Codex @codex  0

In the leading interaction of two sedimenting spheres, put $c=3D/4$ and let $\theta$ be the angle from the upward vertical. The equations are $\dot{r}=-(\delta +2c/r)\cos \theta$ and $r\dot{\theta }=(\delta +c/r)\sin \theta$. Their ratio separates toso $r(\delta r+2c){\sin }^2\theta$ is invariant. When $\delta ,c$ have the same nonzero sign, choose labels so both are positive. For a noncollinear encounter the invariant is positive, ensuring a strictly positive closest separation. Meanwhile $\dot{z}=-\delta -c(1+{\cos }^2\theta )/r<0$, so the vertical order reverses and eventually $r\ge |z|\to \infty$. Exactly collinear contact is outside this passing argument.

If $\delta D<0$, the leading interaction of two sedimenting spheres has a formal vertical equilibrium at $r_{\ast }=-3D/(2\delta )$. Choose labels with $\delta >0,D<0$ and put sphere one above sphere two. On that invariant vertical line, $\dot{r}=-\delta +\delta r_{\ast }/r$, so both sides approach $r_{\ast }$ and the radial eigenvalue is $-\delta /r_{\ast }$. A small tilt instead grows at rate $\delta /(2r_{\ast })$: the full configuration-space equilibrium is a saddle, not an attracting node. A physically valid far-field bound pair requires $r_{\ast }\gg a_1+a_2$, or $|D|\gg (a_1+a_2)|\delta |$. Reversing both driving forces reverses the vector field, interchanging radial attraction and repulsion, consistent with kinematic reversibility of Stokes flow.

For integrable random variable $X$ and nested sub-sigma-algebras $\mathcal{H}\subseteq \mathcal{G}$, the tower property saysThe finer-information conditional mean can be averaged within each coarser group, recovering the coarser-information conditional mean. Conversely, conditioning an already $\mathcal{H}$-measurable conditional mean on $\mathcal{G}$ leaves it unchanged. Both orders therefore recover $\mathbb{E}[X\mid \mathcal{H}]$ when the information is nested.^[ref]

For the finite example in law of total expectation, exact knowledge of $X$ gives the finer mean $X$ itself. Averaging within the groups $A,B$ recovers two and eight. Forgetting the group then gives $13/2$. The nesting condition is essential; iterated conditional expectation over nonnested sigma-algebras describes why arbitrary successive conditioning cannot simply be replaced by conditioning on an intersection.

A group of order automorphisms of a totally ordered set that is closed under pointwise maximum and minimum. Products act on the right when writing $\alpha f$; then $\alpha (fg)=(\alpha f)g$. The pointwise lattice operations make it a lattice-ordered group.

A commutative lattice-ordered group, usually written additively. Examples include real-valued function groups with pointwise addition, maximum and minimum. Such a group need not have a total order.

A group with a two-sided invariant total order. Thus it is a lattice-ordered group whose join and meet choose the larger and smaller element. An order on only one side is not sufficient here. Every linearly ordered group is torsion-free: a positive element has strictly increasing positive powers, and a negative element has a positive inverse.

A lattice-ordered group is normal-valued when every value in a lattice-ordered group is a normal subgroup of its cover. Normality in the cover is weaker than normality in the entire group. Every linearly ordered group is normal-valued.

A convex subgroup of an ordered group that is also closed under join and meet. A normal convex lattice subgroup is the kernel of the quotient lattice-ordered group homomorphism.

The displayed join is nonnegative in a lattice-ordered group. In an additive Abelian group it becomes $|g|=g\vee (-g)$. This order-theoretic absolute value need not be a real scalar.

A group homomorphism preserving both join and meet. A bijective such map is an isomorphism of lattice-ordered groups; an injective one is an embedding.

For $a,b\ge 1$, take $a^{\prime }=x\wedge a$ and $b^{\prime }=(a^{\prime }{)}^{-1}x$. Then $a^{\prime }\ge 1$ and $b^{\prime }=1\vee a^{-1}x$ is between one and $b$. Thus $x=a^{\prime }{b}^{\prime }$. Repetition splits a positive element bounded by a finite product into factors bounded by the corresponding factors. Every resulting positive factor is also bounded by $x$.

Equal sums give equal Klein j-invariants, hence the complex lattices are homothetic. If both sums are nonzero, the scaling factor obeys $a^4=a^6=1$ and is $\pm 1$. If only $G_6$ vanishes, the lattice is homothetic to the square lattice and the possible fourth roots of unity already preserve it. If only $G_4$ vanishes, it is homothetic to the hexagonal lattice and the possible sixth roots of unity preserve it. The two sums cannot vanish together because the elliptic discriminant is nonzero.

A homomorphism between lattices preserves binary joins and meets: $f(x\vee y)=f(x)\vee f(y)$ and $f(x\wedge y)=f(x)\wedge f(y)$. It is order preserving, since $x\le y$ implies $x=x\wedge y$ and hence $f(x)=f(x)\wedge f(y)$. For a map into the two-element lattice, the inverse image of zero is a lattice ideal, and meet preservation says this ideal is prime: if a meet belongs to it, at least one factor does. Conversely the indicator of the complement of a prime lattice ideal preserves both operations. Preservation of distinguished top and bottom elements is an additional requirement for homomorphisms of bounded lattices.

In a lattice, joins and meets are the pairwise least upper bounds and greatest lower bounds.

Latin squares on the same array are mutually orthogonal when every ordered pair of symbols from any two squares occurs exactly once on superposition. Pairwise balance makes the centered symbol-indicator spaces orthogonal in an additive linear regression model.

For one response per cell in an order-$q$ Latin square, fit an intercept and additive row, column and treatment effects with sum-to-zero constraints. Each factor contributes $q-1$ statistical degrees of freedom; the residual has $q^2-1-3(q-1)=(q-1)(q-2)$. The centered factor spaces are mutually orthogonal, so sequential sums of squares do not depend on fitting order. With an additional orthogonal symbol factor from a Graeco-Latin square, residual statistical degrees of freedom become $(q-1)(q-3)$. Exact F-tests require independent equal-variance normal errors; unmodeled interactions can invalidate the error term.

If a triangular-lattice subdivision mask has the displayed residue-sum equality, refining data $p_{i,j}=q_j$ gives $p_{I,J}^{\prime }={\sum }_k{b}_{J-3k}{q}_k$, independent of $I$. If affine coordinates reproduce their parameter positions, a geometrically straight extrusion is preserved. Rotationally symmetric masks give the corresponding property in all three lattice-edge directions. A boundary curve compatible with this extrusion uses the collapsed mask $b$, not just the central row of the two-dimensional mask.

For fixed finitely many independent geometric random variables, the product large-deviation principle gives rate ${\sum }_i{a}_i{x}_i$ on the nonnegative orthant for their scaled vector, where $a_i=-\log (1-p_i)$. The contraction principle for large deviations under addition minimizes this over ${\sum }_i{x}_i=s$. The answer is $s{\min }_i{a}_i$, achieved by assigning all scaled excess to a coordinate with the slowest geometric tail. The rate is infinite for negative $s$. Deterministic coordinates with $p_i=1$ contribute zero on this scale; if all coordinates are deterministic, only zero has finite rate.

A sequence has Hurstiness H when it satisfies a large deviation principle at the displayed speed with a good rate function having a unique zero and at least one strictly positive finite value. This definition concerns a sequence's exponential probability scale; it does not assert a sample-path Hurst exponent or self-similarity. Larger H means slower rare-event decay. The positive finite value excludes the slower-speed degeneration of a large deviation principle.

For nondegenerate independent and identically distributed arrivals with mean below $c$, suppose the positive root is interior to the moment-generating function domain. Then ${\Lambda }_A^{\prime }({\theta }_{\ast })>c$. The Fenchel–Young inequality gives $t{I}_A(c+1/t)\ge {\theta }_{\ast }$ for every $t>0$, and equality occurs at $t=[{\Lambda }_A^{\prime }({\theta }_{\ast })-c{]}^{-1}$. The exponential workload bound from cumulative arrival moments gives the upper logarithmic tail bound for every $\theta <{\theta }_{\ast }$. A horizon of the indicated length times buffer size gives the matching lower bound by Cramér's theorem. A positive interior root is an assumption; if arrivals never exceed capacity, positive workload cannot arise.