I don't see why 48% would imply an accuracy of less than ±1%. When people make predictions or forecasts, they typically don't include error margins when talking to the public. For example, the US 2010 deficit forecast is $1.17 trillion. I have no idea where one could find the margin of error for that forecast.