How the AP Score Predictor Works

1. You enter your raw scores

Every AP exam has a multiple-choice section and a set of free-response questions. After an exam — or a practice test — you know roughly how many multiple-choice questions you answered correctly, and you can estimate each free-response score by comparing your work against the College Board's published scoring guidelines.

The calculator form for each subject is generated from that exam's actual structure, so the number of inputs and the maximum points per question match the real exam rather than a generic template. AP Chemistry asks for three long responses and four short ones; AP Psychology asks for two; AP Statistics includes the 15-point Investigative Task as its own field.

2. Your inputs become a composite score

The College Board combines the two sections into a single weighted composite score before converting anything to a 1–5 scale. This calculator does the same arithmetic using each exam's real point distribution — your multiple-choice count plus the weighted sum of your free-response points.

Because the weighting differs by subject, the same raw performance produces different composites on different exams. On AP Psychology the multiple-choice section carries 100 of the available points against 14 from free response; on AP Calculus the free response carries more than half. Each subject page shows its own exam-format table so you can see exactly how your composite was built.

3. The composite is mapped to a 1–5 score

This is the step where honesty matters most. The College Board does not publish the raw-score cutoffs used to convert a composite into an AP score — for any exam, for any year. The cutoffs are set after each administration through a statistical equating process and are not released publicly. We verified this across multiple research passes, including checking the Course and Exam Description documents where such tables would most plausibly appear. They are not there.

What the College Board does publish is the score distribution for each exam: the percentage of students who earned each score from 1 to 5. Those distributions, together with published pass rates, are what we use to calibrate the thresholds for each subject individually. An exam where a smaller share of students reach a 5 gets a correspondingly higher threshold than one where more students do.

This is why the curve differs by subject rather than applying one generic percentage to everything. The resulting table is shown openly on every subject page so you can see the thresholds being used rather than just the output.

4. Where this estimate is least reliable

The largest source of error is usually not our curve — it is the free-response scores you enter. Self-grading is hard, and students consistently give themselves credit for answers that name the right concept without supplying the reasoning a rubric requires. Grading strictly against the official scoring guidelines will change your estimate more than any adjustment we could make.

The estimate is also weakest at the boundaries. If your composite lands within a point or two of a threshold, the difference between a 3 and a 4 can come down to how a single free-response sub-point was awarded, and no calculator can resolve that.

Finally, curves genuinely move between years depending on how the whole cohort performs. An estimate calibrated on historical distributions cannot anticipate an unusually difficult or unusually easy administration.

5. Who builds this and why

This is an independent project, not a College Board product, and it is free to use. Exam formats and curve calibrations are reviewed against published College Board materials, and corrections are welcome — if something looks wrong for a subject you know well, the contact page goes directly to the person who maintains it. More background is on the about page.