Forecast comparison

Forecasts compared, source by source

Temp

Sources

This page puts several weather sources side by side for the same place and the same day, next to the reliability each one has earned by measuring its forecasts against what actually happened.

Pick a location in the search box above to see the comparison.

Why does every weather site say something different?

Every site starts from a different prediction model, and for the same day and the same town they may not agree: one says 24 degrees and another 28, or one gives rain in the afternoon and the other does not. Here we do not pick one and hide the rest: they all go into the same table, same place and same day, so you can see at a glance where they agree —and there you can trust them— and where they do not.

How do you know which one gets it right?

We keep what each source predicted and, when the day arrives, compare it with what the official stations measured. It is not a score we make up, nor one they award themselves: if a source said 30 degrees and it turned out 27, it is marked down as three degrees off; and so on day after day, until we know how far off each one runs on average in each area.

Can I trust a ten-day forecast?

Less than tomorrow's, and we would rather say so than gloss over it: the further ahead you look, the less any source gets right, and that happens to all of them. So accuracy is measured separately by how many days ahead —one, two, three, five, seven and ten— instead of giving one single score that mixes tomorrow's forecast with next week's. And if there are still not enough measured days for one of those distances, it is left blank: we do not make the number up.