# grading on a curve results in what percent of F's?

It depends where you set your cutting points. for example, if you use Z = -2 as your cutting point, the percent below that point, approximately 2.5%, would get F's. However, the cutting point can be arbitrarily set above or below that point. Where is your cutting point?

I hope this helps. Thanks for asking.

That cannot be answered. It depends on the distribution. In small samples, such as an individual class, the distributions can vary widely. And subject matter, as well as the students, affect the distributions.

I tried many years to keep class average at 80 in chemistry, with average students. My "average" number of failures was about ten percent, but that varied greatly year by year. I had semester where no one failed, and semesters where one third failed. Classes vary in time, and in samples. I can say with great confidence that those who ended up as failures did not know the subject matter, whether it was their fault, my fault, family problems, laziness, or the devil's work. If someone is not competent in the subject matter, no grade but F is fair.

The distribution I sought was an average between 80-85, sixteen percent A's, and a normal distribution over the rest. I found that the actual distribution in any class was not that, however, averaging all classes, it closely approximated. But it did vary year by year.
There have been a lot of academic research in this area, if you are interested. Most of it summarized will give you what I gave above.
As PsyDAG pointed out, the actual statistics depend on the parameters of the curving process.