How Percentile actually works.

Everything starts with a full-length, official-style diagnostic test. It takes about two hours, it is free, and it is honest: you get a predicted score with a realistic range, not a flattering number designed to sell you something.

More important than the score is the map underneath it. The SAT tests a few dozen distinct concepts, from comma splices to quadratic systems. The diagnostic estimates your mastery of each one individually, so day one of studying already knows the difference between "weak at algebra" and "weak at exponent rules under time pressure."

Percentile turns the diagnostic into a day-by-day plan based on three things: your test date, your target score, and your actual weaknesses. A student missing easy grammar points gets a very different plan from one missing hard geometry questions, even if their total scores are identical.

The plan favors whatever closes your score gap fastest. Concepts that appear often on the test and that you miss often in practice float to the top. And the plan is not a contract: it rebuilds itself after every session based on what you actually got right and wrong.

Each day you get a short, focused session: new concepts taught with full lessons, practice questions matched to your level, and reviews of older material timed to arrive just before you would forget it.

That timing is spaced repetition, and it is the most reliably proven technique in learning science. Your brain holds on to information it has to retrieve repeatedly, at growing intervals. Percentile tracks a memory schedule for every concept you have studied and asks you to recall each one right around the moment it is about to slip. Easy concepts come back rarely; slippery ones come back often. You never grind what you already know, and you never silently lose what you learned three weeks ago.

When you miss a question, the AI tutor explains why, in plain language, as many times and as many ways as you need. It has seen every mistake you have made on the platform, so its explanations start from how you think.

Concept drills build skills; full tests build scores. On a regular cadence the plan schedules complete, timed, adaptive practice tests that mirror the real digital SAT's structure, including the way the second module gets harder or easier based on your first-module performance.

After each test you get the same concept-level breakdown the diagnostic gave you, plus an updated score prediction. Across several tests you see a trend line, not a guess, and pacing problems show up in the data before they show up on test day.

There is a difference between knowing an answer and being sure of it. Percentile periodically asks how confident you were, and compares that against whether you were right. Overconfident on comma rules? You will see it, with numbers. Underconfident on geometry you actually know? You will see that too, and stop second-guessing correct answers on test day.

This calibration feeds the score prediction you see on your dashboard, which is why it comes with an honest range instead of a single number. When the range sits where you need it, you are ready, and you know you are ready. That is the whole point: walking into the test with no surprises left.