be treated as a rigorous empirical science. By leveraging tick-by-tick data, intrinsic time, and scaling laws, we can move beyond the failed Efficient Market Hypothesis toward a computational framework that anticipates crises and stabilizes markets through predictive services.

Background: The Infrastructure of Failure

Why do experts fail to predict financial meltdowns? The author, Richard Olsen, argues that the root cause is a lack of R&D. While industries like pharma or engineering reinvest 10% of revenue into research, banks treat finance as a process to be automated rather than a system to be understood. We are currently flying a "global economy" jet using 18th-century maps—specifically, Adam Smith’s theory of the "invisible hand."

The Core Insight: Markets are Heterogeneous

The fatal flaw of classical economics is the assumption of homogeneity. In reality:

• Time Horizons vary: Some traders operate in seconds; others in months.
• Secondary Reactions: Information isn't absorbed instantly; it triggers cascades of reactions that pile up like "freak waves" in the ocean.
• Overshooting: Because there is no fixed reference point for value, markets naturally swing far beyond equilibrium.

Methodology: The Tools of High-Frequency Finance

To fix this, the paper proposes three radical technical shifts:

1. Intrinsic Time vs. Clock Time

Linear clock time is an "artificial" construct in finance. In active periods, many events happen in a second; in quiet periods, nothing happens for minutes. Intrinsic time (or transaction time) contracts and expands the timeline based on the flow of events, revealing the natural dynamics of the marketplace that regular intervals hide.

2. The 18 Scaling Laws

The author identifies a set of "Power Laws" that remain invariant across scales. These laws provide a mathematical backbone to predict how far a price will move and how large an "overshot" will be relative to a given timeframe.

Figure 1: High-frequency FX data acts as a "particle accelerator" for financial physics.

A Research Agenda for the Future

The author calls for a "Wikipedia-like" approach to financial data:

• Open-Source Data Repository: A global, filtered, tick-by-tick database for all instruments.
• Open-Source Modeling (Olsen Routes): An event-based programming language to model complex agent interactions.
• Predictive Weather Maps: Using computational models to trace global capital flows in real-time.

Critical Analysis & Conclusion

Takeaway

The shift from "Efficient Markets" to "Computational Dynamics" isn't just academic; it's a prerequisite for global stability. By understanding the footprints of capital through tick data, we can create trading strategies that "do right and do well"—profiting from overshoots while dampening their destructive impact.

Limitations

While the vision is compelling, the paper acknowledges the "colossal task" of data collection and filtering. Skeptics might argue that the "black box" nature of proprietary institutional algorithms makes a truly comprehensive "weather map" difficult to achieve without regulatory mandates for transparency.

Future Outlook

As we move toward 2026 and beyond, the integration of Agent-Based Modeling and high-frequency scaling laws will likely be the battlefield where the next generation of SOTA financial AI is built.