This effect seems to have a life of its own, and certainly generating a lot of interesting discussion and experiments. I've mentioned previously on such studies that studied this effect more closely. Well now comes another one, and this one has a slight twist to it.
Abstract: Unlike most of the research on the Mpemba effect which has focused on verifying the observation that warm water freezes faster than cold water, our work quantitatively investigates the rates at which hot and cold water cool and the point at which hot water reaches a lower temperature than cold water under a set of external conditions. Using a vacuum pump to cool samples of water initially at different temperatures, we measured reproducible temperature values at which hot and cold water equilibrate. We have confirmed that warmer water indeed cools at a faster rate than colder water and that, surprisingly, this trend continues past the point where the temperatures of the two samples are the same. Our results show that when using optimal initial temperature conditions, the crossover temperature is found to be 2.7 oC whereas our other set of initial conditions gave a crossover temperature of -0.07 oC. These data taken together provide a definite quantitative evidence of the Mpemba effect.
There you have it. We just have to wait and see if this gets published.