Because when I dig further into how to do that, I often get blanks. Saying you will do that and being able to do it are two different things.
You might not understand how low the bar is here. If I ask 100 applicants to write a working function to calculate E=mc^2 for different values of m, 30+ just won't know where to start.
So when I ask how we might solve a problem I am not asking for the abstract 10000 foot view. I'm checking they can write simple code. Filters out a huge number of people, unfortunately.
By reading papers and experimenting and trying things out? Throw in a few books for good measure?
Why are you answering the question you are asking yourself?