The use of ice slurry to cool fish at sea is widespread. Regulations exist for cooling to a particular temperature in a given time. Here we synthesis a quantitative failure assessment of the impact of naturally occurring, chance fluctuations in slurry temperature and fish mass on cooling time. It is demonstrated with Southern Bluefin Tuna (Thunnus maccoyii) (SBT), a premium farmed fish. The assessment is based on the transient cooling model of Davey (Davey, 2015, LWT 60 (1), 308-314) together with a cooling risk factor (p). Cooling is simulated using Monte Carlo (Latin Hypercube) sampling of parameters. Results revealed that 6.93% of freshly harvested SBT could fail to cool to a Regulatory 5 degrees C in <= 12 h, averaged over a long period. Quantitative evidence is used to show variance about the mean fish mass is the controlling parameter, and that cooling failures can be mitigated if the slurry temperature is maintained close to the theoretical minimum of 0 degrees C (not withstanding any acceptable use of brine slurries). Because all practical scenarios that can exist for ice slurry cooling at sea are computed, the new failure assessment is a significant advance over traditional approaches. It will be of immediate practical use. Crown Copyright (C) 2016 Published by Elsevier Ltd. All rights reserved.