It seems the problem is often mis-stated - there is no calculation problem, the problem is in adequately defining/sampling the initial data. It's maybe similar

It seems the problem is often mis-stated - there is no calculation problem, the problem is in adequately defining/sampling the initial data. It's maybe similar to predicting the weather - we can have all the perfect equations in the world for fluid dynamics and heat flow etc, but until we have system-invisible temperature and humidity sensors for every square millimetre of atmosphere and earth volume, we won't be able to predict the weather very accurately or very far ahead.