Does hot water freeze faster than cold?

In 1963, a Tanzanian schoolboy named Erasto Mpemba noticed his hot ice cream mix froze before his classmates' cooled ones, and refused to stop asking why despite his teachers' ridicule. When physicist Denis Osborne visited the school, Mpemba asked him; Osborne tested it, found something to the claim, and in 1969 they published together. They were late by millennia: Aristotle, Francis Bacon, and Descartes had each reported that hot water can freeze faster than cold. This entry is kept for an unusual kind of openness: after two thousand years, science is still arguing about whether the effect is real.

Why it matters

On its face the claim is thermodynamic heresy, since hot water must first become cold water, and should therefore always lag. That is exactly what makes it valuable. If the effect is real, the explanation lives in the difference between a mug of water and the idealized water of textbooks: evaporation, dissolved gases, convection currents, supercooling, the container, the frost under it. If it is not real, the century of positive reports is a case study in how easy it is to measure the wrong thing. Either resolution teaches something, and the second may teach more. Meanwhile the question has escaped the freezer: theorists generalized it to any system relaxing toward equilibrium, asking whether a hotter start can reach the cold state faster.

What has been tried

Kitchen-scale replications succeed and fail in roughly equal measure, which is itself a clue that the effect, if real, is fragile and conditional. The sharpest skeptical study argued that freezing onset is so sensitive to uncontrollable details, especially supercooling and where thermometers sit, that many positive results measured cooling to an arbitrary threshold rather than actual freezing, and found no genuine effect when conditions were controlled. Proposed mechanisms for the positive cases, mass loss to evaporation, gas expelled by heating, vigorous convection sustaining faster heat loss, less supercooling in previously heated water, each has partial support and none commands the field. The generalized version, meanwhile, has been placed on firm ground: in controlled colloidal and spin systems, hotter starting states demonstrably relax faster along shortcuts through the energy landscape, a genuine and now well-studied phenomenon.

Where the edge is

Water, the original defendant, remains unjudged: no protocol reliably produces or forbids the effect, and no mechanism is established for when it appears.

What would count as an answer

A reproducible recipe, specified to the level of container, gas content, and freezer, that makes hot water beat cold on demand, with its mechanism isolated, or a definitive demonstration that it never truly does. Mpemba kept asking. The catalog keeps the question.