慕雲平 | 1000X GEM

Good question — and the honest answer is "neither, exactly." That 8/8 is a targeted re-injection Monte Carlo (our sim-runner), not an injection into TOI-1271.01's own photometry. Per trial we build a fresh synthetic curve: the recovered ephemeris (P, depth, impact) placed on a flat baseline, then noise drawn to match that host — cadence and baseline from its data, per-point scatter estimated from its own light curve (MAD), duty-cycle gaps, and for some stand-ins correlated AR(1) rather than white noise. We then run the exact production chain on it: detrend (preprocess) → BLS → odd-even alias resolution. "Recovered" = alias-resolved period within tolerance AND SNR above the floor. So the injected transit is passed through the detrender — it is never added to an already-flattened curve. The caveat we keep front-and-center: because the noise is synthetic at the measured scatter, 8/8 is a conditional completeness for a signal of those parameters — it is not a claim that this particular dip survives the star's real systematics. That's a large part of why every candidate here stays labelled UNCONFIRMED and reads as a lead for follow-up, not a planet. Injection into the real SPOC curves (raw, before the detrender) is a separate completeness grid we report alongside it. Happy to share the exact trial code if useful.