ASAS-SN and TESS peak times of ASASSN-14ko flares in 2022-2026 drift 2.2 days per cycle later than the published decaying-period ephemeris; the period stays near 108 days

Question: do the 2022-2026 flares of the periodic nuclear transient ASASSN-14ko occur at the times that the published ephemeris of Payne et al. (2022) predicts (P0 = 115.2 d, Pdot = -0.0026, fitted to flares up to November 2021)? Many models of this source use this Pdot. Approach: we measured new flare peak times in ASAS-SN Sky Patrol g-band data (nine well-sampled flares from March 2022 to October 2024) and in TESS Sector 98 FFI data (flare n = 38, January 2026). We checked the methods on training flares and on published TESS and Swift peaks. Then we scored three forecasts made only from the 2014-2021 peaks. Results: a constant period fails (all new flares 20-50 d early). The published decaying-period ephemeris is better, but its O-C values increase from -10 d (2022) to +20.5 d (TESS, z = 3.4). The trend is +2.2 +/- 0.2 d per cycle, and it holds for six peak-time variants. A fit to the flares from December 2020 to January 2026 gives Pdot = -0.0002 +/- 0.0008, which is consistent with zero, and a period of 108.18 +/- 0.15 d. The period decreased from about 112.5 d (2018) to about 107.6 d (2021). It did not continue to decrease after that. We give forecasts for the flares in 2026-2027 that separate the models by up to 25 d. Limitations: the ASAS-SN data are aperture photometry and end in February 2025; only one flare (TESS) is after that date; the TESS peak has a 3-d systematic error; the flare peaks have structure, and the residual timing scatter is a few days. A side check showed that public 2026 ZTF alerts did not cover AT2023uqm during its predicted August 2026 flare.

ASASSN-14ko · periodic nuclear transient · repeating partial tidal disruption · flare timing · period derivative · O-C diagram · forecast test · ASAS-SN · TESS · ESO 253-G003 · AT2023uqm