This paper argues that the ancestor-simulation hypothesis carries an unexamined commitment about physics. An ancestor simulation, in Bostrom's sense, aims at the actual past, and is therefore constrained by whatever record of that past survives: it is a boundary-value problem, not an initial-value problem. Because chaotic dynamics drive any freely running simulation away from the record within days, a faithful reconstruction requires continuous correction — and invisible correction, since the re…
Read moreThis paper argues that the ancestor-simulation hypothesis carries an unexamined commitment about physics. An ancestor simulation, in Bostrom's sense, aims at the actual past, and is therefore constrained by whatever record of that past survives: it is a boundary-value problem, not an initial-value problem. Because chaotic dynamics drive any freely running simulation away from the record within days, a faithful reconstruction requires continuous correction — and invisible correction, since the record contains no credible report of physical law ever failing. I argue that this requirement selects a kind of physics. If our world is a faithful archive-constrained reconstruction of an actual past, then its physics must be irreducibly stochastic, because only stochastic law permits the corrections that faithfulness requires without detectable violation. The claim's form is a selection effect — a feasibility filter — rather than a design choice. Only indeterministic pasts admit faithful reconstruction; inhabitants of a faithful reconstruction should therefore expect to observe indeterministic physics. Simulator psychology plays no role in the argument. The mechanism has a theological ancestor in the quantum divine action tradition; the archive, the filter, and the resulting expectation are what the paper adds. I develop the proposal through three illustrative mechanisms — outcome curation, Everettian post-selection, and checkpoint resampling — locate it among live positions in the foundations of physics, including measurement independence, the retrocausal programme, and the preservation of quantum statistics, and identify its in-principle signature: statistical residue that would track documentation rather than observation, together with a conditional defeater should physics prove fundamentally deterministic. The paper does not argue that we are simulated, and treats no feature of quantum mechanics as evidence. The argument selects for a class of physics of which quantum mechanics is a member; it does not derive quantum mechanics.