Closed timelike curves (CTCs) are a theoretical testbed for how quantum mechanics behaves when causal order breaks down. Two major quantum models exist: Deutsch CTCs, which impose a fixed-point self-consistency condition on the state inside the loop, and postselected CTCs, which model time travel via teleportation and postselection. Although both avoid logical paradoxes, they generally produce different quantum evolutions for the same circuit. This paper compares the two models directly at the circuit level using a small density-matrix and statevector simulator. The Deutsch model is implemented by solving for self-consistent fixed points; the postselected model by applying the corresponding nonlinear postselection map. Identical test circuits, including grandfather-paradox circuits, CNOT, SWAP, controlled-unitary families, and Haar-random unitaries, are run through both models, and the outputs compared using trace distance. The two models are found to disagree generically, not just numerically. The disagreement stems from a structural difference: Deutsch CTCs enforce consistency with mixed fixed-point states, while postselected CTCs renormalize only successful histories. For example, the X-gate "grandfather" circuit yields a well-defined mixed state under the Deutsch rule but an undefined (null-projection) output under postselection. Across 200 random unitaries, the two prescriptions never agree. All simulation code is included for full reproducibility.
Naol Demisse (Sun,) studied this question.