A worked example: assignment 88% (25%), midterm 74% (35%), final 91% (40%)
Multiplying each score by its weight gives 88 × 25 = 2200, 74 × 35 = 2590, and 91 × 40 = 3640. Those three products sum to 8430, and dividing by the total weight of 100 gives a weighted average of 84.3% — a letter grade of B.
The formula: weighted sum divided by total weight, not by 100
The calculator does not simply assume the weights add up to 100 — it explicitly divides the weighted sum by whatever the weights actually total, so relative weights work exactly like percentages. Scores of 91, 84, and 77 entered at equal weights of 1, 1, and 1 (which sum to only 3, not 100) produce (91×1 + 84×1 + 77×1) ÷ 3 = 84% — the same result as entering those weights as 33.3/33.3/33.3, and still a letter grade of B.
A component weighted at zero: assignment 60% (0%), midterm 72% (40%), final 95% (60%)
Setting the assignment weight to 0 removes it from the average entirely, no matter what score is entered for it. With midterm 72% at 40% weight and final 95% at 60% weight, the result is (72×40 + 95×60) ÷ 100 = 85.8% — a letter grade of B — arrived at as if the zero-weight assignment score didn't exist, which is exactly the intended behavior for a student who has, say, only a midterm and a final in a given course.
Why normalization matters more than hitting exactly 100
Because the calculator divides by the actual total weight rather than assuming it is 100, a syllabus that lists weights as "2 assignments, 3 tests, 5 final" (summing to 10) or as raw point totals rather than percentages still produces the correct overall grade without any manual rescaling — the only number that has to be right is each component's own relative weight, not the sum of all of them.