Modeling as Exploration: The 27‑Method Framework for Deeper Understanding

Sharing some recent thoughts on modeling.

1. Modeling is a process of understanding and exploration

Modeling is often seen as a tool to solve problems, but it is also a way to deepen understanding and explore the subject. The insight gained from modeling itself becomes part of the solution.

I have been developing a thinking framework called 27 Magic Method , which expands understanding of a research object along three dimensions—time, space, and analogy—so as to avoid narrow, case‑by‑case thinking. The process starts with the historical background and future trends (time), then examines all relevant factors from a global perspective (space), and finally seeks interdisciplinary analogies to inspire new ideas.

A comprehensive qualitative analysis often suffices without heavy mathematics, yet it can resolve many issues.

The modeling process itself is iterative, consisting of hypotheses, trials, and feedback. Each adjustment deepens our grasp of the problem’s essence, even if the final “optimal” solution is not immediately reached.

2. Progressive Modeling and the "Beautiful" Model

Instead of striving for an overly precise mathematical model from the start, I now prefer to begin with a simple, computable model and gradually add detail and accuracy. This progressive approach yields quicker initial insights and allows refinement toward a model that best reflects reality.

I also favor "beautiful" models—those that are concise, intuitive, and adaptable across contexts while remaining insightful. Such models capture the core of a problem with minimal assumptions and variables.

A key principle is that a model need not be perfect to be meaningful; even an imperfect model can provide valuable insights and guide us toward the problem’s essence.

3. Interaction Between Data and Models

Previously I wrote that data serve the model, but I now view data and models as equal partners. Models give structure and interpretation to data, while data supply the empirical basis and validation for models.

High‑quality data are essential for good models, and models help uncover patterns in data and direct data collection. Their relationship is a two‑way feedback loop that must be continuously adjusted and optimized.

4. Modeling and Philosophy: From "Dao" to "Technique"

Modeling also involves philosophical layers, which I summarize as "Dao, Fa, Ze, Shu, Qi"—the core philosophy, basic principles, specific rules, technical applications, and tools respectively.

The "Dao" (philosophy) defines the ultimate purpose of modeling: to better understand the world and solve real problems. The "Fa" (principles) guide simplification, hypothesis setting, and summarization. The "Ze" (rules) are discipline‑specific modeling guidelines. The "Shu" (technique) includes classic mathematical models and computational methods. The "Qi" (tools) are the programming languages, software platforms, and databases used.