Well, what do you mean, the "accuracy" of political borders? Again, they're all inaccurate, just in different ways. Are you optimizing for length of border, or curvature, or orientation? Better nearer or further from the pole? Arguably there are far more borders nearer the equator than far from it, Africa has hundreds. And the Mercator projection does pretty much the best job at representing political borders length and orientation within Africa than any other.
I’m thinking about accuracy when looking at google maps.
So do you mean that the length and orientation of the borders is the least distorted with the Mercator projection?
I’m actually not sure how a different projection would distort a country’s border. But for example I’m not putting a ruler against my screen and I don’t know if I need to know if a border is oriented North or not.
Mostly my point is, are we rendering maps for real users’ use cases or backwards rationalizing the status quo? Besides the level of effort that would be involved in changing it, can’t think of any real defense of the use Mercator projection itself in its current context.
Some other comments references street maps that are grids, so a non-mercator projection would bend a grid of streets- this makes sense- but again, the main critique is around the relative size of countries- i.e. the look of the map tiles only when zoomed out sufficiently far. Maybe this is the main UX hurdle- in google maps to change the projection above some level of zoom.
