Retrocausation seems to admit of time paradoxes in which events prevent themselves from occurring and thereby create a physical instance of the liar's paradox, an event which occurs iff it does not occur. The specific version in which a retrocausal event is used to trigger an intervention which prevents its own future cause is called the bilking paradox (the event is bilked of its cause).The analysis of Echeverria, Klinkhammer, and Thorne (EKT) suggests time paradoxes cannot arise even in the presence of retrocausation. Any self‐contradictory event sequence will be replaced in reality by a closely related but noncontradictory sequence. The EKT analysis implies that attempts to create bilking must instead produce logically consistent sequences wherein the bilked event arises from alternative causes.Bilking a retrocausal information channel of limited reliability usually results only in failures of signaling. An exception applies when the bilking is conducted in response only to some of the signal values that can be carried on the channel. Theoretical analysis based on EKT predicts that, since some of the channel outcomes are not bilked, the channel is capable of transmitting data with its normal reliability, and the paradox‐avoidance effects will instead suppress the outcomes that would lead to forbidden (bilked) transmissions.A recent parapsychological experiment by Bem displays a retrocausal information channel of sufficient reliability to test this theoretical model of physical reality's response to retrocausal effects. A modified version with partial bilking would provide a direct test of the generality of the EKT mechanism.
General relativity admits closed timelike curves, and quantum mechanics on such a loop is widely believed to defuse the grandfather paradox: in the postselected (P-CTC) model, paradoxical histories are simply assigned probability zero. Here we turn that belief into a sharp, quantitative question. We show that the Doeblin information governing the retrocausal capacity of a noisy quantum channel [Phys.\ Rev.\ Lett.\ \textbf{136}, 230801 (2026)] has a dual description as the \emph{minimal loop amplitude}: the smallest total probability that any legal intervention can leave on a time loop built from that channel. A grandfather paradox is realizable precisely when this amplitude can be driven to zero. This yields a new invariant of a quantum channel --- its \emph{paradox number} n0, the number of independent time loops needed before some entangled intervention annihilates every consistent history. We prove that any channel with full Choi rank has n0=∞ (decoherence protects chronology, no matter how many machines are stacked), that ideal machines have n0=1, and, for the experimentally central amplitude-damping (photon-loss) channel, we solve the problem exactly: n0(γ)=⌈ln2/ln(1+1−γ)⌉, a staircase whose first step is the algebraic threshold γ∗=2(2−1). Equivalently, each loop of transmissivity η contributes log2(1+η) bits toward the single bit that a paradox costs. The proofs are constructive on both sides, and they resolve, for this channel family, the regularization problem left open for the retrocausal capacity.