Writing Theory Sections

1 Why This Page Matters

Many weak theory sections are not mathematically false.

They are weak because the reader cannot tell:

  • what is being claimed
  • under which assumptions
  • why the theorem matters
  • how the proof idea connects to the rest of the paper

The job of a theory section is not just to contain correct statements.

Its job is to make the reader trust that the mathematics is carrying a real part of the paper’s story.

2 What A Theory Section Has To Do

A strong theory section usually does four things clearly:

  1. names the question
  2. states the assumptions without hiding them
  3. tells the reader what the main result actually says
  4. explains why that result is relevant to the paper’s broader contribution

If one of those pieces is missing, the section often feels decorative even when the proof is correct.

3 The Load-Bearing Parts

3.1 Problem Setup

The reader should know exactly what object is being studied:

  • estimator
  • algorithm
  • model
  • dynamical system
  • stochastic process

This setup should use the same names and notation that the rest of the paper uses.

3.2 Assumptions

Assumptions should be visible, not buried.

The important question is not only whether the assumptions are standard, but whether the paper helps the reader understand:

  • which assumptions are structural
  • which are technical
  • which are likely to fail in practice

3.3 Main Theorem Statements

A theorem statement should make three things easy to see:

  • the regime
  • the guarantee
  • the dependence on the important variables

If the theorem is long, the paper should still give the reader a short sentence saying what changed relative to prior work.

3.4 Interpretation

The section should not end at “the theorem is proved.”

It should say what the theorem teaches:

  • better dependence on noise?
  • weaker assumptions?
  • sharper sample complexity?
  • stability under perturbation?
  • justification for an empirical pattern?

That interpretation is often the difference between a theorem section and a proof dump.

4 Common Failure Modes

  • theorem statements appear before the reader knows why those objects matter
  • assumptions are technically correct but never interpreted
  • notation changes between introduction, method, and theorem section
  • the main proof idea is never summarized
  • experiments later claim stronger support than the theorem actually gives

5 A Practical Writing Loop

Before polishing prose, force the section through this loop:

  1. write the main theorem in its shortest honest form
  2. list every assumption in plain language
  3. write two sentences of interpretation for each result
  4. check whether the experiments probe anything the theorem actually talks about

If step 4 fails, the issue is often paper design, not theory writing.

6 How This Connects To The Site

Back to top