Skip to content

Logarithmic Escape Theorem

Statement

Theorem (Logarithmic Escape). For any Dropping Set Dsetk\text{Dset}_k with period P=2ksP = 2^{k-s}, the maximum number of consecutive self-transitions (drops where both nn and dest(n)\text{dest}(n) belong to Dsetk\text{Dset}_k) starting from nn is at most:

mlogP(n)1m \leq \log_P(n) - 1

Equivalently: no orbit can remain in the same dropping set for more than O(logn)O(\log n) consecutive drops.

Significance

The contraction ratio alone would allow long self-chains (e.g., Set13_{13} with ratio 0.949 could theoretically self-loop ~124 times from n105n \sim 10^5). The modular tightening bound is much stronger and independent of the contraction ratio — it depends only on the period PP. This forces orbits to change dropping sets frequently.

Examples

Bounds by set:

kkPPBoundn=106n = 10^6 gives mm \leq
34log4n\log_4 n9
616log16n\log_{16} n4
832log32n\log_{32} n3
13256log256n\log_{256} n2

Verified chains for Dset3\text{Dset}_3 (P=4P = 4):

mmRequired modulusSmallest nnChain
1n1(mod16)n \equiv 1 \pmod{16}17171317 \to 13
2n1(mod64)n \equiv 1 \pmod{64}6565493765 \to 49 \to 37
3n1(mod256)n \equiv 1 \pmod{256}257257193145109257 \to 193 \to 145 \to 109
4n1(mod1024)n \equiv 1 \pmod{1024}102510257695774333251025 \to 769 \to 577 \to 433 \to 325
7n1(mod65536)n \equiv 1 \pmod{65536}655378-element chain

Proof

By induction on chain length mm (modular tightening).

Each self-transition multiplies the required modular constraint by PP.

Claim: mm consecutive self-transitions require nn to satisfy a specific residue class mod Pm+1P^{m+1}.

Base (m=0m = 0): Membership in Dsetk\text{Dset}_k requires nr(modP)n \equiv r \pmod{P}.

Inductive step: Suppose mm self-transitions require nrm(modPm+1)n \equiv r_m \pmod{P^{m+1}}.

Write n=Pm+1a+rmn = P^{m+1} \cdot a + r_m. By the Affine Orbit Structure:

dest(n)=3sPn+C=3sPma+(const)\text{dest}(n) = \frac{3^s}{P} \cdot n + C = 3^s \cdot P^m \cdot a + (\text{const})

For dest to lie in Dsetk\text{Dset}_k mod Pm+1P^{m+1} (not just mod PmP^m), we need:

3sPma(target)(modPm+1)    3sa(target)(modP)3^s \cdot P^m \cdot a \equiv (\text{target}) \pmod{P^{m+1}} \implies 3^s \cdot a \equiv (\text{target}) \pmod{P}

Since gcd(3s,P)=gcd(3s,2ks)=1\gcd(3^s, P) = \gcd(3^s, 2^{k-s}) = 1, this uniquely determines amodPa \bmod P, giving:

nrm+1(modPm+2)n \equiv r_{m+1} \pmod{P^{m+2}}

Since nPm+1n \geq P^{m+1} is required, the chain length satisfies mlogP(n)1m \leq \log_P(n) - 1.

Corollaries

Corollary 1. No orbit can "camp" in a slow-contracting set. Even Set13_{13} (P=256P = 256, ratio 0.949) forces escape after ~2 steps for n<106n < 10^6.

Corollary 2. Combined with the Bit Destruction Bound, this gives progress guarantees: orbits cannot accumulate many low-destruction drops from any single set.