Not yet reviewed by Pith; the record is open.

An uncontroversial observation of adjoint string breaking is proposed, while measuring the static potential from Wilson loops only. The overlap of the Wilson loop with the broken-string state is small, but non-vanishing, so that the broken-string groundstate can be seen if the Wilson loop is long enough. We demonstrate this in the context of the (2+1)d SU(2) adjoint static potential, using an improved version of the Luscher-Weisz exponential variance reduction. To complete the picture we perform the more usual multichannel analysis with two basis states, the unbroken-string state and the broken-string state (two so-called gluelumps). As by-products, we obtain the temperature-dependent static potential measured from Polyakov loop correlations, and the fundamental SU(2) static potential with improved accuracy. Comparing the latter with the adjoint potential, we see clear deviations from Casimir scaling.