Is a snow day predictor accurate?
Editorial Team · November 4, 2026
What a school-closure probability can and cannot tell you, and why this site does not publish a hit rate.
A snow-day predictor is a weather translation, not a crystal ball. It takes a public forecast and turns snowfall, freezing rain, wind, and wind chill into two numbers: the chance we would assign to a school closure, and the chance we would assign to bus trouble. Those numbers are only as good as the forecast, the timing, and the huge gap between a weather model and a bus yard.
What we will not claim
We do not publish an accuracy percentage. Doing that honestly would require a verified, national record of board and consortium decisions, matched to the forecast that was available the night before, for many winters. We do not have that dataset, and we will not invent a hit rate to fill the box. Anyone who shows a precise success percentage without showing the record is asking you to trust a slogan.
The score is also capped at 97 on purpose. A forecast is not a guarantee, and boards are allowed to surprise you.
Where the estimate is useful
The overnight and early-morning snow total is the part a forecast often handles well enough to plan around, especially inside two days. If the model shows a clear 12 cm falling through the night in a city that rarely sees snow, you should take the evening seriously: pack a bag, talk to your manager, and know which site you will refresh at dawn. The methodology shows the exact curve, including the lower tolerance we use on the B.C. south coast and the higher tolerance in snow-belt cities.
Freezing rain, narrow lake-effect bands, and the last twenty kilometres of a rural route are the weak spots. A squall can bury one concession road and miss the forecast point. Ice can start an hour earlier than the hourly codes suggested. Confidence on this site drops after the first couple of days for that reason. Later in the week the number is a sketch.
How to use it without being fooled
Treat a high score as a prompt to prepare, and a low score as a prompt not to panic. Do not treat either one as permission to skip the board's site. If the morning decision differs from the score, the decision is the one that counts. The interesting cases are the split ones: buses cancelled, school open. That pattern is common enough in Ontario that we score it separately, and it is explained in bus cancellations versus school closures.
If you want to judge this site, judge whether the factors on the page match the forecast you can see at Environment and Climate Change Canada, and whether we told you plainly what we cannot know. That is a fair test. A fake percentage is not.
Editorial Team. The Editorial Team writes and maintains this site. We document the predictor, check school-board and transportation links against the organizations' own websites, and correct pages when those sites change. We are not a school board, a transportation consortium, or a government office.