Clever analogy, but it oversimplifies the problem. There is a precise standard unit for height, but it's impossible to create a precise standard unit for knowledge.

Problem: measurements are noisy, biased, and sometimes vague.

Solution: stop measuring stuff?

Height is a good measurement because it is standardized and objective: I stand on my heels, fully erect, adjacent to a ruler. The marking parallel with the top of my head is my height.

What we should be doing is trying to bring grading to a similar level of objectivity and standardization, not throwing it away completely.

I thought that standardization would be the solution, too, until I realized that that would just make it worse. There are whole industries around getting people higher marks on standardized tests like SAT, LSAT, MCAT, etc.

It really depends what role marks play. If they are used exclusively as a guide to the student of how much they are absorbing, then they could be useful because people would have no incentive to game them.

For a sufficiently well designed test, the easiest route to higher marks would be actually understanding the material. The SAT sucks, and is easy to prep for. The only way to prep for the "standard" (1) calculus exam given at most universities is to actually get a basic understanding of the material.

If people are standing on their toes in a height test (perhaps because Princeton Review told them to), the problem is not the use of standardized inches. The problem is that we aren't forcing people to stand on their heels. And even a bad test isn't useless; most people on their toes are shorter than me on my heels, and the bad height test will reflect this.

(1) Even though there is no official inter-university standard calc test, they are pretty much all the same.

Ok, I agree on that point. As long as the institutions are honest (or dishonest institutions are punished), a test that forces students to stand on their heels is a good solution.