33.42 turns microlensing variability into a time-series mosaic tightening court: after a shared bandpass response model, a shared time reference, and a frozen subsegment segmentation dictionary, microimage subsegments must reproduce same-direction smooth common terms across bands, stations, pipelines, epochs, and environment strata, with near-achromatic behavior, zero-lag co-occurrence, and feed-forward hits from lens-plane and source-structure layers; under V08-compatible tighten, this remains only a microlensing residual ledger rather than a new potential, propagation, or cosmology verdict.
Use this section as a compact machine-readable EFT reference.
33.42 rewrites microlensing variability from “continuous noise plus spikes” into an alignable time-series mosaic of microimage subsegments: the issue is not whether a light curve looks complicated, but whether a same-direction smooth common term survives after chromatic, sampling, and timing ledgers are stripped away. Under compat adjudication, this chapter is tighten.
The readout package is five-fold: a subsegment-level common-term consistency index, near-achromatic cross-band robustness, zero-lag co-occurrence across stations and methods, mosaic order stability with ratio checks, and environment-and-structure feed-forward hit rates plus repeatability across epochs. What matters is not one pretty spike, but whether these measurements cohere as one microlensing residual ledger.
The workflow freezes a shared bandpass response model, a shared external time reference, source-structure and blending controls, and a derivative/curvature/pulse-motif segmentation dictionary before arbitration begins. Global relay observing, dual processing paths, stratified stacks by lens and source conditions, and feed-forward prediction cards keep calibration, segmentation, and adjudication from collapsing into one post-hoc pipeline.
False synchronization must collapse under band and subsegment label permutations, time reversal, method swaps, parameter scrambles, quiet reference windows, and band-edge holdouts. If the pattern stays equally “significant” in quiet windows or after dictionary and band-edge stress tests, it belongs to sampling, calibration, or method ledgers instead of the mosaic court.
The pass line requires at least two bands, two stations, two pipelines, multiple epochs, and multiple subsegments to reproduce a same-direction smooth common term with zero-lag closure, near-achromatic behavior, and environment/structure stratification that survives holdouts. Sampling and window-function aliasing, chromatic calibration or blending drift, and intrinsic source chromatic structure remain the named ledgers that must absorb the pattern first.
The chapter ends with one microlensing residual ledger only: if the mosaic common term survives as a near-achromatic, zero-lag, cross-epoch result that feed-forward cards can hit, it strengthens the lens residual chain and routes onward to 33.43. If not, it returns to sampling, chromatic calibration, blending, and model ledgers without being allowed to decide potential ontology or cosmology.