Shock revival in core-collapse supernovae (CCSNe) may be due to the neutrino mechanism. While it is known that in a neutrino-powered CCSN, explosion begins when the neutrino luminosity of the proto-neutron star exceeds a critical value, the physics of this condition in time-dependent, multidimensional simulations are not fully understood. \citet{Pejcha2012} found that an `antesonic condition' exists for time-steady spherically symmetric models, potentially giving a physical explanation for the critical curve observed in simulations. In this paper, we extend that analysis to time-dependent, spherically symmetric polytropic models. We verify the critical antesonic condition in our simulations, showing that models exceeding it drive transonic winds whereas models below it exhibit steady accretion. In addition, we find that (1) high spatial resolution is needed for accurate determination of the antesonic ratio and shock radius at the critical curve, and that low resolution simulations systematically underpredict these quantities, making explosion more difficult at lower resolution; (2) there is an important physical connection between the critical mass accretion rate at explosion and the mass loss rate of the post-explosion wind: the two are directly proportional at criticality, implying that, at criticality, the wind kinetic power is tied directly to the accretion power; (3) the value of the post-shock adiabatic index $\Gamma$ has a large effect on the length and time scales of the post-bounce evolution of the explosion larger values of $\Gamma$ result in a longer transition from the accretion to wind phases.

Recently, we developed a pair of meshless finite-volume Lagrangian methods for hydrodynamics: the 'meshless finite mass' (MFM) and 'meshless finite volume' (MFV) methods. These capture advantages of both smoothed-particle hydrodynamics (SPH) and adaptive mesh-refinement (AMR) schemes. Here, we extend these to include ideal magneto-hydrodynamics (MHD). The MHD equations are second-order consistent and conservative. We augment these with a divergence-cleaning scheme, which maintains div*B~0 to high accuracy. We implement these in the code GIZMO, together with a state-of-the-art implementation of SPH MHD. In every one of a large suite of test problems, the new methods are competitive with moving-mesh and AMR schemes using constrained transport (CT) to ensure div*B=0. They are able to correctly capture the growth and structure of the magneto-rotational instability (MRI), MHD turbulence, and the launching of magnetic jets, in some cases converging more rapidly than AMR codes. Compared to SPH, the MFM/MFV methods exhibit proper convergence at fixed neighbor number, sharper shock capturing, and dramatically reduced noise, div*B errors, and diffusion. Still, 'modern' SPH is able to handle most of our tests, at the cost of much larger kernels and 'by hand' adjustment of artificial diffusion parameters. Compared to AMR, the new meshless methods exhibit enhanced 'grid noise' but reduced advection errors and numerical diffusion, velocity-independent errors, and superior angular momentum conservation and coupling to N-body gravity solvers. As a result they converge more slowly on some problems (involving smooth, slowly-moving flows) but more rapidly on others (involving advection or rotation). In all cases, divergence-control beyond the popular Powell 8-wave approach is necessary, or else all methods we consider will systematically converge to unphysical solutions.