Is there really a W to τν puzzle?

According to the Particle Data Group, the measurements of ${\cal B}(W^+ \to \tau^+ \nu_\tau)$ and ${\cal B}(W^+ \to \ell^+ \nu_\ell)$ ($\ell = e,\mu$) disagree with one another at the $2.3\sigma$ level. In this paper, we search for a new-physics (NP) explanation of this $W \to \tau\nu$ puzzle. We consider two NP scenarios: (i) the $W$ mixes with a $W'$ boson that couples preferentially to the third generation, (ii) $\tau_{L,R}$ and $\nu_{\tau L}$ mix with isospin-triplet leptons. Unfortunately, once other experimental constraints are taken into account, neither scenario can explain the above experimental result. Our conclusion is that the $W \to \tau\nu$ puzzle is almost certainly just a statistical fluctuation.