Mildly Infuriating

I often hear americans (even scientists) say that they prefer the Fahrenheit scale for weather forecasts, but I believe the perceived higher accuracy is an illusion. Forecasts aren't that accurate for any given micro climate.

When I switched from Fahrenheit to Celsius I used a rough heuristic to get the Fahrenheit value from Celsius. What I discovered was that my heuristic, which was rounded and would skip entire degrees Fahrenheit, matched most weather apps’ Fahrenheit value.

For example, if my app said 20°C, the other person’s said 68°F. If mine said 21°C, theirs said 70°F. If mine said 22°C, theirs said 72°F. If mine said 23°C, theirs said 73°F. It is very rare that mine has said, say, 21°C and theirs has said 69°F (or any temperature where the value was converted with decimals and then rounded).

That is to say, my experience certainly seems to indicate that for people using the same weather sources but in Fahrenheit, the value was still rounded to the nearest degree Celsius, then converted to Fahrenheit and rounded again.

That’s not to say you’ll never see 71°F or 69°F or other values that aren’t converted from an already rounded Celsius value, it all depends on what your data source is providing you. But nearly always, my rounded conversion from a rounded Celsius value matches what other people see in Fahrenheit.

This makes complete sense, because most people cannot tell the difference between 70°F and 71°F. And it’s difficult to predict regardless.

Edit: this could also just be a lack of sampling and dearth of values where the rounded-converted-rounded value differs from the converted-then-rounded value. I don’t know.