So I wouldn't characterize it as a data issue per se, but can see your point about how under certain circumstances the two should match. If it's raining outside right now, and the forecast is for a '50% chance of rain' -- shouldn't the forecast be '100% chance of rain'?
Actually, no it shouldn't... One thing to keep in mind is that the percentage chance of rain is not technically the percentage chance that YOU will get rain, but rather they are predicting that it WILL rain some amount over 50% (or whatever percentage they give) of the forecast area. So raining at your particular time does not indicate it should be 100% chance.
This is actually why people think meteorologists are so incorrect. For instance, in our area, about 80% of the population lives in about 25% of the coverage / forecast area. So when they predict, say, a 40% chance of rain and they actually get it right with 40% of their coverage area receiving precipitation they did a good job. Only thing is if that 40% all was outside of the main population area then 80% of the people think they got it wrong.
Anyway, this might explain some discrepancies as well as MAYBE explain some others -- if an extended forecast, for instance, is a more general prediction over a larger area then the conditions and precipitation could differ. Obviously some WAG, but without knowing their models and such I don't know that we should expect everything to match up.
Adam