Stephen J. Dubner and Steven D. Levitt’s new book Think Like A Freak (New York: William Morrow, 2014) is out today. Dubner described it in April as a book about creativity and problem-solving. This book may interest the structured analytic techniques (PDF) community I saw at ISA 2014. The Wall Street Journal published a book excerpt here.
I missed the Saturday sale of political science publishers at ISA 2014.
One of the books on my post-conference reading list is Anne L. Clunan‘s book The Construction of Russia’s Resurgence (John Hopkins University Press, 2009). Jack Snyder‘s original paper for RAND in 1977 focused on Soviet strategic culture and the socialisation of politico-military elites during nuclear detente negotiations (PDF). Clunan’s book looks like a useful social construction approach to contemporary issues of leadership and national image that Snyder, Colin S. Gray, Ken Booth and others explored in the first generation of strategic culture scholarship.
Post-ISA 2014, I am delving into formal models and the scientific method. I’m reading Patrick Thaddeus Jackson on scientific models of international relations; writing declarative statements in SWI Prolog; and considering the potential microfoundations of my PhD topic on strategic culture. This is all new territory for theory-testing.
This evening I looked at the first chapter of David Aronson’s book Evidence-Based Technical Analysis (Hoboken, NJ: John Wiley & Sons, 2006). Technical analysis (TA) usually involves pattern recognition (Edwards & Magee); geometric angles and waves (Gann and the Elliott Wave); or institutional money flows (Wyckoff). Aronson suggests the majority of TA approaches involve superstitious, magical thinking. In contrast Karl Popper’s falsifiability provides a way to develop what Jackson would describe as neopositivist TA models.
Aronson’s book tests 6402 trading rules (some significance tests are here). He uses a binary structure to codify each of the trading rules: (1) +1 is a long recommendation; and (2) -1 is a short recommendation (pp. 16-17). For Aronson, “An investment strategy based on a binary long/short rule is always in either a long or short position in the market being traded” (p. 17). This binary structure enables Aronson to combine Boolean logic and Popperian falsifiability in order to test each of the 6402 trading rules. Thresholds (pp. 17-18) mean Aronson can transform the binary trading rules to create If-Then nested loops of declarative rule conditions.
Aronson’s binary structure assumes that traders are trading in the market at all times – just switching between long and short positions. However, prop traders and high-frequency traders may close-out positions – such as at end-of-day to avoid overnight exposure and gap risk. Some TA proponents like Richard D. Wyckoff note that close-out positions can also have strategic uses: first accumulating a position and then profit-taking via selling to trend-followers.
My initial solution was to change Aronson’s binary structure into a trinitarian trade rule. The additional rule / outcome: (3) 0 means the trader is out of the market. This necessitates a sell order closes out any current market positions. This could be done either as a declarative rule condition or as a nested If-Then-Else loop.
One benefit of the scientific method is more rigorous exploration of how such formal models work.