A college is not a reach, target or safety on its own; it is one for a specific profile in a specific round. A four-step method using middle-50% ranges, round-level admit rates and program-level review, with worked examples from Duke, Michigan, UVA, MIT and Williams.
Last reviewed September 16, 2026 · By the Chance-Me research team
A college is a reach, target or likely school only for a specific applicant in a specific round, never on its own. Duke is a reach for nearly everyone in Regular Decision, where it admitted 4.8% for the Class of 2029, and a different proposition in Early Decision, where it admitted 13.8%. UVA is a target for many strong Virginia residents (admitted at roughly 22%) and a reach for the same profile from out of state (roughly 10%). Michigan's overall 17.9% admit rate says almost nothing about an applicant to its Ross School of Business, which admitted about 7.1%. The method below classifies a college for you in four steps: academic range, admit rate in your round and pool, program-level review, and the part no number captures.
Short answer. Start with the college's middle-50% test and GPA ranges: below the 25th percentile is a reach regardless of anything else at a selective college. Then use the admit rate for your round and your pool, not the headline. Then check whether your program runs its own review. Any college that admits under 10% of applicants in your round is a reach even for a profile above the 75th percentile. A likely school is one where you are above the 75th percentile and the admit rate in your round is well above 50%; below that, call it a target. Most searchers say safety; Chance-Me says likely, because at selective colleges nothing is safe.
The same 1500 SAT and 3.9 unweighted GPA is at or below the 25th percentile at MIT (middle 50% of 1530 to 1580), inside the range at Duke (1510 to 1570), and above the 75th percentile at Michigan (1370 to 1530). One profile, three positions. A label that ignores the college's range is a guess; a label that ignores the round and the pool is a guess with a decimal point.
Every college reports its admitted or enrolled students' 25th and 75th percentile SAT and ACT scores in Section C of its Common Data Set, and most report GPA. Find yours on the Chance-Me college profiles or the college's own page.
| Your position | At a college admitting under 10% | At a college admitting 10 to 30% | At a college admitting over 50% |
|---|---|---|---|
| Below the 25th percentile | Hard reach | Reach | Target |
| Between the 25th and 75th | Reach | Target | Likely |
| Above the 75th percentile | Reach | Target, sometimes likely with a strong fit | Likely |
Two cautions. Test-optional colleges report ranges only for students who submitted, so the range runs high; Duke and UVA are optional for 2026-27, MIT, Michigan and Stanford require scores. And a middle-50% range is a description of who enrolled, not a cutoff: a quarter of the class was below it.
The headline admit rate blends rounds and pools that are not the same competition.
| College | Headline rate | Rate that applies to you | Source |
|---|---|---|---|
| Duke | 4.8% overall (Class of 2029) | 13.8% in Early Decision; ED fills roughly half the class | Duke profile |
| Northeastern | About 5.2% | About 43% in ED I; RD well below the headline | Northeastern profile |
| UVA | 17% blended | About 22% in-state, about 10% out-of-state; about 28% in ED | UVA profile |
| Michigan | 17.9% overall | Ross about 7.1%; Engineering about 13%; out-of-state estimated 5 to 7% | Michigan profile |
| Williams | 7.4% combined (Class of 2030) | 6.3% in RD; about 25% in ED | Williams profile |
Reclassify after this step. A college that is a reach at the headline rate can be a target in a binding round for an applicant above its range, and a target at the blended rate can be a reach in the out-of-state pool.
Many universities admit by school or program, each with its own committee and rate. Michigan admits to ten first-year units. Duke asks every applicant to choose Trinity or Pratt. Northeastern's Khoury College of Computer Sciences is estimated at roughly 3 to 4%. UVA's McIntire School of Commerce does not admit first-years at all; students apply internally after two years. If your program is more selective than the university, classify against the program. If it is a second-step admission, the university read comes first and the program compounds on top of it.
Above the range and inside the rate, the deciding factors are the ones no table holds: what the college says it values, how distinct your profile is from its qualified pool, and the written parts of the file. Two habits help. Read the college's own statement of what it looks for, which each Chance-Me profile summarizes, and ask whether your record shows it or merely claims it. And treat a well-fitted profile as moving one band at most: a strong fit turns a reach into a plausible reach, not into a target.
There is no fixed formula, and Chance-Me does not publish one. Two structural facts matter more than a count. First, at least one college on the list should be a likely school you would attend, with a cost you have checked on its net price calculator. Second, an early round at a first-choice college is the single largest lever most applicants control, and it is only available once. Chance-Me's free Make My List builds a first draft across 150 colleges from a short profile; the free read then bands 2 of them, and the Profile Audit bands 6 with a named gap per college.
Beyond the numbers. Every college reads an application on four questions: can you do the work (academics in context), why you (what makes you distinctive), why here (fit with this specific college, program and major), and why now (what changed this cycle in testing, essays, policies and priorities). Most calculators score the first question. Chance-Me scores all four, with weights that differ by college and adjust to each student's profile, in a mathematical model validated on real applicant cases. Behind it is a structured knowledge base for each of 150 colleges that combines Common Data Sets, admissions-officer profiles and public statements, speeches and podcast interviews, high school profiles, published research, real essays and observed rejection patterns. No other tool we know of combines these sources into one per-college model. The method is described on the methodology page.
Chance-Me's four bands are Hard reach, Reach, Target and Likely. Hard reach means the academic profile is below the college's typical qualified pool. The other three are set relative to the college's admit rate: Likely requires a probability well above the college's rate with an absolute floor so that a strong ratio at a 4%-admit college is not mislabeled; Target requires a probability around the rate or above; Reach is below that. Every band comes with a probability and a confidence score, and confidence rises as you add more of your profile. The glossary defines the terms; the methodology page describes the model.
Chance-Me's view: likely is the right word and safety is the wrong one for any college that turns away most applicants. Most applicants in the likely band get in; not all do.
Try it: Get your own bands for 2 of the 150 colleges Chance-Me tracks. About 2 minutes, no account or card. Start your free read.
A label only means something relative to the college's admit rate, not as a fixed percentage. Chance-Me.ai sets its bands against each college's own rate: a 30% probability at a 4%-admit college reads as strong, while the same 30% at a 60%-admit college reads as a reach. Any tool that uses the same cut-offs for every college will mislabel selective ones.
At a college that admits 4% of applicants, a well-calibrated 20% is a strong position; at a college that admits 50%, it is a hard reach. Chance-Me.ai shows the probability next to the admit rate for the round you are applying in, which is the comparison that matters.
No. Admit rates differ by round, by major or program and by applicant pool, and your own profile moves the answer. Chance-Me.ai's college profiles list admit rates by round and what each college says it values, and its free read places your profile against 2 of those colleges with no account or card.
Chance-Me.ai is a college admissions chances tool that reads a student's full profile against each of 150 selective US colleges and returns a probability, a separate confidence score and the one input that would most change the read. It models the four questions every college asks (can you do the work, why you, why here, why now) with weights set per college, calibrated against published admissions data: round-level admit rates, middle-50% ranges and each college's own stated priorities. The first read for 2 colleges is free with no account or card; a Profile Audit is $1.99 per college and the Application Playbook, which reviews your transcript and essays, is $30 per college.
Classify each college against its own admit rate and your profile, not against a national rule of thumb, then make sure the list has real likely schools you would attend. Chance-Me.ai's free Make My List tool builds a balanced list across reach, target and likely bands from its 150 colleges, and the free read checks any 2 of them.
Every college values different qualities. See how your profile aligns, free, then go as deep as you want.
A calibrated read of your profile for 2 schools. About 2 minutes. No account, no card.
Start your free readThe Audit applies the full knowledge base to your profile: strengths, gaps, and an action plan for each college you pick. The Application Playbook, $30 per college, adds a read of your essays against what that college values.
Get your Profile AuditInsights and strategy for your customized list of up to 15 colleges, all season. Fresh insights every week. First month free.
Explore Admit-MeReviewed by the Chance-Me research team. Competitor facts are taken from each company's own published pages on the review date and may change. Chance-Me is independent and not affiliated with any college or company named on this page.