SAT scoring
Every SAT score you have ever seen started as a count of correct answers and then went through two transformations. Understanding them tells you how many questions you can actually afford to miss, and why your friend's "score calculator" disagrees with yours by twenty points.
FunSAT scores your practice automatically: raw count, scaled section scores, a 400–1600 total, and an approximate percentile — using the published curve you select.
Open the free score calculator →Your raw score is simply the number of questions you answered correctly. One point each, every question weighted the same at this stage, and — critically — no penalty for a wrong answer. The old quarter-point deduction for guessing was removed years ago and has not come back.
The practical consequence: never leave a question blank. A blank and a wrong answer cost exactly the same, and a blind guess among four options is worth a quarter of a point in expectation. On a section where you are running out of time, spending the last fifteen seconds filling in every remaining bubble is free expected points.
Each section's raw score is converted to a scaled score between 200 and 800 using a conversion table built for that specific test form. The two scaled section scores — Reading & Writing, and Math — add together for a total between 400 and 1600.
This conversion is what people loosely call "the curve", and it is the reason scores are comparable across test dates. It is not a curve in the classroom sense: it is not affected by how other students did on your test date. It is set in advance through equating, so that a given scaled score means the same ability regardless of which form you sat.
Curves differ by form, but the shape is consistent. This illustrates a typical Math section:
| Raw score (of 44) | Approximate scaled score |
|---|---|
| 44 | 800 |
| 42 | 770 |
| 40 | 740 |
| 36 | 690 |
| 32 | 640 |
| 28 | 600 |
| 22 | 540 |
| 16 | 480 |
| 10 | 410 |
Illustrative only. The exact table varies by test form, which is why any calculator is an estimate unless it uses your specific form.
Notice how the curve steepens at the top. Between raw 44 and raw 40 you lose 60 points for four questions — 15 points each. Between raw 32 and raw 28 you lose 40 points for four questions — 10 points each. Errors near the top of the scale cost roughly fifty percent more than errors in the middle. This is the mathematical reason why high scorers should prioritise eliminating careless mistakes over attempting the hardest questions.
| Target total | Approximate questions missed (whole test) |
|---|---|
| 1600 | 0, sometimes 1 |
| 1550 | 2–4 |
| 1500 | 4–8 |
| 1400 | 12–16 |
| 1300 | 20–25 |
| 1200 | 28–34 |
| 1100 | 36–42 |
Ranges are wide because the split between sections matters and because curves vary. Still, the headline is useful: a 1400 does not require near-perfection. It tolerates well over a dozen errors.
The digital SAT is section-adaptive. Each section has two modules. Everyone gets the same first module; your performance on it determines whether your second module is the easier or the harder version.
Scoring then accounts for which module you received. A student who takes the harder second module and misses six questions can outscore a student who takes the easier one and misses three, because the questions were not worth the same. Two implications:
| Section score | Approximate percentile |
|---|---|
| 800 | 99th |
| 750 | 98th |
| 700 | 93rd |
| 650 | 85th |
| 600 | 73rd |
| 550 | 58th |
| 500 | 40th |
Colleges see both section scores, not only the total. A balanced 1300 and a lopsided 1300 are read differently, particularly by programs with a quantitative focus. More on reading your score against a college list.
Many colleges superscore: they take your best Reading & Writing from any sitting and your best Math from any sitting and combine them. A student who scores 700/650 and then 660/710 has a 1350 and a 1370 on paper, but a 1410 superscore. Check each college's policy — it is usually on the admissions site under testing — because where superscoring applies, a retake carries almost no downside.
One raw point per correct answer, no penalty for wrong ones. Each section's raw score converts to 200–800 on a curve specific to that test form, and the two add to a 400–1600 total.
No. Answer every question, including blind guesses.
Roughly four to eight across the whole test, depending on the curve and how the misses split between sections.
Each test form has its own conversion table, so calculators built on different published practice tests differ by 10 to 30 points. Any raw-to-scaled estimate is an approximation unless it uses your exact form.
Questions are weighted by difficulty and your second module's difficulty depends on your first module, so two students with identical correct-answer counts can receive different scaled scores.
FunSAT is a free browser-based SAT and ACT prep app: digital SAT and ACT practice tests, unofficial score calculators, a built-in Desmos graphing calculator, college admissions chances from official U.S. Department of Education data, and a scholarship search. Start free practice →