How we calculate chances, and how accurate they are
School Match simulates the city’s admissions process from the DOE’s public data. We fitted the model on one admissions year, then tested it on the next year without changing it. On that test it stayed within 5 percentage points of all three citywide results the DOE publishes.
What data do you use?
Only public data from the NYC Department of Education: seats, applicants, admissions methods and priorities from the DOE directory, published citywide match results, and school reports for SAT, graduation and demographics. Admission requirements come from school pages on MySchools and school websites. We do not use private surveys.
How does the estimate work?
For each program we look at seats, applicants, its admissions method and your priority group, then simulate the match for an unknown or entered random number. That produces a chance and a range. The range matters as much as the number: it shows how much the answer could move given what is unknown.
How accurate is the model?
We fitted four parameters on Fall 2023 data and tested on Fall 2024, which the model had not seen. Before the test we wrote down a limit of 5 percentage points against each DOE result.
- Got their first choice: model 46.2%, DOE 51.0%, a gap of 4.8 points.
- Got one of their top three: model 79.3%, DOE 77.0%, a gap of 2.3 points.
- Got any choice on their list: model 97.6%, DOE 96.0%, a gap of 1.6 points.
- Across 658 programs, chance falls as competition rises (Spearman correlation −0.945).
- None of the 145 programs with fewer than 1 applicant per seat got a chance under 80%, and none of the 38 programs with more than 10 applicants per seat got a chance over 40%.
What does that test not prove?
It checks citywide totals and the ordering of programs. The DOE does not publish individual outcomes, so we cannot measure how often a single family’s estimate came true. Treat each number as a planning estimate, not a forecast about your child.
What can’t the model predict?
- Audition and portfolio results: these depend on the child’s work, so we do not give a percentage.
- Specialized high schools: we compare a practice score with past cutoffs, without forecasting this year’s cutoff.
- Programs with their own essays, interviews or tests: the outcome depends on how the school scores them.
- Changes in demand: seats and applicants come from the last published year and move every year.
Why does the DOE say 700+ programs and you say 943?
We count every program code in the DOE directory file, including 59 District 75 inclusive programs, 49 ASD and ACES programs and 23 transfer school programs that most families will not apply to. The DOE’s “more than 700” refers to high school program options on its own page; it does not publish a breakdown, so the exact difference is our inference.
Are you connected to the DOE?
No. School Match is independent and not affiliated with NYC Public Schools. Offers are made only by the DOE, through MySchools.
Backtest run September 5, 2026.