Tuesdays are not invented. Probability of precipitation, on the other hand, is one of the most widely misread numbers in public communication, and almost nobody is told what it means.
The U.S. National Weather Service defines probability of precipitation as the chance that a measurable amount of precipitation — at least one hundredth of an inch — will fall at any given point in the forecast area during the forecast period. “Any given point” is the load-bearing phrase. The number is about your location, not about duration and not about coverage.
That rules out the two most common readings. A 70% chance of rain does not mean it will rain for seventy per cent of the day. It also does not mean rain will fall over seventy per cent of the region. It means that if you are standing somewhere in that area during that window, there is roughly a seven-in-ten chance measurable rain reaches you.
A forecaster who says 70% and sees no rain has not necessarily failed. A forecaster who says 70% a hundred times and sees rain twenty times has.
There is a second layer that explains why forecasters sometimes seem to hedge oddly. The figure combines the forecaster’s confidence that precipitation will occur somewhere in the area with the expected fraction of the area it would cover. High confidence in scattered showers and moderate confidence in widespread rain can produce the same headline number by different routes, which is why two days with identical percentages can feel completely different.
Understanding this changes what counts as being wrong. A forecaster who says 70% and sees no rain has not necessarily failed; three times in ten, that is exactly what 70% predicts. What would constitute failure is a forecaster who says 70% a hundred times and sees rain twenty times. That property is called calibration, and it is measurable. Meteorology is unusual among forecasting professions in that it scores itself against outcomes systematically and publishes the results — forecast verification is a formal discipline with international guidance behind it.
The machinery underneath has also changed in a way the public wording has not caught up with. Modern forecasts are largely produced by ensembles: the same model run many times from slightly different starting conditions, because small differences in initial state grow rapidly in a chaotic atmosphere. The spread across those runs is itself information. When the ensemble members agree, confidence is high; when they diverge, the forecast is genuinely uncertain, and the percentage is reporting that honestly rather than evasively.
This is why the standing complaint that forecasters are always wrong is difficult to sustain. Multi-day forecast accuracy has improved steadily over decades, to the point where a modern several-day outlook is comparable to a much shorter-range forecast from a generation ago. What has not improved is the interface. A single percentage compresses confidence, coverage and timing into one number, and then a graphic renders it as a cartoon cloud.
So the practical upgrade is small. Read the percentage as being about you, at a point, over a window. Treat a low number as real information rather than an all-clear. And when it matters, look at the hourly breakdown rather than the daily summary, because that is where the timing the number cannot express actually lives.