A personal scale, not a published index. Each point is one outfit, so a single number never has to say "or"; deciding whether the day needs more than one outfit is what the three numbers at the top are for.

Read bottom to top it is a ladder of clothing rather than of degrees, which is why the temperatures are unevenly spaced: adding a jacket buys more degrees than swapping a t-shirt for a long-sleeve shirt. The ends are anchored on not going out at all, so a day that merely wants shorts is a 9, not a 10.

The temperatures below are felt temperatures — what the air does to a person standing in it, wind and humidity and sun included. They are not air temperatures, and on a windy or sunny day the two are a long way apart.

• 10 dangerous heat — avoid outdoors felt 105°
• 9 hot — shorts and a light t-shirt felt 94°
• 8 warm — shorts and a t-shirt felt 78°
• 7 pleasant — jeans and a t-shirt felt 73°
• 6 mild — jeans and a long-sleeve t-shirt felt 70°
• 5 neutral — jeans and a light jacket felt 61°
• 4 chilly — jeans and a warm jacket felt 53°
• 3 cold — coat felt 46°
• 2 freezing — winter coat felt 37°
• 1 brutal — heavy winter coat, hat and gloves felt 16°
• 0 dangerous cold — avoid outdoors felt -34°

Felt temperature is Steadman's apparent temperature in its radiation form, the one the Australian Bureau of Meteorology publishes: AT = Ta + 0.348e − 0.70v + 0.70Q/(v + 10) − 4.25, where e is vapour pressure, v the 10 m wind and Q the net radiation a body absorbs. It was chosen over the usual "feels like" because those are a heat index above about 80° spliced to a wind chill below about 50°, and plain air temperature in between — which is precisely the range this page is for. Steadman's stays continuous across the whole band and takes wind, humidity and sunlight as real inputs.

Q is built as the standard radiation budget for a standing person: the beam scaled by Fanger's projected-area factor, plus diffuse sky and ground-reflected shortwave over the body's sky and ground view factors, plus a longwave term using Brunt's sky emissivity blended toward a black body by cloud cover. The sun and shade figures differ only in Q — shade drops the direct beam and reflects off shaded ground. Q is expressed relative to a radiatively neutral environment, so an overcast night lands on Steadman's non-radiation formula rather than drifting away from it.

Data is Open-Meteo, picked because it is the only free source that publishes direct normal, direct horizontal and diffuse radiation hourly — without those three the sun/shade split cannot be computed at all. No API key, so this server holds no secret. Yesterday's column comes from the same endpoint's past days, which is model analysis for hours that have already happened, not a reading from a thermometer down the street.

The sun figure assumes nothing is between you and the sun. Under an awning, on the north side of a building, or in a street canyon, read the shade column. Real days are spent moving between the two, which is why both are shown.