My way of thinking is to think of a globe and unzip from the south and north pole all the way to the equator. This needs to happen at nearly infinite points to turn it into an undistorted 2D image. Because the poles start on opposite ends of the world with a distinct point, the further away you get from them, the more these sections grow as they near the equator and shrink as they get closer to ther opposite pole. That’s how with limited lines, “sharp ovals” are created from unzipping this globe. The issue is that you will have massive gaps at the poles because you turned a single point into the unrolled width of the equator. A 2D map attempts to distort the poles into a map with no gaps but must still compensate for the massive expansion in areas near the equator. The very common world atlas with a mostly ovular shape with the widest side following the equater is one attempt of showing the distortion. The many ways of compensating for this unzipping and then removing blank spaces is to distort different latitudes in different ways. It can easily result in over representing landmass near poles which consequently underrepresents land near the equator. The best way to combat this distortion (in my experience) is to limit the focus to such a small area that turning it flat leads to negligible curve affect. That won’t work for a globe so go 3D unfortunately. I’m not an expert, but I do have upper education in a related field that works with 3D models.