🔢 mathematics

An improved bound on the number of dot products determined by a finite point set in the plane

This paper improves the lower bound for the number of distinct dot products determined by a finite set of points in the Euclidean plane to approximately $|P|^{2/3 + 7/1425}$ by extending the work of Hanson, Roche-Newton, and Senger.

Original authors: Michalis Kokkinos

Published 2026-06-15

📖 4 min read🧠 Deep dive

CC BY 4.0

Original authors: Michalis Kokkinos

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine you have a collection of dots scattered on a flat sheet of paper. Now, imagine picking up any two dots, drawing a line from the center of the paper to each of them, and calculating a specific number based on how those two lines relate to each other. In math, this calculation is called a dot product.

The big question this paper asks is: If you have a huge number of dots, how many different dot product numbers can you possibly create?

The Problem: Counting the Unique Numbers

Think of the dots as guests at a party. Every time two guests interact, they produce a unique "handshake number" (the dot product). If you have 1,000 guests, you might think you could get 1,000,000 different handshake numbers. But in reality, many pairs might produce the same number.

Mathematicians want to know the minimum number of unique handshake numbers guaranteed to exist, no matter how you arrange the dots.

The Old Rule: For a long time, the best known rule was that if you have $N$ dots, you are guaranteed at least $N^{2/3}$ unique numbers. (If you have 1,000 dots, that's roughly 100 unique numbers).
The Previous Improvement: A few years ago, researchers managed to squeeze that number slightly higher, adding a tiny bit of extra "growth" to the exponent.
This Paper's Goal: The author, Michalis Kokkinos, wanted to see if he could squeeze that exponent even higher, proving that there are more unique numbers than anyone thought possible.

The Strategy: Organizing the Chaos

To solve this, the author doesn't look at the dots randomly. He organizes them like a military formation.

The "Spokes" Analogy: Imagine the dots are arranged on lines that all radiate out from the center of the paper (like spokes on a wheel).
The Sweet Spot: The author focuses on a specific, tricky scenario where the dots are arranged in the most efficient way possible to hide unique numbers. He assumes there are about $\sqrt[3]{N}$ lines (spokes), and each line holds about $\sqrt[3]{N^2}$ dots. This is the "worst-case scenario" where the math is hardest.
The Intersection Trick: He then looks at where these lines cross a specific vertical line on the paper. This creates a smaller, manageable group of dots that still represents the whole group.

The "Super-Expander" Engine

The core of the proof relies on a mathematical tool called a "superquadratic expander."

The Metaphor: Imagine you have a set of numbers. If you mix them together in a specific way (adding 1, multiplying, etc.), a "normal" set might grow a little bit. A "super-expander" is a set that, when mixed, explodes in size much faster than expected.
The Breakthrough: The author uses a recently discovered, more powerful version of this "expander" (found in a 2024 paper by other mathematicians). This new tool is like upgrading from a bicycle to a rocket ship. It allows him to prove that the set of unique numbers must grow faster than the old rules predicted.

The Calculation: Squeezing the Result

The author uses a series of mathematical "squeezes" (using inequalities named after mathematicians like Ruzsa and Plünnecke).

He takes the "expander" result, which shows massive growth.
He connects this growth back to the number of unique dot products.
He calculates exactly how much the exponent can be pushed up.

The Result

The paper concludes with a new, tighter bound.

Old Bound: $N^{2/3}$
Previous Best: $N^{2/3 + \text{tiny bit}}$
This Paper's Bound: $N^{2/3 + \frac{7}{1425}}$

While $\frac{7}{1425}$ looks like a small number, in the world of high-level mathematics, this is a significant victory. It proves that no matter how cleverly you arrange your dots, you cannot hide as many duplicate numbers as you thought. There are simply more unique "handshake numbers" than previously believed.

Summary

In simple terms, this paper is a mathematical audit. It took a known rule about counting unique numbers generated by points on a plane and used a newer, more powerful calculator (the super-expander) to prove that the minimum count is slightly higher than we thought. It doesn't change how we build bridges or treat diseases; it simply refines our understanding of the fundamental geometry of numbers.

Technical Summary: An Improved Bound on the Number of Dot Products Determined by a Finite Point Set in the Plane

Problem Statement
The paper investigates the lower bound for the number of distinct dot products determined by a finite set of points $P$ in the Euclidean plane $\mathbb{R}^2$ . Let $\Lambda(P) = \{p \cdot q : p, q \in P\}$ denote the set of distinct dot products. The central question is to determine the asymptotic growth of $|\Lambda(P)|$ in terms of $|P|$ .

This problem is motivated by the Erdős distinct distances problem, which was resolved by Guth and Katz [1] with a lower bound of $N/\log N$ . While similar bounds are conjectured for dot products, the best known general lower bound prior to this work was $|P|^{2/3}$ , derived from a simple application of the Szemerédi-Trotter Theorem. Previous work by Hanson, Roche-Newton, and Senger [3] improved this exponent to $2/3 + 1/3057$ . This paper aims to further refine the constant in the exponent.

Methodology and Framework
The proof strategy relies on a combination of incidence geometry, additive combinatorics, and specific structural assumptions about the point set $P$ . The methodology proceeds through the following logical steps:

Structural Reduction via Dyadic Decomposition:
The authors first restrict attention to a "worst-case" configuration where the number of distinct dot products is minimized. Using Lemma 4.2 (a pinned dot product result derived from Szemerédi-Trotter), they assume the number of lines through the origin covering $P$ , denoted $|L|$ , is approximately $|P|^{1/3}$ . If $|L|$ were larger, the bound would follow trivially.
Through dyadic decomposition and the Pigeonhole Principle, the authors extract a large subset $P' \subseteq P$ and a subset of lines $L' \subseteq L$ such that each line in $L'$ contains approximately $|P|^{2/3}$ points.
Fixing a Coordinate and Slopes:
To facilitate the analysis, the authors further restrict the set to $L'' \subseteq L'$ containing lines passing through points with a fixed $x$ -coordinate $x_0$ . This creates a framework where $L''$ contains roughly $|P|^{1/3}$ lines, and each line contains roughly $|P|^{2/3}$ points. Let $S$ be the set of slopes of these lines.
Exploiting Product Set Growth:
The core of the argument involves analyzing the growth of product sets derived from the slopes $S$ . The authors utilize a "superquadratic expander" result (Theorem 4.1), an improvement over the techniques in [3] which incorporates the Elekes-Szabó Theorem and computer algebra. This theorem guarantees the existence of two slopes $s, s' \in S$ such that a specific ratio of product sets involving shifts of $S$ grows as $|S|^{31/12}$ .
Intersection and Inequality Application:
The proof defines an intersection set $Z$ of scaled $x$ -coordinates from points lying on lines with slopes $s$ and $s'$ . The authors relate the size of the dot product set $\Lambda(P)$ to the sizes of the sets $Z(sS+1)$ and $Z(s'S+1)$ .
By applying the Ruzsa Triangle inequality and the Plünnecke-Ruzsa inequality (Lemmas 4.3 and 4.4), the authors bound the size of the expander ratio in terms of $|\Lambda(P)|$ and $|Z|$ . Crucially, Lemma 4.5 is used to establish a lower bound on the size of the intersection $Z$ based on the multiplicative energy of the coordinate sets.

Key Contributions and Results
The primary contribution of the paper is the derivation of a new lower bound for the number of distinct dot products.

Main Theorem (Theorem 1.1): For any finite set $P \subseteq \mathbb{R}^2$ , the number of distinct dot products satisfies:
$|\Lambda(P)| \gtrsim |P|^{2/3 + \frac{7}{1425}}$
where the notation $\gtrsim$ indicates a lower bound up to a polylogarithmic factor.
Improvement over Previous Work: This result improves the exponent constant from the previous $1/3057$ (achieved in [3]) to $7/1425$ . The improvement is achieved by utilizing a stronger "superquadratic expander" bound (Theorem 4.1) and optimizing the application of Ruzsa-type inequalities to minimize the loss in the exponent during the derivation.

Significance and Claims
The paper claims that the improvement in the exponent is significant because it demonstrates that the threshold of $2/3$ can be pushed further in the real Euclidean setting. The authors note that while finite field and complex plane variants of this problem remain at the $2/3$ threshold, the real setting allows for these refinements due to specific structural properties (such as the superquadratic expansion of certain product sets).

The work does not claim to solve the problem completely (i.e., reaching the conjectured $|P|^{1-\epsilon}$ bound) but rather represents a quantitative refinement of the current best-known lower bound. The authors emphasize that the argument relies heavily on the specific geometry of the real plane and the availability of the Elekes-Szabó Theorem, which distinguishes this approach from methods applicable in finite fields or complex settings.