Why Simple Models Outperform Complex Ones: Lessons from Mathematical Modeling

Mathematical modeling is the process of converting real‑world problems into mathematical form and solving them through models. The author shares insights from Dr Wu Jun’s *The Beauty of Mathematics* (3rd edition), highlighting several experiences.

A correct mathematical model should be simple in form. Ptolemy’s model is classic but includes many complex factors such as celestial positions and trajectories, making it hard to understand and apply. In contrast, a simple model contains only necessary variables and relationships, making it easier to grasp and use while accurately describing the problem’s essence.

An initially imperfect model can be more accurate than a finely crafted wrong one, but if the overall direction is right, we should persist. Modeling is an iterative process; by continuously correcting and improving the model, its accuracy increases. For example, the heliocentric model was less accurate than the geocentric model at first, yet through scientists’ effort and ongoing observation it eventually proved correct.

Large amounts of accurate data are crucial for research and development. When building a mathematical model, we need sufficient accurate and reliable data as a foundation. Such data help validate the model’s effectiveness and accuracy, providing more precise predictions. Therefore, collecting and analyzing abundant data is a key step in mathematical modeling.

A correct model may be disturbed by noise and appear inaccurate; instead of applying a makeshift fix, we should locate the source of the noise, which may lead to major discoveries.

Additional insights from the book include:

…first solve 80% of users’ problems, then gradually address the remaining 20%; this is one secret of industrial success. Many failures stem not from lack of talent but from the wrong approach—pursuing an all‑encompassing solution that cannot be completed.

A good algorithm should be like an AK‑47: simple, efficient, reliable, and easy to understand, not obscure.

Technology consists of “technique” (methods) and “principle” (theory). This book aims to teach the principle, not the technique. Specific search techniques quickly become common and then obsolete; chasing technique alone leads to a hard‑working but unsustainable career.

Many who ask for shortcuts to “technique” overlook that true mastery requires thousands of hours of training and effort. In improving Google Search quality, the author analyzed far more than 1,020 poor results daily; most search engineers cannot achieve this, hoping a single algorithm will solve everything, which is unrealistic.

Truth is always simple in form, never complex or ambiguous.

Choosing simple solutions makes it easier to explain each step’s rationale, facilitating debugging and future improvements.

“Technologies with correct design thinking may still fail due to non‑technical factors; but technologies lacking correct design thinking always fail.

The charm of mathematics lies in simplifying complex problems.

Many mathematical methods seem useless at first but become invaluable over time, which explains why many people devote their lives to studying mathematics.

Regarding education: elementary and middle school students need not spend excessive time reading; their social experience, life skills, and early aspirations shape their lives. Intensive study in middle school can be compressed in university, where comprehension is stronger. Learning is lifelong; over‑studying early can diminish motivation later, and missing developmental stages cannot be compensated.

Information redundancy ensures security. Historically, information and uncertainty reduction are linked; in English, “information” and “intelligence” share the same word. A single bit of information can outweigh thousands of troops in war.

Dr Wu Jun’s *The Beauty of Mathematics* is an engaging and accessible book that offers many inspirations for mathematical modeling and practical work; the author highly recommends it.