/-
Copyright 2026 The Formal Conjectures Authors.
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
https://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
-/
module
public import FormalConjecturesUtil
public import FormalConjectures.ErdosProblems.«211»Erdős Problem 1069
References:
[Er87b] Erdős, P., Some combinatorial and metric problems in geometry. Intuitive geometry (Siófok, 1985) (1987), 167-177.
[Sa87] Sah, C.-H., The rich line problem of P. Erdős. (1987), 123-125.
[SzTr83] Szemerédi, E. and Trotter, Jr., W. T., Extremal problems in discrete geometry. Combinatorica (1983), 381-392.
@[expose] public sectionopen EuclideanGeometrynamespace Erdos1069
The $k$-rich lines of a finite set s of points in the plane: the lines of ℝ² which
contain at least $k$ points of s.
def richLines (s : Finset ℝ²) (k : ℕ) : Set (AffineSubspace ℝ ℝ²) :=
{L | IsLine L ∧ k ≤ ((s : Set ℝ²) ∩ (L : Set ℝ²)).ncard}
For $k \geq 2$ there are only finitely many $k$-rich lines, since each is spanned by two of
the points, so Set.ncard counts them.
@[category API, AMS 5 52]
theorem richLines_finite (s : Finset ℝ²) {k : ℕ} (hk : 2 ≤ k) : (richLines s k).Finite := s:Finset ℝ²k:ℕhk:2 ≤ k⊢ (richLines s k).Finite
s:Finset ℝ²k:ℕhk:2 ≤ k⊢ richLines s k ⊆ ↑(Erdos211.determinedLines s)
s:Finset ℝ²k:ℕhk:2 ≤ kL:AffineSubspace ℝ ℝ²hL:IsLine Lhcard:k ≤ (↑s ∩ ↑L).ncard⊢ L ∈ ↑(Erdos211.determinedLines s)
All goals completed! 🐙Given any $n$ points in $\mathbb{R}^2$, the number of $k$-rich lines (lines which contain $\geq k$ of the points) is, provided $k\leq n^{1/2}$, $$\ll \frac{n^2}{k^3}.$$
Erdős [Er87b] describes this as a conjecture of himself, Croft, and Purdy. It was proved by Szemerédi and Trotter [SzTr83]. We take $k \geq 2$, since for $k \leq 1$ there are infinitely many $k$-rich lines.
@[category research solved, AMS 5 52]
theorem erdos_1069 :
∃ C : ℝ, 0 < C ∧ ∀ (n k : ℕ) (s : Finset ℝ²), s.card = n → 2 ≤ k →
(k : ℝ) ≤ Real.sqrt n →
((richLines s k).ncard : ℝ) ≤ C * (n : ℝ) ^ 2 / (k : ℝ) ^ 3 := ⊢ ∃ C, 0 < C ∧ ∀ (n k : ℕ) (s : Finset ℝ²), s.card = n → 2 ≤ k → ↑k ≤ √↑n → ↑(richLines s k).ncard ≤ C * ↑n ^ 2 / ↑k ^ 3
All goals completed! 🐙When $k=n^{1/2}$ the lattice points show that there can be $\geq(2+o(1))n^{1/2}$ many $n^{1/2}$-rich lines: for $n=m^2$ the $m\times m$ grid has $2m$ lines with $m$ points each. Erdős thought that perhaps this is best possible, but Sah [Sa87] gave a construction achieving $\geq (3+o(1))n^{1/2}$ many $n^{1/2}$-rich lines.
@[category textbook, AMS 5 52]
theorem erdos_1069.variants.lattice (m : ℕ) (hm : 2 ≤ m) :
∃ s : Finset ℝ², s.card = m ^ 2 ∧ 2 * m ≤ (richLines s m).ncard := m:ℕhm:2 ≤ m⊢ ∃ s, s.card = m ^ 2 ∧ 2 * m ≤ (richLines s m).ncard
All goals completed! 🐙end Erdos1069