Learning Cadence editorial, , 6 min readHow we work
You see 40% beside a rain symbol. Should you expect a short shower, a wet afternoon or rain in only part of town? The number alone cannot tell you all three. It describes a probability. Read amount and timing separately.
This guide explains how to read that probability. The forecasts and counts below are invented teaching examples. They illustrate how to interpret a label. Check your local official forecast and warnings for an actual outing.
The short answer
A chance at a place, over a stated time.
In a US National Weather Service point forecast, 40% means a 40% probability of at least 0.01 inch of precipitation at that location during the stated period. It does not say that rain lasts for 40% of the day or covers exactly 40% of the area. Read the location, time period and provider definition together.
The method
Read the label before the percentage.
Keep the event, place and time together. Otherwise two correct numbers can answer different questions.
01
Find the place
Use the forecast location or grid point that matches your destination. A nearby observation station records weather at its own location.
02
Find the period
An hourly probability and an afternoon probability cover different intervals. Check the local time and date, especially when comparing apps.
03
Find the event
Check what the provider counts as measurable precipitation. The NWS uses the 0.01 inch threshold discussed here. Other services may define the event differently.
04
Read amount and timing separately
Look for the rainfall amount, hourly detail and official warnings. Probability alone does not describe intensity, duration or the consequences of rain.
What the forecast actually promises.
The NWS defines this probability at a forecast point for a stated period, with a threshold of 0.01 inch. That is 0.254 millimeters. A 40% forecast leaves both a wet and a dry outcome possible; a dry afternoon does not, by itself, show that the probability was wrong.
For an intuition check, imagine 100 comparable forecasts that each assign a 40% chance to the same kind of event. If those probabilities are well calibrated, about 40 of the events would occur over many such cases. Calibration describes the long-run pattern. The next ten days can produce a different count.
Original example: a point forecast says 40% chance of measurable precipitation between noon and 6 p.m. You cannot turn that into 2.4 hours of rain. Multiplying six hours by 0.40 treats a probability as a fraction of time, which changes the question.
Nor can you infer 40% of a city will get wet. Spatial coverage and confidence can be inputs to a forecasting method, but the displayed point probability is not a map showing exactly which streets will receive rain.
A 90% chance also does not mean heavy rain. A high chance of crossing a small measurement threshold and a forecast of a large rainfall amount are different statements. Read the amount field to understand what quantity is forecast.
Why two hourly percentages do not add up.
Original example: the forecast shows a 40% rain chance for each of two successive hours. Adding 40% and 40% to get an 80% chance across the two hours can count the same wet situation twice.
Imagine 100 equally weighted possible weather scenarios. In one illustration, the same 40 scenarios are wet in both hours. Then 40 out of 100 have rain in at least one hour. In another illustration, 40 scenarios are wet only in the first hour and a different 40 only in the second. Then 80 out of 100 have rain in at least one hour. Both illustrations show 40% for each individual hour.
The missing information is the overlap between the wet scenarios. If the hours were independent, the calculation would be 1 minus (0.60 times 0.60), or 64%. The hourly percentages alone leave that independence assumption unproven. Use the provider's forecast for the full period instead of manufacturing one by adding hours.
Check the service behind your app.
Do not assume every provider uses the same threshold, time interval or display convention. Open the explanation attached to your forecast. An icon, a daily amount range and an hourly probability can summarize different things.
Australia's Bureau of Meteorology distinguishes rain likelihood from possible rainfall amounts and explains that new observations can change a forecast. Its guide also describes how localized rain can differ from a nearby station observation. These limits make checking the location and label especially useful.
Original practice forecast: 40% chance of measurable precipitation at Park A from noon to 6 p.m. For each statement, decide whether the forecast supports it before reading the answers.
1. It will rain for 2.4 hours. 2. Rain above the stated threshold is possible at Park A in that interval, but not certain. 3. Exactly 40% of the park must receive rain.
Answer check: only statement 2 follows. Statement 1 confuses probability with duration. Statement 3 confuses point probability with an exact coverage map. None of the three tells you the rainfall intensity.
Transfer case: a different app shows 30% at 2 p.m. and 50% at 3 p.m. A friend says the two-hour chance is 80%. What would you ask? One reasonable answer is: Do the wet scenarios overlap, and what period does each label cover? Without that information, adding the percentages is not justified.
Use the forecast for the question you have.
For an ordinary walk, you may care about whether to carry rain gear. For a picnic, the timing and rainfall amount may matter more. The same probability can lead to different choices because the cost of getting wet and the available alternatives differ. There is no universal percentage at which everyone should cancel.
A rain percentage is not a substitute for warnings about lightning, flooding or other hazards. Use current official guidance for safety decisions. This article teaches interpretation; it does not assess the weather at your location.
Learn the ideas behind the forecast.
If you want to understand weather or probability beyond one percentage, try Cadence. We make the app. Describe what you want to understand and answer the setup questions to build a course around your starting point.
Adapt the suggested weather goal below to your own question. Cadence builds your course after setup; it does not provide forecasts. Cadence offers lessons and practice at your pace. The exercise here is complete without the app.
Free rain probability worksheet
Check a forecast without changing its meaning.
Use the fictional example or copy only public forecast labels. No location permission or account is needed.
RAIN PROBABILITY WORKSHEET
Learning Cadence | Original exercise | 31 October 2026
FORECAST LABELS
Provider:
Location named in forecast:
Time interval and time zone:
What event counts as rain:
Probability:
Amount and timing fields, separately:
PRACTICE
A fictional NWS point forecast gives a 40% chance from noon to 6 p.m.
Does it promise 2.4 wet hours? Why?
Does it promise exactly 40% of the park gets rain? Why?
What does it actually say?
ANSWER CHECK
No: probability is not duration.
No: point probability is not an exact area map.
A 40% chance of precipitation at or above the provider threshold at the point during the period. Both wet and dry outcomes remain possible.
TRANSFER
Two hours each show 40%. In 100 equally weighted scenarios, the same 40 are wet in both hours. How many have rain in at least one hour?
Answer: 40, not 80. If different groups of 40 were wet in each hour with no overlap, the answer would be 80. Hourly values alone do not tell you the overlap.
Use current official forecasts and warnings for actual plans. This is a learning exercise, not weather advice.
Put it into practice
Understand the probability behind the number
If you want to learn the weather and probability behind a forecast, try Cadence. Describe your goal and answer the setup questions. Cadence is our app.
Not as a general reading of an NWS point forecast. It gives a probability at the forecast point during a stated interval, not a map of exactly which fraction of the area will get wet.
Does 100% mean rain all day?
No. A probability of the event occurring during the period does not specify how long it lasts or how much falls. Read the timing and amount fields.
Can I add the hourly rain chances?
No. Wet outcomes can overlap across hours. The individual percentages do not tell you that overlap. Use the provider's full-period forecast.
Is this a live weather forecast?
No. The examples and worksheet teach probability. Use your local official forecast and warnings for current conditions.
See it in concrete terms
Explore example courses for your goal.
Adapt this starting goal to the weather questions you want to understand. Answer the setup questions in Cadence; this is not a fixed forecast course.
We separated the event, point and interval, then wrote original duration, coverage and overlapping-hour counterexamples with checked answers.
What we checked directly: On October 31, 2026 we read the NWS point-probability definition and Bureau of Meteorology rain guide. We checked 0.01 inch equals 0.254 millimeters, 40 of 100 equals 40%, and the independent two-hour illustration gives 64%.
Limits of this review: We did not test a forecast provider, measure calibration, observe readers or validate learning gains. The 100 scenarios are invented illustrations, not weather data.
