I have read about the hot water freezing phenonemon in the past. I agree with its theory, under certain controlled conditions. It might go against common reasoning, but if you work it out step by step it kind of makes sense.....
For a given amount of mass, the total cooling duty in BTU/hr to freeze water is composed of a few factors. First there is sensible heat.......this is the cooling requirement to bring the water from its initial temperature to it's freezing point (~32 °F). Then there is latent heat. Latent heat is the cooling duty required to go from liquid to solid @ its freezing point temperature (32 F). Quick analysis tells us that if the water is warmer, the sensible heat (cooling)requirement will be greater thus requiring more cooling duty (BTU/hr).
Now lets take a standard heat transfer equation. DUTY (BTU/Hr) = U*A*DT, where U is the heat transfer coefficient, A is the contact area, and DT is the differential temperature. It's easy to see that if you had a tray of warm water and cold water in the freezer in similar trays, that U and A would be the same. Since the warm water is warmer, the DT is greater relative to the cold ambient conditions of the freezer, hence the amount of duty (cooling) that is absorbed is at a greater rate.
When the cold water tray forms a skin, that acts as an insulating layer and the warmer water in the center of the tray takes longer to cool and freeze. Since the warm water is absorbing heat (cooling) at a faster rate, it cools more homogeneously, and in the right conditions will freeze at a faster rate......
Hope this makes sense!