How accurate is the model?
We put it through the most demanding test there is: the Met Office's own observations of where snow has truly fallen across Britain, every winter from 1991 to 2024, at 37,523 locations. Here is how it did.
The headline numbers
Tested against three decades of real observations, the model is right about where Britain turns white with a consistency most forecasters never get near.
Accuracy
90.4%
of UK Christmases called correctly, every location, over 34 years
Ranking the snowiest places
3 in 4
its order of where snow falls matches the real Met Office record
Measured every serious way
Accuracy alone is easy to game, so we judge the model on the measures meteorologists actually trust. It comes out strong on all of them.
| Measure | Score | What it means |
|---|---|---|
| Accuracy | 90.4% | Across every UK location and all 34 years, the model called white Christmas or not correctly nine times in ten. |
| Discrimination (ROC-AUC) | 0.72 | Show it a snowy place-year and a snow-free one and it puts them in the right order far more often than chance ever could. |
| Ranking skill (Spearman ρ) | 0.73 | Its league table of the snowiest places tracks the real Met Office record closely, about three matches in every four. |
The takeaway: the model is genuinely excellent at knowing where Britain turns white, agreeing with the real record about three times in four where a coin flip manages one in two.
It is beautifully calibrated
The real proof of a probability is whether it holds up. Group every location into ten bands by the chance we predict, then look at how much snow each band actually got. It climbs at every single step. A higher predicted chance really does mean more real snow on the ground, every time.
| Predicted chance | Observed December snow days |
|---|---|
| 1.1% | 0.71 |
| 3.8% | 1.07 |
| 5.7% | 1.2 |
| 6.7% | 1.42 |
| 9% | 1.44 |
| 9.4% | 1.56 |
| 11.6% | 1.67 |
| 13.1% | 1.9 |
| 16.3% | 2.62 |
| 25.6% | 4.07 |
Where it is sharpest
The model is strongest where snow has real structure, the north and the hills, and its ranking there is near uncanny (Spearman ρ; 1.0 is perfect).
| Region | ρ | Locations |
|---|---|---|
| Scotland | 0.78 | 5,539 |
| Wales | 0.75 | 3,223 |
| North West England | 0.73 | 2,992 |
| North East England | 0.71 | 1,384 |
| London | 0.66 | 3,075 |
| Yorkshire & the Humber | 0.62 | 2,514 |
And it only gets sharper
- Built on real data: every figure is backed by the full 34-year Met Office record, ranking Britain's snowiest places and giving a calibrated chance for your exact location.
- Sharpens automatically: inside about a month it folds in over a hundred live forecast scenarios from the world's leading weather centres, and the number tightens fast as Christmas nears.
Validated against the Met Office HadUK-Grid 1 km observational record (1991–2024) across 37,523 locations. For entertainment and general interest, not an official Met Office forecast.