---
title: 'Lopsided Sets: Asymmetry in Mathematics'
url: https://www.emergentmind.com/topics/lopsided-sets
type: topic
---

# Lopsided Sets: Asymmetry in Mathematics

Searching arXiv for the cited papers to ground the article.
Lopsided sets is a polysemous technical term whose meaning depends strongly on context. In combinatorics and discrete geometry, it denotes families of sign vectors \(L \subseteq \{-1,+1\}^E\) introduced by Lawrence and independently rediscovered by Dress under the name *ample sets*; these families satisfy equality cases in the Dress–Pajor inequality and admit geometric realizations as weakly convex cubihedra in the intrinsic \(\ell_1\)-metric [2603.27835]. In additive combinatorics, the phrase refers to subsets of \(\mathbb Z/n\mathbb Z\) for which the sumset and difference set have unequal sizes [1608.03209]. In amoeba theory, lopsidedness is the dominance of one monomial magnitude over the sum of the others, yielding outer approximations to polynomial amoebas [1608.08663]. In observational astrophysics, a “lopsided set” is a satellite system whose projected angular distribution around an isolated host departs significantly from isotropy [2406.11684]. The common thread is asymmetry relative to a natural symmetric baseline, but the underlying objects, metrics, and theorems are domain-specific.

## 1. Terminological scope

In the cited literature, the term appears in several technically distinct senses.

| Domain | Underlying object | Lopsidedness criterion |
|---|---|---|
| Combinatorics and cube complexes | \(L \subseteq \{-1,+1\}^E\) | Equality in Dress’s criterion; equivalent to Lawrence lopsidedness |
| Additive combinatorics | \(A \subseteq \mathbb Z/n\mathbb Z\) | \(|A+A| \neq |A-A|\) |
| Amoeba theory | \(x \in \mathbb R^n\) for a Laurent polynomial \(f\) | One monomial strictly dominates the others |
| Astrophysics | Satellite system around a host galaxy | Angular distribution is significantly asymmetric |

The combinatorial meaning is historically primary in the supplied corpus. Lawrence introduced lopsided sets in 1983 through the study of intersection patterns of convex sets with orthants of \(\mathbb R^E\), while Dress later defined ample sets by equality in a combinatorial inequality and the two notions coincide [2603.27835]. The other usages retain the same qualitative intuition—deviation from balance, isotropy, or non-dominance—but they do not describe the same mathematical structure.

## 2. Combinatorial lopsided sets and Dress’s ampleness

Let \(E\) be finite and let \(\{-1,+1\}^E\) denote the set of sign vectors \(s:E\to\{-1,+1\}\). For \(L \subseteq \{-1,+1\}^E\) and \(A \subseteq E\), the paper defines the projection and full-fiber operators
\[
L_A := \{ t \in \{\pm1\}^{E-A} : \text{some extension } s \in \{\pm1\}^E \text{ with } s|_{E-A}=t \text{ lies in } L \},
\]
\[
L^A := \{ t \in \{\pm1\}^{E-A} : \text{every extension } s \in \{\pm1\}^E \text{ with } s|_{E-A}=t \text{ lies in } L \}.
\]
These induce two simplicial complexes on \(E\):
\[
\overline{X}(L) := \{ A \subseteq E : L_{E-A} = \{\pm1\}^A \},
\]
\[
\underline{X}(L) := \{ A \subseteq E : L^A \neq \emptyset \}.
\]
Here \(\overline{X}(L)\) records shattered sets and \(\underline{X}(L)\) strongly shattered sets [2603.27835].

The fundamental inequality is the Dress–Pajor inequality
\[
\#\underline{X}(L) \le \#L \le \#\overline{X}(L).
\]
Dress called \(L\) *ample* when
\[
\#L = \#\overline{X}(L).
\]
The paper also states that ample sets satisfy the “sparseness” equality \(\#L=\#\underline{X}(L)\), as proved in earlier work, and that Lawrence’s lopsidedness condition is equivalent to ampleness [2603.27835].

Lawrence’s formulation uses the complement \(L^*=\{\pm1\}^E\setminus L\): \(L\) is lopsided if for all \(A,B\subseteq E\) with \(A\cap B=\emptyset\) and \(A\cup B=E\), either \(A\in \underline{X}(L)\) or \(B\in \underline{X}(L^*)\). This condition encodes a global asymmetry in the occupancy of hypercube fibers. The paper further recalls a family of equivalent characterizations from earlier work, including superisometry of all fibers, superconnectivity, hereditary Euler characteristic \(1\), and total asymmetry [2603.27835].

The examples in the paper illustrate both sides of the dichotomy. For \(E=\{1,2,3\}\), the set \(L=\{s\in\{\pm1\}^3:s(1)=+1\}\) is ample, with \(\#L=4=\#\overline{X}(L)\), while \(\{\pm1\}^3\) with the two constant maps removed is not ample: \(\#L=6\) lies strictly between the sizes of the shattered and strongly shattered complexes [2603.27835].

## 3. Cubihedra, \(\ell_1\) geometry, and metric characterizations

The central geometric realization associates to \(L \subseteq \{\pm1\}^E\) the cubihedron
\[
|L| := \bigcup \{ F \subseteq H(E) : F\cap\{\pm1\}^E \subseteq L \},
\]
where \(H(E)=[-1,+1]^E\) is viewed as a cube complex. The vertices of \(|L|\) are precisely \(L\), its \(1\)-skeleton is the induced subgraph \(G(L)\) in the hypercube graph, and its facets are the maximal faces contained in \(|L|\) [2603.27835].

The ambient metric is the \(\ell_1\) metric
\[
d(x,y)=\|x-y\|_1=\sum_{e\in E}|x(e)-y(e)|.
\]
If \(L\) is connected, \(|L|\) carries an intrinsic path metric \(d_{|L|}\). The paper calls a subset \(K\subseteq \mathbb R^E\) *weakly convex* if it is complete and Menger-convex, equivalently if the intrinsic path metric equals the ambient \(\ell_1\) metric on \(K\). Weak convexity is therefore a path-\(\ell_1\)-isometric condition [2603.27835].

A principal theorem gives the metric characterization of ampleness. For \(L \subseteq \{\pm1\}^E\), the following are equivalent:

1. \(L\) is ample.  
2. \(|L|\) is weakly convex.  
3. \(|L|\) is sign-convex.  
4. \(\operatorname{Baryc}(L)\) is sign-convex.  
5. \(L\) is isometric in \(H^{(1)}(E)\) and every face of \(|L|\) is gated in \((|L|,d_{|L|})\).  
6. \(d_{|L|}\) and the ambient \(\ell_1\) metric coincide on \(\operatorname{Baryc}(L)\).  
7. For any two parallel faces of \(|L|\) with barycenters \(t_1,t_2\), one has \(d_{|L|}(t_1,t_2)=d(t_1,t_2)\) [2603.27835].

This theorem relocates ampleness from a counting identity to a metric-geometric property of cube complexes. In particular, ample sets are exactly those sign-vector families whose cubihedra are isometric subspaces of \(\ell_1\)-spaces in the sense used by the paper. The non-ample six-vertex example above becomes a solid hexagon whose intrinsic metric exceeds the ambient \(\ell_1\) metric on certain pairs, explicitly demonstrating failure of weak convexity [2603.27835].

## 4. Barycenter systems, orthants, and oriented-matroid-like structure

Faces of the hypercube are encoded by barycenter sign vectors in \(\{-1,0,+1\}^E\). If
\[
F=H(A)\times s_0,
\]
then the barycenter \(t\) satisfies \(t(e)=0\) for \(e\in A\) and \(t(e)=s_0(e)\) for \(e\in E-A\). The set \(\operatorname{Baryc}(L)\) of barycenters of all faces of \(|L|\) is upward closed under the partial order \(0\prec \pm1\), and the paper emphasizes the identity
\[
\operatorname{Baryc}(\uparrow J \cap \{\pm1\}^E)=\uparrow J
\]
for any \(J \subseteq \{\pm1,0\}^E\) [2603.27835].

The paper introduces the signed-circuit axiom (SCA): a subset \(J \subseteq \{\pm1,0\}^E\) satisfies SCA if for all \(t',t''\in J\) and \(e\in E\) with \(t'(e)t''(e)=-1\), there exists \(t\in J\) such that \(t(e)=0\) and
\[
t(f)\in \{0,t'(f),t''(f)\}\quad \text{for all } f\in E.
\]
For upward closed \(J\), SCA is equivalent to \(0\)-convexity, and on such sets \(0\)-convexity is equivalent to sign-convexity [2603.27835].

This leads to a barycenter characterization: for \(J \subseteq \{\pm1,0\}^E\) and \(L:=\uparrow J\cap\{\pm1\}^E\), the following are equivalent: \([J]\) is weakly convex, \(\uparrow J\) is sign-convex, \(\uparrow J\) is \(0\)-convex, \(J\) satisfies SCA, \(\uparrow J\) is an isometric subgraph of the grid graph \(G(\{\pm1,0\}^E)\), and \(L\) is ample with \([J]=|L|\) and \(\uparrow J=\operatorname{Baryc}(L)\) [2603.27835].

The oriented-matroid analogy is made through cocircuits. The minimal elements of \(\operatorname{Baryc}(L)\) under \(\prec\) form
\[
\operatorname{Cocirc}(L):=\operatorname{Min}(\operatorname{Baryc}(L)),
\]
and \(\operatorname{Baryc}(L)=\uparrow \operatorname{Cocirc}(L)\). The paper states that \(L\) is ample if and only if \(\operatorname{Baryc}(L)\) satisfies SCA, equivalently if \(\operatorname{Cocirc}(L)\) does, and similarly for the complementary circuit system \(\operatorname{Circ}(L):=\operatorname{Cocirc}(L^*)\) [2603.27835].

The resemblance to oriented matroids is structural rather than complete. SCA is presented as an analogue of weak elimination, but ample sets do not require global symmetry \(t\in R \Rightarrow -t\in R\), nor the existence of the zero covector. The paper situates them instead within the broader framework of COMs, with ample sets characterized as those COMs whose cells are cubes [2603.27835].

## 5. Realizability, projection, and related constructions

Lawrence’s original orthant-intersection viewpoint is recovered through the construction
\[
L(K):=\bigcup_{r\in K}\operatorname{Sign}(r)=\{s\in\{\pm1\}^E:K\cap O(s)\neq \emptyset\},
\]
where
\[
O(s):=\{r\in \mathbb R^E:r(e)\cdot s(e)\ge 0 \text{ for all } e\in E\}.
\]
The paper proves that \(L \subseteq \{\pm1\}^E\) is ample if and only if there exists a weakly convex \(K \subseteq \mathbb R^E\), equivalently a compact weakly convex \(K \subseteq H(E)\), such that \(L=L(K)\). One direction uses \(K=|L|\); the converse derives ampleness from weak convexity of coordinate-hyperplane sections and superisometric fibers [2603.27835].

Projection provides another equivalent characterization. If \(A\subseteq E\), let \(|L|_A\) denote the orthogonal projection of \(|L|\) to \(H(E-A)\), and let \(|L_A|\) be the cubihedron of the projected sign family \(L_A\). Then \(L\) is ample if and only if
\[
|L|_A = |L_A| \quad \text{for all } A\subseteq E,
\]
equivalently if
\[
\dim(|L|_A)=\dim(|L_A|)\quad \text{for all } A\subseteq E
\]
[2603.27835].

These results connect lopsided sets to several established structures. The paper explicitly places them at the intersection of hypercube combinatorics, cube complexes, and \(\ell_1\)-embeddability; it also notes earlier connections to median graphs, CAT(0) cube complexes, convex geometries, antimatroids, and conditional antimatroids [2603.27835]. A plausible implication is that ampleness functions as a unifying criterion for when discrete hypercube data admit faithful \(\ell_1\)-geometric realization without distortion under passage to faces, fibers, and projections.

## 6. Other established uses of “lopsided sets”

In additive combinatorics, a subset \(A \subseteq \mathbb Z/n\mathbb Z\) is called lopsided when its sumset
\[
A+A:=\{a_1+a_2 \bmod n : a_1,a_2\in A\}
\]
and difference set
\[
A-A:=\{a_1-a_2 \bmod n : a_1,a_2\in A\}
\]
have different sizes. The paper distinguishes sum-dominated sets, difference-dominated sets, and balanced sets. For random \(A\) formed by including each residue independently with probability \(p(n)=o(1)\), it proves a phase transition near \(p(n)\asymp n^{-1/2}\): if \(p(n)=o(n^{-1/2})\), then \(|A-A|/|A+A| \to 2\) almost surely; if \(p(n)=c\,n^{-1/2}\), then \(|A-A|/|A+A| \to 1+e^{-c^2/2}\); and if \(\sqrt{\log n}\,n^{-1/2}=o(p(n))\) with \(n\) prime, then \(|A+A|=|A-A|=n\) almost surely [1608.03209]. Here “lopsided” means imbalance between additive and subtractive expansion rather than a sign-vector property.

In amoeba theory, lopsidedness is defined pointwise for a Laurent polynomial
\[
f(z)=\sum_{\alpha\in A} c_\alpha z^\alpha.
\]
For \(x\in \mathbb R^n\), write
\[
m_\alpha(x):=|c_\alpha|e^{\langle \alpha,x\rangle}.
\]
Then \(f\) is lopsided at \(x\) if there exists \(\beta\in A\) such that
\[
m_\beta(x)>\sum_{\alpha\in A\setminus\{\beta\}} m_\alpha(x).
\]
The lopsided amoeba \(\mathcal{LA}(f)\) is the set where \(f\) is not lopsided, and it contains the true amoeba \(\mathcal A(f)\). Purbhoo’s approximation scheme uses cyclic resultants \((f;r)\), with \(\mathcal{LA}((f;r))\) converging uniformly to \(\mathcal A(f)\) as \(r\to\infty\); the paper then develops fast dyadic algorithms for computing \((f;2^k)\) and reports speedups from \(4\times\) to roughly \(30{,}694\times\) relative to general resultant computation in the examples shown [1608.08663]. In this setting, lopsidedness is a dominance inequality that certifies non-membership in the amoeba.

In observational astrophysics, lopsidedness refers to angular asymmetry in the projected distribution of satellite dwarf galaxies around an isolated host. The cited study analyzes 47 MATLAS systems and 21 ELVES systems using six metrics calibrated against \(10^5\) isotropic Monte Carlo realizations per host, with only angular positions randomized and radial distances fixed. The preferred diagnostic is the wedge metric
\[
N_{\mathrm{wedge}}(\alpha)=\max_{\phi_0}\sum_{i=1}^N I\!\left(|\operatorname{wrap}(\phi_i-\phi_0)|\le \alpha/2\right),
\]
with a meta-\(p\)-value to control scanning over opening angles \(\alpha\). Under this metric, about \(16\%\) of tested systems are significantly lopsided, compared to the \(5\%\) null expectation at \(p\le 0.05\); combining all six metrics raises the fraction to about \(21\%\) [2406.11684]. The paper further reports that outer satellites dominate the signal and that restricting all systems to satellites within \(200\) kpc reduces the wedge-significant fraction from \(16.2\%\) to \(7.4\%\) [2406.11684]. Here the term names a statistically significant anisotropy in phase-space projection.

Across these domains, the same word marks an excess on one side of a benchmark: imbalance between orthants, between sumset and difference-set sizes, between monomial contributions, or between angular sectors. This suggests a family resemblance centered on asymmetry detection, but the formal definitions are not interchangeable.

Source: https://www.emergentmind.com/topics/lopsided-sets