Reassessment of the reported correlations between gravitational waves and neutrinos associated with SN 1987A

Correlations of considerable apparent significance have been reported between data taken by two bar-type gravitational wave detectors and particle events recorded in the Mt. Blanc, Kamiokande, and IMB particle detectors during a 2-h period near the explosion of Supernova 1987A. In particular, the correlations among the gravitational wave detectors and the Mt. Blanc neutrinos were claimed to have a chance probability of less than ${10}^{\mathrm{-}6}$. If this low probability implies that the correlations are a real physical effect, then new physics will be required to explain them. However, one of the statistical tests used to establish these correlations is seriously flawed, and most others were devised a posteriori and contain considerable freedom to make choices that affect the probability of finding correlations. By a careful consideration of these free parameters, and by applying similar analysis methods to simulated pseudorandom data sets, we show that the actual frequency with which correlations similar to those in the Mt. Blanc data would occur in random data streams is between 0.1% and 10%. Moreover, if the Mt. Blanc correlations were real, then one would expect them in the other particle detectors. After inspecting the evidence, we also conclude that there are no physically significant correlations of the Mt. Blanc–type between the gravitational wave detectors and the Kamiokande and/or IMB particles. This makes it very likely that the Mt. Blanc correlations are due, not to any physical effect, but simply to chance.