Forecasts deal in generalizations at the macro level for a collection of micro events
. As rebel_heart indicated it is a question of means/averages across a fairly large area.
For example, a 30% chance of rain means that the models show in cases where pressure, humidity, temperature, etc. were similar it rained 3 in 10 times. Or for the optimists - 7 in 10 times it did not rain.
There are times the marina has been worse then the bay and 1 mile was the difference between a tight sail and a bucking bronco. I had a sailing partner run aground in a microburst when it barely sprinkled 2 miles away.
This Analysis of Local Weather reporting
is a good indicator of the issues involved. For those wanting to avoid reading it? Beyond 3 days is a crapshoot, the odds are more a guess than a prediction.
Also, the Butterfly effect does exist. Tens of millions of changes in land, sea, air over a 1000 mile range can have a significant and often unpredictable effect. Think of the air temperature change in Kansas when soil over 100,000 square miles is tilled. That semi-green reflective surface is now brown and holding heat. Rain clouds rise, hill features are overcome, etc.
So should we bother looking at weather? Yes. The data NOAA and other services use are statistical models and while a 30 day outlook is nothing to bet your life on, come Thursday, the weekend is almost always right. (Our brains are wired to remember anomalies - so we often think unexpected things happen more often than they really do)