PowderCow
How a Powder-o-Meter score is built
We'll try to be as open as we can. This is our input so far.
Every resort gets one number, 0 to 11, for how good the skiing is about to be. It is
computed on a server from live Open-Meteo weather, in one place
(netlify/functions/_meter.js), and the numbers below are read out of that
running code rather than written down by hand.
Two hard floors first
- 0 — bare ground. Under 10 cm underfoot including everything still coming, and under 10 cm in the last 48 hours. Not a bad day: no day.
- 1 — rocks and regret. There is snow, but under 30 cm of it, under 5 cm fell all week and nothing is coming. Sun and terrain get no vote.
Otherwise
The score starts at 1, and snow × dryness × calm, plus bluebird, plus terrain is added. Then it is capped and rounded.
Snow — up to 6 points
A centimetre is worth different amounts depending on when it falls: a quarter for 7–3 days ago, full value for the last 48 hours, double for the next 48. Beyond that it is drawn on the chart and scores nothing, because a five-day forecast is not something to book a flight on.
It is a curve rather than a line — 5 weighted cm earns 1.3, 20 cm earns 3.4, 45 cm maxes it out — because real days are mostly small, and a straight line would waste the top half of the scale.
Dryness — a multiplier, 0.55 to 1.25
The only term that can push a day above full value. Measured across the days that actually got 5 cm or more. Snow at 0.5 °C or warmer keeps 55% of its value; at −5 °C it keeps all of it; between −10 and −18 °C it earns a quarter more. Colder than −26 °C the crystals stop sliding and the curve eases back down.
Calm — a multiplier, 0.55 to 1
Strongest wind on any snowing day. 30 km/h or less builds powder; 70 or more builds slab.
Bluebird — a flat bonus
A day with 15 cm or more, immediately followed by a day with 2 cm or less and at least five hours of sun. Averages lie; timing matters.
Terrain — up to 2 points
A fixed editorial number per resort. It never changes with the weather. See the honest caveat below.
Base depth, and the caps
Base only matters when it is thin: under 30 cm the score cannot pass 4, under 60 cm it cannot pass 6. Above that it is shown but not weighed — 60 cm packed under new snow skis like five metres. Fifty fresh centimetres on rocks is rocks.
The 10 and the 11
A 10 needs at least 20 cm of real new snow behind it, so "drop everything" cannot be assembled out of terrain and sunshine alone.
The 11 is a verdict rather than a sum: 30 cm or more across the storm, at −5 °C or colder, in calm wind, on a base of 80 cm, with sun on the way. Hit every one and it is an 11 outright. Backtested against two full seasons it happens a handful of times a year, somewhere on Earth.
The numbers, straight from the running code
| Constant | Value | What it does |
|---|---|---|
| W_OLD | 0.25 | a centimetre that fell 7–3 days ago |
| W_FRESH | 1 | a centimetre from the last 48 h |
| W_COMING | 2 | a centimetre due in the next 48 h |
| SNOW_SAT_CM | 45 | weighted cm that maxes the snow term |
| SNOW_EXP | 0.7 | the curve; below 1 spends the scale where real days are |
| BARE_CM | 10 | below this, with nothing falling, the score is 0 |
| DEAD_WEEK_CM | 5 | less than this in 7 days counts as no new snow |
| DRY_COLD_C | -5 | at or below this, snow keeps full value |
| DRY_WET_C | 0.5 | at or above this, snow keeps 55% |
| COLD_PEAK_C | -10 | coldest snow earns its biggest bonus here |
| COLD_PEAK | 1.25 | and that bonus is this much |
| WIND_OK_KMH | 30 | calm enough for full value |
| WIND_BAD_KMH | 70 | gale; 55% value |
| SUN_BONUS | 2 | the bluebird bonus, added whole or not at all |
| DEPTH_GATE_LOW | 30 | base below this caps the score at 4 |
| DEPTH_GATE_MID | 60 | base below this caps the score at 6 |
| TEN_CM | 20 | new snow needed before a 10 is possible |
| ELEVEN_CM | 30 | storm total needed for an 11 |
| ELEVEN_TEMP_C | -5 | snow temperature needed for an 11 |
| ELEVEN_WIND | 18 | wind ceiling for an 11 |
| ELEVEN_BASE_CM | 80 | base needed for an 11 |
What each score is called
| Score | Label |
|---|---|
| 0 | Bare ground |
| 1 | Rocks & regret |
| 2 | Ice rink |
| 3 | Stay home |
| 4 | Groomer day |
| 5 | Decent turns |
| 6 | Fun out there |
| 7 | Worth a stop |
| 8 | Worth a detour |
| 9 | Worth a flight |
| 10 | Drop everything! |
| 11 | SELL YOUR HOUSE!! |
Where this is weakest
Terrain is the one number on the site that is not data, and it is uneven. The southern resorts were scored by hand. The northern ones started from measured vertical drop within a few kilometres of the village, which is a poor stand-in for steeps, trees and off-piste and is wrong in both directions: Jackson Hole came out at 2 of 10 because the measuring box never reached the summit, while resorts with long lift-served vertical over groomed pistes came out at 9.
Around forty well-known mountains now carry a hand-written score instead, and everything else keeps the measured one. That leaves a long tail still judged by vertical drop. Terrain is worth 2 points of 11, so the damage is bounded — but if you think one of these is wrong, you are probably right.