Grading systems are obviously broken. Why isn't anyone fixing them?

I am working on a series of educational apps for iOS and other mobile platforms. Part of this is coming-up with a grading system that makes sense. The objective isn't to simply provide a grade that corresponds to how a student tested. That would simply reflect how someone did on that particular test. The goal is to grade knowledge, not test results.

Having lived and studied in three different cultures in different parts of the world I can say that, from my experience, grading in the US and abroad is pretty much the same. Other parts of the world might use a 0 to 10 scale while US schools like the A through F symbols. In the end the result is the same: You are grading test-taking performance, not knowledge.

Any entrepreneur knows full-well that test performance does not matter. How many times have you failed before succeeding? Failure is a part of learning. Some might argue that failure is the start of learning.

Yet another example is that of a skilled athlete, say, a gymnast. It might take weeks or months of failure to master one particular move that lasts barely past one second. After hundreds of failures the gymnast gets it right and it is part of their skill set. Countless other examples abound.

The problem with test-style grading is that it does not encode acquired knowledge. It encodes isolated and aggregate test scores. And, to add further insult to injury, you get to average test scores at the end of the year, semester or quarter to get your grade. The resulting number is meaningless.

Imagine that we applied this to the aforementioned gymnast. They would get grades below 5 or 6 most of the time due to constant failures while learning. Then, once they figure it out they start getting 10's. Well, if you average hundreds of scores at 5 or below with a few dozen 10's you'll get an average score that does not represent what this person can now actually do. And therein lies the problem.

So, I've been looking at grading systems that allow one to place greater weight on success when compared to failure in order to encode acquired knowledge rather than average test scores.

I'd be interested to learn from others who may have traveled this road and what they may have learned in the process.

An even greater topic is how to go about changing the way schools grade for the benefit of all involved.

Thank you,

-Martin

PM: x@y.z where x = martin y = algoshift z = com