Type A/B shadow pullbacks as a structured odd-covering problem

Status: theory bridge / prior-art orientation

Date: 2026-08-14

Claim boundary: this note does not claim a solution of the Erdos-Selfridge odd covering problem, Direct-Shadow Completeness, Lopez Type A/B coverage, or the Erdos-Straus conjecture.

This note connects the FCF/CENTL shadow program to classical covering-system theory.

Read with:

• DIAMOND.md
• DIRECT-SHADOW-COMPLETENESS.md
• DIRECT-SHADOW-K1000.md
• SHADOW-COVER-GEOMETRY.md
• PRIOR-ART.md

1. Pullback formulation

For a fixed admissible Type A/B candidate (k,h,t), write

Every earlier layer j<k induces a forbidden parameter system

Because every 4j-1 is odd, every nontrivial pullback modulus q_j is odd.

Thus union-shadowing of the candidate is exactly the statement that the structured family

covers all integers.

A direct shadow is the degenerate local case where one earlier layer alone supplies all residue classes modulo its q_j.

Direct-Shadow Completeness asks whether, for this special Type A/B-generated family, collective covering can ever occur without such a local complete layer.

2. Relation to the Erdos-Selfridge odd covering problem

Classical covering-system theory studies finite unions of congruence classes that cover all integers. A famous open problem of Erdos and Selfridge asks whether a covering system can exist with all moduli odd, distinct, and greater than one.

Relevant sources include:

• Song Guo and Zhi-Wei Sun, On odd covering systems with distinct moduli, arXiv:math/0412217.
• Jackson Hopper, On covering systems of integers, arXiv:1705.04372.
• Joshua Harrington, Yewen Sun, Wing Hong Tony Wong, Covering systems with odd moduli, arXiv:2104.00602.
• Chris Bispels et al., A further investigation on covering systems with odd moduli, arXiv:2507.16135.

A 2026 formalization by Ibrahim Mian and Shayaan Siddique, arXiv:2607.25628, gives kernel-checked finite exclusions for the distinct-odd-modulus problem while explicitly treating the general problem as open.

The Type A/B pullback problem is not identical to the Erdos-Selfridge problem:

1. pullback moduli q_j can repeat;
2. one earlier layer can contribute multiple residue classes R_j at the same modulus;
3. the residues are not arbitrary: they are affine pullbacks of divisor-generated trap sets

1. the moduli themselves arise as quotients of 4j-1 by gcds with a fixed candidate modulus L;
2. prime-realization requires a reduced uncovered class, not merely an uncovered integer.

Therefore a proof of DSC-P would establish a no-cover theorem for a special arithmetic subclass of odd covering systems, not resolve the general odd covering problem.

3. Why this bridge matters

The covering-system viewpoint gives precise language for the obstruction we are trying to understand.

General proper congruence classes can collectively cover the integers. Consequently,

is nowhere near enough in an arbitrary covering problem.

But in the exact Type A/B computations through k=1000, every directly novel candidate admits a reduced uncovered progression.

Thus the research question is now:

What arithmetic property of the divisor-generated Type A/B pullback family prevents it from behaving like an arbitrary odd covering system?

That question should be attacked using both the special algebra of T_j and known covering-system obstructions.

4. A prime-power coordinate model

Let

Then the parameter line modulo Q decomposes by CRT into prime-power coordinates

Each forbidden system (q_j,R_j) depends only on the coordinates corresponding to prime powers dividing q_j.

This turns Direct-Shadow Completeness into a finite constraint-satisfaction problem on an odd prime-power product.

A useful taxonomy is:

• unary constraints: q_j is a prime power, so the event depends on one coordinate;
• binary constraints: q_j has two distinct prime factors;
• higher-support constraints: q_j has three or more distinct prime factors.

Preliminary diagnostics on difficult certified candidates show a strong concentration in unary and binary support. This suggests that the pullback family may have a low-complexity local core even while the raw number of earlier constraints is large.

5. Canonical coordinate experiment

For each prime-power coordinate ell^a of Q:

1. combine all unary Type A/B constraints acting only on that coordinate;
2. choose an allowed local residue, preferring 1 whenever it survives;
3. combine the local choices by CRT;
4. count the remaining violated multi-prime constraints;
5. attempt to repair them by changing the smallest possible number of prime-power coordinates while preserving all unary constraints.

This is not merely a faster witness search. It asks whether the global survivor can be constructed locally.

If every directly novel candidate admits a survivor after a bounded number of local repairs, that would suggest a much sharper theorem than raw DSC-P.

6. Candidate theorem hierarchy

Coordinate-Core Conjecture

Every directly novel Type A/B pullback family has a prime-power coordinate assignment satisfying all unary constraints and all but finitely controlled multi-prime obstructions.

Bounded-Repair Conjecture

There exists an absolute or structurally bounded repair number B such that every directly novel candidate can be made globally avoiding by changing at most B prime-power coordinates from a canonical unary-safe assignment.

DSC-P

Every directly novel candidate has a reduced avoiding parameter class and therefore infinitely many exact-depth prime realizations.

A proof of a strong bounded-repair theorem could imply DSC-P by an explicit construction.

7. Proof mechanisms to investigate

1. Unary saturation: classify when the combined prime-power-only constraints can cover a coordinate. Determine whether such local saturation is equivalent to a direct shadow or forces one elsewhere.
2. Binary graph structure: build the graph whose vertices are prime-power coordinates and whose edges are binary pullback constraints. Study degeneracy, cores, cycles, and whether the forbidden edge relations have a common algebraic orientation.
3. Residue 1 bias: the trap fact 1 notin T_j may survive affine pullback in a weakened coordinate form, making the all-ones assignment a natural base point except on a small exceptional set of coordinates.
4. Minimal-cover contradiction: assume a minimal Type A/B union cover and apply classical covering-system necessities to its odd moduli, then use the special divisor-generated residues to contradict minimality.
5. Character signatures: test quadratic and higher multiplicative characters of the allowed and forbidden local residues.
6. Prime-power valuation signatures: test whether forbidden pullbacks force incompatible valuation patterns around r+Ls+e or r+Ls+4e.

8. Research significance

This bridge sharpens the novelty target.

The generic concept of a congruence covering is classical and must not be claimed as new. The candidate novelty is that the Lopez Type A/B system appears to generate a special family of odd covering problems with an unexpectedly strong local-to-global noncoverage phenomenon.

If universal DSC-P is proved, one defensible interpretation would be:

Type A/B first-hit realizability admits a complete local obstruction theory because the associated divisor-generated odd pullback systems cannot collectively cover unless a direct shadow is already present.

That would be a theorem about a structured subclass of odd covering systems and, simultaneously, a theorem about the exact-depth geometry of the Erdos-Straus Type A/B system.

9. Immediate work

The project should now run two attacks in parallel:

• continue exact candidatewise falsification beyond k=1000;
• mine the certified k<=1000 bundle for the prime-power coordinate invariant that explains the absence of covers.

The second attack is now higher value than merely extending the numerical range.