An Altarelli Cocktail for the MiniBooNE Anomaly?

We critically examine a number of theoretical uncertainties affecting the MiniBooNE short-baseline neutrino oscillation experiment in an attempt to better understand the observed excess of electron-like events. We re-examine the impact of fake charged current quasi-elastic (CCQE) events, the background due to neutral current $\pi^0$ production, and the single-photon background. For all processes, we compare the predictions of different event generators (GENIE, GiBUU, NUANCE, and NuWro) and, for GENIE, of different tunes. Where MiniBooNE uses data-driven background predictions, we discuss the uncertainties affecting the relation between the signal sample and the control sample. In the case of the single-photon background, we emphasize the uncertainties in the radiative branching ratios of heavy hadronic resonances. We find that not even a combination of uncertainties in different channels adding up unfavorably (an "Altarelli cocktail") appears to be sufficient to resolve the MiniBooNE anomaly. We finally investigate how modified background predictions affect the fit of a $3+1$ sterile neutrino scenario. We carefully account for full four-flavor oscillations not only in the signal, but also in the background and control samples. We emphasize that because of the strong correlation between MiniBooNE's $\nu_e$ and $\nu_\mu$ samples, a sterile neutrino mixing only with $\nu_\mu$ is sufficient to explain the anomaly, even though the well-known tension with external constraints on $\nu_\mu$ disappearance persists.