Three-ends Furstenberg estimates and approximate rectangles
Listen
Radio episode about this paper
Transcript
Introduction to the show: ident: Mathematics. The day in mathematics: new theorems, new proofs, and what they open.
Ada: We're looking at the paper: "Three-ends Furstenberg estimates and approximate rectangles".
Emmy: Here's the gist of the paper. We prove a three-ends incidence estimate whose multiplicity exponent is sharp for one-dimensional Katz–Tao point sets The paper develops techniques based on rectangular nonconcentration and its incidence theory to obtain strong quantitative bounds on the number of approximate rectangles with vertices belonging to four separated camps, each retaining a non-concentration condition.
Ada: Incidence Estimates for Tubes The authors introduce a notion of three-ends incidence, defined as a collection (P, T) being (M, c)-three-ends if for every p ∈ P, there are subfamilies T1(p), T2(p), T3(p) ⊂ the tubes incident to p such that Ti(p) ≥ M for all i = 1, 2, 3, and ∠(T, T′) ≥ c if T ∈ Ti(p), T′ ∈ Tj (p), i ≠ j They establish a sharp lower bound on the number of tubes in a (δ, s, t, δ−ϵ)-Furstenberg set as T ≳ η δ / ηMδ−f(s,t) Furthermore, they define conditions like (P is a (δ, s, C)-KT set) which are used in incidence theorems such as Theorem 2.3 and Theorem 2.4.
Emmy: Bounding Approximate Rectangles The paper derives results concerning the number of approximate rectangles formed from four separated sets S1, S2, S3, S4 in
0, 1: 2.
Ada: Theorem 1.5 states that R ≲ η δ − O(ϵ) δ − 7/3 Proposition 4.2 shows that the collection P is a (δ, s, C)-RKT set where C ≲ δ − O(ϵ) s Σ P.
Paper summary: Ada: So wrapping up this discussion on "Three-ends Furstenberg estimates and approximate rectangles," Demeter and O’Regan have really given us a precise tool for controlling incidence in these complex point configurations.
Emmy: They've established that the three-ends property, which involves those three families of tubes around each point, is powerful enough to generate a strong quantitative bound on the number of approximate rectangles formed by four separated sets.
Kolya: It seems like this work bridges the gap between abstract geometric intuition about Furstenberg sets and concrete counting problems involving rectangles in low-dimensional spaces.
Ada: Exactly. The sharpness of the exponent five over four for M in Theorem one point three shows that their method for finding these estimates is quite robust, and it’s not just an artifact of the specific construction they used to show it's sharp <ref:2611.17293#pg3>.
Emmy: And they’ve shown that when s is less than or equal to one, the energy formulation outperforms other linear energy results, which gives us a better handle on the structure when dealing with these lower density points.
Kolya: The implications are that we have a very concrete way to measure how much geometric structure—like non-concentration—translates into bounds on combinatorial objects like approximate rectangles.
Ada: It’s a strong confirmation that the three-ends hypothesis is a useful tool for manufacturing stronger non-concentration assumptions in this area of mathematics.
Emmy: Ultimately, the paper provides a pathway to get those sharp estimates we were looking for by using incidence theory in a very specific way.
Conclusion: Ada: So, we're wrapping up on these three ends Furstenberg estimates and approximate rectangles paper by Demeter and O’Regan.
Emmy: It really boils down to using those three families of tubes around each point to get a much tighter control over how many approximate rectangles you can form from four separated sets.
Kolya: The authors prove this using techniques based on rectangular nonconcentration, which gives them these strong quantitative bounds on the number of these rectangles.
Ada: It's about showing that those three-ends conditions create a stronger version of nonconcentration, which is what lets them get an exponent five over four for M in their main theorem.
Emmy: And the sharpness of that exponent confirms they didn't just pick a convenient number; the construction they used really forces that bound.
Kolya: From my side, what this means is we have a better handle on these kinds of point configurations when dealing with lower density points, especially when s is less than or equal to one.
Ada: Exactly. It suggests that the three-ends hypothesis isn't just theoretical; it’s a concrete tool for proving things about geometric structures in these sets.
Emmy: And it helps us understand how much structure you can expect to see from those point sets when you look at their energy estimates, like the E4 delta P ones.
Kolya: It opens up a way to verify some of those conjectures about the fourth energy, which is pretty cool because it connects these incidence bounds to other hard problems.
Ada: Yeah, it's a lot of interconnected ideas here. We’ll look at how this new control over rectangles fits into the bigger picture of additive combinatorics next.
CIPRIAN DEMETER, WILLIAM O’REGAN
Swedish Research Council · Institut Mittag-Leffler
math.CO, math.CA
Submitted: 2026-11-12
Updated: 2026-11-12
License: http://creativecommons.org/licenses/by/4.0/
Importance score: 100/100
The gist: We prove a three-ends incidence estimate whose multiplicity exponent is sharp for one-dimensional Katz–Tao point sets The paper develops techniques based on rectangular nonconcentration and its…
Key concepts
- Three-ends Incidence
- This concept defines a collection of points and tubes where for every point, there are three subfamilies of incident tubes, each having at least M elements. Crucially, the angles between any tube from one subfamily and any tube from another must be at least c.
- Approximate Rectangles
- These are geometric structures formed by four separated sets in a 2D space. The paper counts how many such rectangles exist within these sets, providing bounds on their number based on the non-concentration properties of the underlying point sets.
- Rectangular Nonconcentration
- This is a technique used to prove strong quantitative bounds. It involves analyzing how points are distributed across various scales and orientations, ensuring that no small local region concentrates too many points, which leads to sharp estimates on incidence counts.
- (δ, s, C)-RKT Set
- This is a specific condition applied to the point set P. It means the set satisfies certain density and non-concentration requirements related to parameters δ and s. Proving this property allows the authors to apply powerful incidence theorems for bounding geometric structures.
Terminology
Summary
We prove a three-ends incidence estimate whose multiplicity exponent is sharp for one-dimensional Katz–Tao point sets The paper develops techniques based on rectangular nonconcentration and its incidence theory to obtain strong quantitative bounds on the number of approximate rectangles with vertices belonging to four separated camps, each retaining a non-concentration condition. p. 0 At the heart of the argument lies rectangular nonconcentration and its incidence theory introduced earlier by the authors. As an application,
we obtain a strong quantitative bound on the number of approximate rectangles with the
vertices belonging to four separated camps, each retaining a non-concentration condition.Where the paper says it
Incidence Estimates for Tubes
The authors introduce a notion of three-ends incidence, defined as a collection (P, T) being (M, c)-three-ends if for every p ∈ P, there are subfamilies T1(p), T2(p), T3(p) ⊂ the tubes incident to p such that Ti(p) ≥ M for all i = 1, 2, 3, and ∠(T, T′) ≥ c if T ∈ Ti(p), T′ ∈ Tj (p), i ≠ j They establish a sharp lower bound on the number of tubes in a (δ, s, t, δ−ϵ)-Furstenberg set as T ≳ η δ / ηMδ−f(s,t) Furthermore, they define conditions like (P is a (δ, s, C)-KT set) which are used in incidence theorems such as Theorem 2.3 and Theorem 2. p. 0 Then
|T | ≳η δ
ηMδ−f(s,t)
,
where
f(s, t) = min
s + t
2
, t, 1
. Some further research directions have centred upon weakening some of the assumptions in
the statement of Theorem 1.Where the paper says it
Bounding Approximate Rectangles
The paper derives results concerning the number of approximate rectangles formed from four separated sets S1, S2, S3, S4 in [0, 1]2. Theorem 1.5 states that R ≲ η δ − O(ϵ) δ − 7/3 Proposition 4. p. 1 2 CIPRIAN DEMETER AND WILLIAM O’REGAN
see Proposition 2. This bound may be seen to be sharp if one considers a collection P
maximally clustered, that is, being contained in a ball of radius ∼ δ|P|
1/2
. Consider a collection T of δ-tubes and a collection P of δ-points in [0, 1]2
. Assume that P
is a (δ, s, C)-RKT set.Where the paper says it
Where the paper says it
Sharp Exponents and Constructions
The sharpness of the exponent 5/4 for M in Theorem 1.3 is demonstrated by a multiscale construction which yields a bound P ∼ δ − 2/3 for M The analysis shows that the exponent 5/4 cannot be improved without adding new hypotheses Similarly, the sharp exponent for Conjecture 6. p. 2 Indeed,
δ
−1/2M5/4 = δ
−1/2
(δ
−2/3
)
5/4 = δ
−4/3
,
which is exactly the size of each tube family in (1. Thus the exponent 5/4 cannot be
improved in general without adding new hypotheses. 3 to the
attention of WOR. CD thanks Shukun Wu for pointing out to him the train-track example
relevant to two-ends.Where the paper says it
Where the paper says it
Energy Estimates and Decoupling
The paper explores energy estimates on the paraboloid P2 = (x, y, x2 + y2). Theorem 1.7 provides an equivalent formulation for the approximate rectangle count E4,δ(P1, P2, P3, P4) as E4,δ(P1, P2, P3, P4) ≲ η δ − O(ϵ) δ − s (10−3σ) / 4−σ. p. 4 Energy estimates on the paraboloid. The δ-approximate rectangle count admits
an equivalent formulation. 6) was found by the authors. A proof of Proposition 1. At the heart of the argument lies rectangular nonconcentration and its incidence theory introduced earlier by the authors. As an application,
we obtain a strong quantitative bound on the number of approximate rectangles with the
vertices belonging to four separated camps, each retaining a non-concentration condition.Where the paper says it
Where the paper says it
Where the paper says it
Projection Theorems and Final Bounds
The paper presents projection theorems, such as Proposition 1.9, which provides a lower bound on the number of points in the projection Nδ(πx(P)) ≳ η δ − O(ϵ) C − 3/4 M1 + σ / 4 δ s / 4 P. p. 0 All distances between line directions are measured modulo π. The following sharp lower bound up to exponent for the number of tubes was obtained by
[RW23] building on the work of [OS26b]. The manuscript was written by humans although we have used AI in
an assistive role, which did include writing some short and prescriptive parts. For example,
all the text in the next section preceding the statement of Proposition 2. THREE-ENDS FURSTENBERG ESTIMATES AND APPROXIMATE
RECTANGLES
CIPRIAN DEMETER AND WILLIAM O’REGAN
Abstract. We prove a three-ends incidence estimate whose multiplicity exponent is sharp
for one-dimensional Katz–Tao point sets.Where the paper says it
Where the paper says it
Where the paper says it
Conjectures and Examples
The paper discusses two conjectures regarding the fourth energy E4,δ(P), Conjecture 6.1 concerning separated fourth energy, which suggests an exponent of 2s for s ≤ 1 and a bound of δ − a(s)−η where a(s) = max[2s, 3s − 1]. p. 4 5) for the separated energy outperforms the optimal exponent for
the linear energy. Also, the sharpness of the upper bound (1. 3 to the
attention of WOR. CD thanks Shukun Wu for pointing out to him the train-track example
relevant to two-ends.Where the paper says it
Where the paper says it
Decoupling Bound Verification
Proposition 6.3 verifies Conjecture 6.2 when s ≤ 1 by showing that the train-track lower bound matches the result of Proposition 6. p. 7 The lower bound is immediate; the upper bound follows since for any pair (T, S) with the
conditions stipulated above, the admissible p will live in a ball of radius ∼ δ/c, which may
trivially contain ≲ c
−2 points from P. The result follows immediately. We call P uniform at prescribed dyadic scales if, at each scale ρ, the quantities |P ∩ Q|
are comparable independently of Q ∈ Dρ(P). We use scales separated by a fixed factor 2−T
,
apart from a possibly shorter terminal block, with T = T(ϵ) sufficiently large. A qualitative version of Theorem 1. 8, assuming additionally |Θ| ≥ δ
−τ
for a fixed τ > 0
and taking ϵ sufficiently small in terms of s, τ , yields a gain δ
−η
for some η = η(s, τ ) > 0.Where the paper says it
Where the paper says it
Where the paper says it
References
[BD15] Jean Bourgain and Ciprian Demeter. The proof of the l2 decoupling conjecture. p. 5 7], see also [OS26a, Theorem 1. Their proof strategy
has some similarities to ours.Where the paper says it
[DO25] Ciprian Demeter and William O’Regan. Incidence estimates for quasi-product sets and applications. p. 0 THREE-ENDS FURSTENBERG ESTIMATES AND APPROXIMATE
RECTANGLES
CIPRIAN DEMETER AND WILLIAM O’REGAN
Abstract. We prove a three-ends incidence estimate whose multiplicity exponent is sharp
for one-dimensional Katz–Tao point sets. WOR is supported in part by an NSERC Alliance
grant administered by Pablo Shmerkin and Joshua Zahl. 1
arXiv:2611.Where the paper says it
Where the paper says it
[DW25] Ciprian Demeter and Hong Wang. Szemerédi–Trotter bounds for tubes and applications. p. 3 4 with P
′ gives the desired result. Let us now show some applications, that have motivated our study of the three-ends
phenomenon. Provided this is
not the case, in some weakly quantitative way, then one may improve upon (1. This was
established in [OS26a] and we invite the reader to view their well-written and clear paper.Where the paper says it
Where the paper says it
[FR24] Yuqiu Fu and Kevin Ren. Incidence estimates for α-dimensional tubes and β-dimensional balls in R2, J. p. 0 THREE-ENDS FURSTENBERG ESTIMATES AND APPROXIMATE
RECTANGLES
CIPRIAN DEMETER AND WILLIAM O’REGAN
Abstract. We prove a three-ends incidence estimate whose multiplicity exponent is sharp
for one-dimensional Katz–Tao point sets.Where the paper says it
[Orp23] Tuomas Orponen. Additive properties of fractal sets on the parabola. p. 4 Getting the optimal exponent
2 would be a significant target. The interest in the energy formulation goes back to a closely related δ-discretized energy
problem on the parabola
P
1 = {(t, t2
) : |t| ≤ 1}.Where the paper says it
[OS26a] Tuomas Orponen and Pablo Shmerkin. Furstenberg-type estimates under mild non-concentration assumptions. p. 4 THREE-ENDS FURSTENBERG ESTIMATES AND APPROXIMATE RECTANGLES 5
in a long thin rectangle inside such a tube. Thus the non-concentration can only be exploited
in the forms permitted by available Furstenberg-type incidence theorems.Where the paper says it
[OS26b] Tuomas Orponen and Pablo Shmerkin. Projections, Furstenberg sets, and the ABC sum-product problem. p. 0 THREE-ENDS FURSTENBERG ESTIMATES AND APPROXIMATE
RECTANGLES
CIPRIAN DEMETER AND WILLIAM O’REGAN
Abstract. We prove a three-ends incidence estimate whose multiplicity exponent is sharp
for one-dimensional Katz–Tao point sets.Where the paper says it
[OWY26] William O’Regan, Shukun Wu, and Guangzeng Yi. Two-ends Furstenberg inequality for transversal families and applications to Fourier decay. p. 2 THREE-ENDS FURSTENBERG ESTIMATES AND APPROXIMATE RECTANGLES 3
more or less along the long directions of the horizontal rectangles (with the small angular
variation permitted by their eccentricity), the blue lines run along the long direction of the
larger vertical block, and the dark lines go through one of the clusters. The three direction
families are quantitatively transverse. WOR is supported in part by an NSERC Alliance
grant administered by Pablo Shmerkin and Joshua Zahl. 1
arXiv:2611.Where the paper says it
Where the paper says it
[OY25] Tuomas Orponen and Guangzeng Yi. Large cliques in extremal incidence configurations. p. 0 A (δ, s, t, C)-Furstenberg set is a collection T of δ-tubes in the plane, where there
exists a (δ, t, C)-set of points P in the plane so that through each point p ∈ P there exists a
(δ, s, C)-set T (p) ⊂ T of tubes incident to p. For planar point configurations we work in a fixed bounded region, normalised to [0, 1]2
.Where the paper says it
[PS92] J´anos Pach and Micha Sharir. Repeated angles in the plane and related problems, J. p. 0 Introduction
Much of the work in approximate incidence geometry has centred on the study of Furstenberg sets. A (δ, s, t, C)-Furstenberg set is a collection T of δ-tubes in the plane, where there
exists a (δ, t, C)-set of points P in the plane so that through each point p ∈ P there exists a
(δ, s, C)-set T (p) ⊂ T of tubes incident to p. We prove a three-ends incidence estimate whose multiplicity exponent is sharp
for one-dimensional Katz–Tao point sets. At the heart of the argument lies rectangular nonconcentration and its incidence theory introduced earlier by the authors.Where the paper says it
Where the paper says it
[RW23] Kevin Ren and Hong Wang. Furstenberg sets estimate in the plane. p. 0 Introduction
Much of the work in approximate incidence geometry has centred on the study of Furstenberg sets. A (δ, s, t, C)-Furstenberg set is a collection T of δ-tubes in the plane, where there
exists a (δ, t, C)-set of points P in the plane so that through each point p ∈ P there exists a
(δ, s, C)-set T (p) ⊂ T of tubes incident to p. WOR is supported in part by an NSERC Alliance
grant administered by Pablo Shmerkin and Joshua Zahl. 1
arXiv:2611.Where the paper says it
Where the paper says it
[Shm23] Pablo Shmerkin. A non-linear version of Bourgain’s projection theorem, J. p. 6 THREE-ENDS FURSTENBERG ESTIMATES AND APPROXIMATE RECTANGLES 7
Acknowledgements. We are indebted to Kevin Ren and Pablo Shmerkin for sharing a
proof of a non-quantitative version of Theorem 1.Where the paper says it
[Tao11] Terence Tao. An Epsilon of Room, II: Pages from Year Three of a Mathematical Blog, American Mathematical Society, Providence, RI, 2011
[WW24] Hong Wang and Shukun Wu. Restriction estimates using decoupling theorems and two-ends Furstenberg inequalities. p. 4 THREE-ENDS FURSTENBERG ESTIMATES AND APPROXIMATE RECTANGLES 5
in a long thin rectangle inside such a tube. Thus the non-concentration can only be exploited
in the forms permitted by available Furstenberg-type incidence theorems. WOR is supported in part by an NSERC Alliance
grant administered by Pablo Shmerkin and Joshua Zahl. 1
arXiv:2611.Where the paper says it
Where the paper says it
[WW26] Hong Wang and Shukun Wu. Two-ends Furstenberg estimates in the plane, Math. p. 0 THREE-ENDS FURSTENBERG ESTIMATES AND APPROXIMATE
RECTANGLES
CIPRIAN DEMETER AND WILLIAM O’REGAN
Abstract. We prove a three-ends incidence estimate whose multiplicity exponent is sharp
for one-dimensional Katz–Tao point sets. Provided this is
not the case, in some weakly quantitative way, then one may improve upon (1. This was
established in [OS26a] and we invite the reader to view their well-written and clear paper.Where the paper says it
Where the paper says it
[WZ26] Hong Wang and Joshua Zahl. Sticky Kakeya sets and the sticky Kakeya conjecture, J. p. 0 THREE-ENDS FURSTENBERG ESTIMATES AND APPROXIMATE
RECTANGLES
CIPRIAN DEMETER AND WILLIAM O’REGAN
Abstract. We prove a three-ends incidence estimate whose multiplicity exponent is sharp
for one-dimensional Katz–Tao point sets.Where the paper says it
Pruning redundant reasoning tokens
The argument exploits bilinear transversality through the three-ends hypothesis to produce a quantified rectangular Katz–Tao condition This permits bilinear incidence counting inside rectangles of arbitrary orientation and eccentricity The authors use Lemma 3.2 to control the products of local tube counts at every coarse scale, showing that T1 (Q)T2 (Q) ≤ δ − O(ϵ) T2 Nρ(P1) This leads to the bound P1 ≥ 2 − m P
How the training signal is built
The proof of Proposition 4.4 involves a complex counting process where rectangles are encoded as hyperedges H, and a selection of a 'good' corner is performed to bound R The final application of Proposition 4.1 uses the lower bounds on selected tubes and circles to conclude the estimate R ≲ δ − O(ϵ) N2 M3 j T j T j − 1
Where the gains come from
The gains stem from exploiting bilinear transversality through the three-ends hypothesis to manufacture a stronger form of non-concentration This allows for a quantified rectangular Katz–Tao condition, which is then applied to incidence theorems with explicit dependence on the RKT constant The use of three families of tubes instead of two leads to a gain in the exponent 5/4 for M and improves the bound compared to two-ends results. The analysis also shows that for s ≤ 1, the energy formulation outperforms the optimal exponent for linear energy. p. 4 We will see that when
s ≤ 1, our exponent (1. 5) for the separated energy outperforms the optimal exponent for
the linear energy.Where the paper says it
The decoupling bound verifies Conjecture 6.2 when s ≤ 1 by showing that the train-track lower bound matches Proposition 6. p. 6 3 to the
attention of WOR. CD thanks Shukun Wu for pointing out to him the train-track example
relevant to two-ends. A qualitative version of Theorem 1. 8, assuming additionally |Θ| ≥ δ
−τ
for a fixed τ > 0
and taking ϵ sufficiently small in terms of s, τ , yields a gain δ
−η
for some η = η(s, τ ) > 0. For any subarc J ⊂ I of length at least δ, maximality
gives |Θ ∩ J| ≤ (|J|/ρ)
ϵ
|Θ ∩ I|. Rescaling proves the stated Frostman bound.Where the paper says it
Where the paper says it
Where the paper says it
Conclusion
The final result is that R ≤ C(√rw/δ)s where C ≲ δ − O(ϵ) s Σ P. This confirms P is a (δ, s, C)-RKT-set The entire framework successfully manufactures the stronger form of non-concentration required for the result. p. 4 Thus the non-concentration can only be exploited
in the forms permitted by available Furstenberg-type incidence theorems. The three-ends
result is designed precisely to manufacture that stronger form of non-concentration.Where the paper says it
Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.
Improvements for AI systems
-
Enhanced incidence geometry capabilities for lower dimensional structures: The system can prove sharp lower bounds on three-ends incidence estimates for one-dimensional Katz–Tao point sets, which is crucial for analyzing complex geometric configurations in low dimensions.
-
Strong quantitative bounds on approximate rectangles: The AI can derive strong quantitative bounds on the number of approximate rectangles with vertices belonging to four separated camps, each retaining a non-concentration condition, by applying the three-ends incidence estimate.
-
Improved analysis of energy estimates on the paraboloid: The system can solve for an equivalent formulation of approximate rectangle counts using energy estimates on the paraboloid, providing quantitative bounds like E4,delta(P1, P2, P3, P4) ≲ϵ δ −O(ϵ)δ−s(10−3σ)4−σ.
-
Advanced projection theorems for point sets: The AI can apply and generalize projection theorems to establish bounds on the number of points in a set after projecting onto arcs, yielding estimates like Nδ(πx(P)) ≳ϵ δ O(ϵ) Θ σ/4p P.
-
Robust handling of complex hypergraph refinement: The system can perform sophisticated hypergraph refinement arguments, including dyadic pigeonholing and pruning at all scales, to obtain a delta O epsilon fraction of the original rectangles in the proof of Proposition 1.6.
-
Verification and correction of complex proofs: The system can identify substantial errors in existing proofs by comparing them against established theorems, such as correcting a error found in the original proof of Proposition 1.6 by modifying the statement to use Proposition 4.1.
-
Generalization of incidence theorems using mixed structures: The system can apply mixed three-ends configurations involving both tubes and circles to derive bounds on approximate rectangles, as shown in Proposition 4.2, which yields P is a (delta, s, C)-RKT set with C ≲ delta −O(epsilon) s Σ P.
Abstract
We prove a three-ends incidence estimate whose multiplicity exponent is sharp for one-dimensional Katz-Tao point sets. At the heart of the argument lies rectangular non-concentration and its incidence theory introduced earlier by the authors. As an application, we obtain a strong quantitative bound on the number of approximate rectangles with the vertices belonging to four separated camps, each retaining a non-concentration condition.
Sources
- Incidence estimates for quasi-product sets and applications
- Furstenberg-type estimates under mild non-concentration assumptions
- Two-ends Furstenberg inequality for transversal families and applications to Fourier decay
- Furstenberg sets estimate in the plane
- Restriction estimates using decoupling theorems and two-ends Furstenberg inequalities
Related papers
- Multicolor Ramsey numbers of ordered matchings
- Exponential random graph models with soft clique constraints
- A Full-Sequence Quantitative Gap Between the Chromatic and Cochromatic Numbers of a Random Graph
- Ehrhart positivity for paving positroids
- Nonlinear lower-tail large deviations at criticality
- An involution on Dyck paths in the region at most (,) that interchanges area and dinv