I have a Bachelor's degree in printing ("Graphic Communication") and one of the classes I took was called "Design with Type." I learned all sorts of fascinating minutia about fonts. One of our assignments was creating a one sheet with a poem or some other text along with an image behind it. Besides the layout challenge, we had to use a system where the kerning was turned off and we had to manually code ALL of the kerning.

So far, I haven't had the patience to take the test GD linked to to see if I retained any of those skills.

Another mostly useless factoid I learnt was that in Postscript (and most proper vector-based typesetting protocols), the details of each character vary ever so slightly depending on the point size; this was demonstrated by outputting the same character (same font obviously) in a 6 point and a 72 point size, then blowing up the 6 point character to the same visual size as the 72 point character to display them side-by-side. When comparing, you would see where the stroke might vary, the transition of a crossbar was more filled in on one, etc.

Another assignment was programming in raw Postscript coding. While searching for a visual illustration of the font size difference, I ran across the image below. I remember creating an image exactly like the one on the left side of this image. EXACTLY!

89. If that's all that kerning involves and it can be done that well with about 8 seconds of effort on each one (I assume that's not actually the case) then typesetters need to stop proclaiming they've accomplished anything all that noteworthy.

93. I completely blew one of them.

80%

The Type Fight sight isn't loading correctly for me so I'm not sure I am correct about what it is doing, but how does one decide which font is better without the context of a specific use?

Most fonts are good fonts for something (except comic sans).

Shape Type, way harder than the Kerning Game. I only got 71%.

I got 71% too!