How We Turn a College's Published Early Decision Rate Into an Estimate for You

When you pick a college on our site and enter a test score, you get a line like "Early Decision 12.8% · Regular 3.4%." People reasonably ask what that is. The short answer is that your early decision odds, as we show them, are two public numbers from the college's own filing, taken down by one factor if your score sits under its range. There is no hidden model. This post walks through the whole calculation so you can decide how much weight to give it.

The two numbers from the filing

Every college that publishes a Common Data Set reports, in section C1, how many students applied and how many were admitted, and in section C21, how many applied Early Decision and how many of those were admitted. From those four counts we get two rates:

• Early Decision rate: early admits divided by early applicants.
• Rate for everyone else: all admits minus early admits, divided by all applicants minus early applicants.

For Brown, from its 2025–26 filing, those are 17.9 percent (907 of 5,062) and 4.8 percent. Nothing is estimated yet. These are the counts the college itself reported.

The one factor we add

Colleges also report the middle 50 percent of test scores among enrolled first-years, the 25th and 75th percentiles: in section C9 of the same filing, and to the federal government, which publishes it in the College Scorecard. We read it from the Scorecard. For Brown's SAT that range is 1510–1580. By definition, a quarter of the enrolled class scored below that range, half inside it, and a quarter above.

Applicants are not distributed that way. We assume the applicant pool splits 35 percent below the range, 45 percent inside it and 20 percent above, which is our one assumption, and we state it here so you can disagree with it. Dividing the enrolled shares by the applicant shares gives three factors, and we cap them at 1:

Why cap them? The rate a college reports counts every admit, including students admitted for reasons a test score can't show: recruited athletes, legacy applicants and others the filing doesn't separate. A score inside or above the range doesn't make you one of them, so we never print an estimate above the rate the college reported. That makes the estimates deliberately conservative: weighted by the applicant shares, they come to a little less than the published rate, not all of it. Any single estimate is also capped at 90 percent, because no filing supports a near-certainty.

The Brown example, worked

A student with a 1480 is below Brown's 1510–1580 range, so the factor is 0.71.

• Everyone-else rate 4.8% × 0.71 = 3.4% if applying Regular Decision.
• Early Decision rate 17.9% × 0.71 = 12.8% if applying Early Decision.

That is the whole "Early Decision 12.8% · Regular 3.4%" line. The distance between the two is what applying early is worth to that student at that college, as far as the college's own numbers can say.

A student with a 1550 would be inside the range, and a 1590 above it. Both get Brown's own rates as they stand: 4.8 percent regular and 17.9 percent early. Those are the only two answers the estimate can give for Brown. It is deliberately coarse.

What it ignores

Everything that is not a test score. Grades, course rigor, essays, recommendations, the major you apply to, and the connections that concentrate inside early pools, recruited athletes and legacy applicants among them. The filing does not separate any of those, so neither can we. The estimate describes how the college treated people in your score band last year, not how it will read your application.

It also inherits every limit of the filing. If a college's early pool is 396 applicants, as Vermont's was, one extra admit moves the rate a quarter of a point. If a college does not publish early counts at all, as Michigan and USC do not, there is no estimate, and the page says so rather than guessing.

Why two numbers and not one

The two numbers are not an error margin. One is the regular round and the other is the early round, for the same student at the same college. The gap between them is the early round's value to you, in points. A small gap means the early round changes little there; a large one means it changes a lot. That is why we show both rather than collapsing them into one figure that would hide the only decision the estimate can help with.

Where every number comes from

Two sections of the college's own Common Data Set, C1 for all applicants and admits and C21 for early applicants and admits, and the score range the college reports to the federal government, from the College Scorecard. Nothing else. We print the filing year next to the figures because a 2024–25 filing describes students who heard back in December 2024. We do not use grade point averages, we do not score essays, and we do not run a chance-me model on top of the filing. If the college has not published an early count, the estimate is not shown.

How to use a number like this

Not as a forecast. As a comparison. The estimate is built the same way for every college, so when it says a 1480 sits below the range at one college and inside it at another, that difference is real and comes from the filings. When it says the early round is worth about 9 points at Brown and about 17 points at Tufts for the same below-range student, that difference is real too. Use it to see which of your schools the early round moves most, and whether your score is the thing holding you back at any of them.

The absolute number is the college's own rate, or that rate taken down for a score under its range, and it can't see the rest of your application. The differences between colleges are the useful part.

Start with the school you're leaning toward: what it published, and your estimate there, are free. The full answer for your whole list is the paid document.