In French high schools, grading is in the range [0,20], and typically follows a Gaussian distribution centered at 10 or 11 and with standard deviation about 2.
The correspondence with US letter grades is roughly as follows in my opinion:
A+ 16 and above
A- 14 and above
B 12 and above
C 10 and above
Fail below 8 (makeup exam between 8 and 10).
In US high schools, it seems that grading is roughly as follows:
A 93 and above
B 85 and above
C 78 and above
Fail below 70 (D between 70 and 77, whatever D means)
Weird because most of the information is in the neighborhood of 100.
It makes more sense to think of the complementary measure, i.e. the distance to 100:
A 7 and below
B 15 and below
C 22 and below
Fail above 30 (D between 23 and 30).
This mean that in France, exams are design so that the typical student will be able to answer slightly more than half of the questions, while in the US exams are designed so that the typical student will be able to answer most of the questions.
No wonder my students at Brown say that my exams are particularly challenging!
From an information-theoretic point of view, the French system yields the most information.