Osczevski and Bluestein's calculations repaired many of the gross inaccuracies of the Siple-Passel system. Instead of basing their estimates on the rate at which a bottle of water freezes in the wind, they modeled heat loss using the approximate size and shape of the human face. (They figured that's where we lose the majority of our heat, since we keep the rest of our bodies pretty well covered-up in winter.)

Next, they assumed a baseline wind speed of 3 mph, matched to the average person's walking speed. That accounted for the fact that we create a little breeze for ourselves just by moving around. (The old Siple-Passel system used a value of 4 mph.) They also corrected for the standard wind-speed readings collected at airports, which are measured where gusts tend to be especially strong, at 33 feet above the ground. The new system converted these values to their rough equivalents at a more natural height of 5 feet.

And finally, they tried to account for variations in skin resistance, which determines how well our bodies are insulated from the cold. Someone with a thick, protective coating—an obese person, for example—would lose heat more slowly than someone who's all skin and bones. In other words, the same outside conditions will produce a different wind chill—i.e., a different rate of heat loss—depending on how fat you are. Since better-insulated people are more vulnerable to frostbite, Osczevski and Bluestein assumed that we've all got very thick skin. That way they could gear their readings toward the people who would be most likely to suffer damage from the cold—i.e., the 5 percent of the population with the highest skin resistance.