How we count
The rules behind each number on this site: who counts, who is left out, and why the students who did not move up stay in every average.
- Source
- Official SAT results sheet
- Measured
- First diagnostic to official SAT
- Years
- 2023, 2025 and 2026
- Platform
- Elmy
- 238
students who count
- 219 of 238
scored higher on the official SAT
- 19
scored the same or lower, kept in every average
From the first diagnostic to the official SAT.
Scroll through the method, or select a stage.
An illustration of the method · not a student’s data
Start from the diagnostic.
Before a single lesson, we know exactly where you stand. You see which skills cost you points, ranked.
Teach the gaps. Check every lesson.
Your hours go exactly where they move your score the most. A short check after each lesson shows what is fixed, and what just slipped. When the data changes, your plan changes with it.
Measure on the official SAT.
Real test conditions, start to finish. The official score is read against the first diagnostic. The distance is the Effect.
Numbers are only worth showing if the counting behind them is plain.
Here is ours.

The source
One record: our official SAT results sheet.
It lists every official SAT result students have sent us since 2023, next to their TestPrep diagnostic where we have it, and the progress between the two.
Every number in these stories is computed from that sheet, and can be checked against it line by line. The sheet was last read on 27 October 2026.
Two moments, nothing between
Only two scores are in the Effect: the first diagnostic, taken on day one, and the official SAT, the score the student sends us.
The checks after each lesson, the halfway test and the full practice tests steer the plan. They are in no average.
- Diagnose
- Analyze
- Plan
- Teach
- Check
- Adapt
- Simulate
01 · The diagnostic. Before a single lesson, we know exactly where you stand. No time wasted on what you already know.
See the full engineThe rules
- A result counts when both scores are on the sheet: a TestPrep diagnostic and an official SAT total on the 400 to 1,600 scale. A diagnostic alone proves nothing, and neither does an official score without a starting point.
- The gain is the sheet's own. Each result's gain is its progress on the sheet: the official score minus the diagnostic beside it. Nothing is added from other records, and nothing on the sheet is left out.
- Nobody is dropped for a low result. The 19 students who scored the same or lower stay in every average. That is why the share who scored higher, 219 of 238, sits next to the gain.
- Gains are per student. The average is the mean of 238 individual gains, with the median beside it.
- Years are kept apart. The record covers 2023, 2025 and 2026, and the report shows each year on its own. The 2024 results on file came without diagnostics, so they are not counted.
What we do not publish
- School-by-school results, without the school's written permission.
- Results by instructor. They stay inside TestPrep.
- The records themselves. They hold names.

The full record, read four ways, is the report: The TestPrep Effect, measured on the SAT.
If a number on this site ever changes, this post will say what moved and why.
About the numbers. 238 students with a TestPrep diagnostic and an official SAT score on record, 2023, 2025 and 2026, from records last updated on 27 October 2026. Gains are averages; the median gain is 150 points. Every result on our official SAT results sheet with both scores counts, the official score against the diagnostic beside it, and the students who did not move up are in every average.
We diagnose before we teach.
Where would
you start?
Start with a free full-length practice test. We read it with you, then build the plan around what it shows.