Mpemba effect: warmer water can freeze faster than colder water

I have no idea what this has to do with HN, but, one of the best "life hacks" I've ever learned was from my Iraq tours: evaporative cooling.<p>If you're outside on a hot day and your drink is hot (in our case, it was palettes and palettes full of bottled water), get a wash rag or a skivvy shirt wet, wrap it around your drink container, and let it sit out in the sun. In a little while, voila, cold drink.<p>Apropos of nothing.

And nobody understands exactly why.<p>The British Royal Society offered a 1,000 pound prize recently for an explanation, so far this is the winner, though it's not complete:<p><a href="http://www.rsc.org/mpemba-competition/mpemba-winner.asp" rel="nofollow">http://www.rsc.org/mpemba-competition/mpemba-winner.asp</a>

I can perceive this as nothing but a horrible, horrible failure to account for unseen variables.

Didn't they prove this happens only if the water is impure?<p>If the water is purified, ie. steam distilled, the warm vs cold effect disappears?

Back in my high school chemistry class this effect was noted and explained in the following way:<p>Ice needs to form a particular crystalline structure when it freezes. This is what makes ice float in water. The molecules in warm water are moving faster and therefore have a greater chance to wander into this correct formation. The cold water molecules are moving more slowly and simply take more time to find the correct alignment.<p>There wasn't really any proof to this theory, but it was followed with a bunch of anecdotal evidence, like Canadian's washing their cars with cold water, etc.

Newton's cooling law can be written as t = -ln(T(t) - Te)/k + ln(To - Te)/k. T(t) is the object's temperature at t, Te is the surrounding environmental temperature that's assumed constant, To is the object's starting temperature, and k is a constant scaling factor related to the material of the object in question. Since this effect is contrary to that law, the law is wrong, i.e. too simplistic. It's not hard to imagine that the surrounding Te is changed by the object or that k of water+container can change depending on its temperature or purity...

The controller/thermostat on a stabilized freezer turns on and off in a steady state as it reaches setpoint and then drifts away from it when the motor is off. Filling your ice cube trays with hot water forces the thermostat on sooner, makes it stay on longer because of hysteresis in the controller settings and so freezes faster. That is the only demonstrable reason I have ever seen.

Warm water has less disolved gas, and provides less bouyancy so particulates that would lower the freezing point settle to the bottom. These two factors make the water freeze faster.<p>This isn't "unexplained" unless you never took a college level physics course, which is pretty typical of the people who edit WikiPedia pages.

A friend of mine swears that if you're making say a gin &#38; tonic, filling the glass with ice will result in less ice melting and therefore your drink will be colder but less watery.

If I recall correctly, ice frozen after being heated has fewer air bubbles. I wonder if there is a correlation.

This is clearly madness.