The Mpemba Effect: The 2,000-Year-Old Mystery Behind Why Hot Water Can Freeze Before Cold

Picture two identical glasses of water, one boiling hot and one straight from the tap, going into the freezer at the same time. Common sense says the cold one should freeze first — it’s already got a head start, right? Not always. Sometimes the hot water wins, and this small thermal absurdity has been driving philosophers and physicists up the wall for over two thousand years.

It’s called the Mpemba effect, and while it sounds like a lab prank, it’s a real, documented phenomenon that in 2026 finally got a serious explanation, courtesy of a supercomputer simulation.

The student who talked back to his teacher

The official story starts in Tanzania in 1963. Erasto Mpemba was a secondary school student making homemade ice cream with his classmates: they mixed milk with sugar, boiled it, then rushed it into the freezer. In a hurry not to lose his spot, Mpemba shoved his mixture in while it was still hot, while his classmates waited for theirs to cool first. Result: his froze first. When he told his physics teacher, the story goes, the response was basically “that’s Mpemba physics, not real physics.”

Luckily, the kid didn’t let it go. Years later he put the same question to a visiting physicist, Denis Osborne, who actually took it seriously enough to test it in the lab. Surprise: under certain conditions, hot water really could freeze before cold water. The pair published the finding together in 1969, and the effect has carried Mpemba’s name ever since. The funny part is he wasn’t even the first to notice it — Aristotle had written down something similar centuries earlier, and so had Francis Bacon and René Descartes, none of them quite sure what to make of it.

A puzzle with too many loose pieces

For decades, scientists threw explanations at the wall without ever fully agreeing on one: that hot water evaporates and loses mass (so there’s simply less water left to freeze), that boiling drives out dissolved gases (changing how ice crystals form), that it sets up convection currents that spread the cold more evenly, or that cold water is more prone to “supercooling” without actually solidifying. The trouble was that no single explanation fit every experiment, and some labs couldn’t even reproduce the effect at all. For a while, more than one physicist suspected the Mpemba effect might just be a statistical mirage.

What changed in 2026

This is where the new research comes in. A team at the Jawaharlal Nehru Centre for Advanced Scientific Research (JNCASR) in India built the first supercomputer simulation able to reproduce the freezing process atom by atom, and published their results in Communications Physics, part of the Nature group.

According to their model, the key isn’t so much how much heat the water loses, but the path it takes while cooling down. Before turning into actual ice, water passes through unstable in-between states — molecular “pit stops,” so to speak — that can delay the formation of the very first ice crystals, a process physicists call nucleation. Depending on its starting temperature, water can get stuck at that pit stop for longer or shorter periods. The simulation suggests that, under the right conditions, hot water finds a shortcut to nucleation and dodges the delays that cold water gets stuck with, crossing the finish line first.

There’s an extra twist, too: the team found a similar pattern showing up in other physical systems that have nothing to do with water, which hints that the Mpemba effect might be a general feature of how matter switches from liquid to solid, not some quirk unique to H₂O.

Is this good for anything besides dinner-party trivia?

Beyond the fun anecdote, understanding the kinetics of these phase changes has genuinely useful applications — from improving cooling and thermal management in electronics, to fine-tuning industrial processes where controlled cooling matters, like manufacturing certain materials or cryopreservation.

So next time someone insists hot water can never freeze faster than cold, you’ve got the receipts: we’ve been arguing about it for two thousand years, a Tanzanian high schooler was right all along, and now a supercomputer has proven it atom by atom. Not bad for something that started as an excuse to skip the line for the freezer.