Spike Structures, Finite-State Exclusion, and Bounded-Exponent Collatz Cycles

Using an accelerated Collatz map, we introduce a structural framework for analyzing hypothetical nontrivial cycles. The odd positive integers partition into three families B, D, C according to their division exponents under the map: elements of B = {8n+1} always yield exponent 2, elements of D = {4n+3} always yield exponent 1, and elements of C = {8n+5} always yield exponents ≥ 3. Within C, a further decomposition modulo 24 into three subfamilies C_1, C_2, G reveals that their associated linear forms exhibit cyclic 2-adic lifting. High powers of 2 dividing these forms rotate systematically through explicit congruence relations. We call this phenomenon spike structures. 

This framework yields two main results. First, any cycle with bounded division exponents must project to a directed cycle in a finite residue graph, reducing cycle search to finite combinatorics. We verify computationally that no cycles exist with all exponents bounded by 10. Second, in the dynamically balanced regime (average exponent equal to 2), forbidden exponent transitions and forced residue constraints combine to exclude all nontrivial bounded-exponent cycles. We verify computationally that no such cycles exist for lengths 50–200 with exponents bounded by 20. 

This approach reframes bounded-exponent cycle search as structural exclusion rather than brute-force enumeration.