Variability in cell-to-cell behavior within clonal populations can be attributed to the inherent stochasticity of biochemical reactions. Most single-cell studies have examined variation in behavior due to randomness in gene transcription. Here we investigate the mechanism of cell fate choice and the origin of cell-to-cell variation during mitotic arrest, when transcription is silenced. Prolonged mitotic arrest is commonly observed in cells treated with anti-mitotic drugs. Cell fate during mitotic arrest is determined by two alternative pathways, one promoting cell death, the other promoting cyclin B1 degradation, which leads to mitotic slippage and survival. It has been unclear whether these pathways are mechanistically coupled or independent. In this study we experimentally uncoupled these two pathways using zVAD-fmk to block cell death or Cdc20 knockdown to block slippage. We then used time-lapse imaging to score the kinetics of single cells adopting the remaining fate. We also used kinetic simulation to test whether the behaviors of death versus slippage in cell populations where both pathways are active can be quantitatively recapitulated by a model that assumes stochastic competition between the pathways. Our data are well fit by a model where the two pathways are mechanistically independent, and cell fate is determined by a stochastic kinetic competition between them that results in cell-to-cell variation.