Consider what the detection actually turned on. Montarges expected to fail. "Honestly, I thought we did not have the sensitivity to detect Betelgeuse B as it was predicted," he says, and the image exists because the companion is more massive, and therefore brighter, than the models had it [6]. The prior figure was roughly one solar mass; the observations put it at about two to three [7], a revision of a factor of two to three [14]. That puts the earlier failures in a different light: they were tests against a fainter object than the one that is apparently there.
The other constraint was geometry. Two papers published in 2024 predicted that the companion would reach its greatest apparent separation from the supergiant in December 2024 [4]; the team took that date, observed with SPHERE on the VLT in the Atacama, and recorded light from the companion itself rather than an indirect signature [8]. Several months of processing followed [5]. That is why roughly a hundred years of watching produced suspicion and not a picture [3]. For most of the orbit, a faint star sits projected against a naked-eye supergiant [17].
What the release does not give is an orbital period, which is the number you would want in order to test the companion against the brightness variation that ESO says it could help explain [12]. An inference is available, and it is only an inference: if maximum separation fell in December 2024 and the team expects the candidate on the far side about a year after the August 2026 announcement [10], the gap between opposite extremes is near 33 months, implying an orbit of roughly five and a half years [16]. Observing windows rather than orbital mechanics may set that schedule, so treat it as arithmetic, not a result.
Two different brightness phenomena are in play. The Great Dimming, which set off talk of an imminent explosion, was attributed by a Montarges-led VLT study to dust blocking part of the star [11]. The companion is offered against the longer-running variability that motivated the hypothesis a century ago in the first place [3][12].
For supernova modelling, the useful quantity is the mass, and the mass is loose. Two to three solar masses is a range half again as wide as its own floor [7], and the release commits only to the companion possibly influencing the future explosion [12]: no separation, no period, no interaction described. Note also that about 20 months separate the observation from the announcement [15], of which the release accounts for several months of analysis. The confirmation test is cheap and specific, one epoch on the other side of the star [10], and Gemini North already had a possible direct detection [9]. The standard for this class of result is a sequence of independent looks, and the sequence is one look short.
