Free nightly offsets limit a 40-visit GJ 3543 signal test to 61.82 percent balanced accuracy

We test the information left by a stable 9.161-day sine wave when each night has a free velocity offset. A weighted range bound applies to any fixed schedule with nightly windows at most four hours. For 40 visits, amplitude 2.73 m/s and independent Gaussian errors of 1.6401 m/s, it gives at most 61.82 percent balanced accuracy against no signal, even with a known waveform. Power at a 1 percent false alarm rate is at most 4.23 percent. This is a uniform limit over allowed offsets, not the accuracy of choosing between two specified nonzero aliases. At the same noise and window, 95 percent balanced accuracy needs at least 1198 visits; this condition is not sufficient. Random weighted designs, direct offset fits, Gaussian simulations and a sharp control check the proof. A public HARPS table supplies the error scale and visit-count check. We do not measure real offset variance, find a planet or give independent instrument confirmation. Code, frozen input, hashes, tables, figure and offline steps are included.

astronomy · GJ 3543 · radial velocity · nightly offsets · observation design · Gaussian test · information bound