Can hot water really freeze faster than cold water?

Sometimes, under specific conditions — but the effect is inconsistent, and after decades of experiments there is still no agreed explanation. It is known as the Mpemba effect, after a Tanzanian schoolboy who noticed that hot ice cream mix froze before cold mix.

The honest state of the evidence is that some careful experiments reproduce it and others do not. Attempts to establish it as a robust, repeatable phenomenon have produced conflicting results, and a widely discussed 2016 study failed to observe it at all when the starting conditions were tightly controlled. That does not mean nobody has ever seen it; it means the effect depends heavily on details that are hard to hold constant.

Several mechanisms have been proposed, most of them plausible in particular setups rather than universally true. Hot water loses mass to evaporation, so less water remains to freeze. Warm containers can melt through frost in a freezer and make better thermal contact with the shelf. Dissolved gases, which come out of solution when water is heated, change how the liquid conducts heat and how it convects. Supercooling — water staying liquid below 0°C until something triggers crystallisation — introduces large random variation, and previously heated water may nucleate differently.

What matters is that "freezing" has to be defined. If you measure the moment ice first appears, you get different answers than if you measure when the sample is fully solid.

Practically, you should not boil water to make ice faster. The reliable variables are surface area, container material and freezer airflow — a shallow metal tray beats a deep plastic tub regardless of starting temperature.