2027 Rank Order List Strategy for IMGs

How the Match actually works

The NRMP algorithm is applicant-proposing. In plain terms: it takes your list and tries to place you in your #1 first. If that program doesn't rank you, or fills with applicants it prefers, the algorithm moves to your #2, then your #3, and so on — only ever moving down your list, never sideways for strategy. A match becomes final only when every applicant's list has been worked through.

That one design fact drives everything below. Because the algorithm always attempts your most-preferred program first and only descends when a program you preferred has already turned you away, there is no way to help yourself by ranking a program you like less above one you like more. Doing so can only match you to the lesser program before your preferred one is ever tried. Ranking your honest preferences is not just the ethical move — for an individual applicant it is the mathematically optimal one. (This is the well-established strategy-proofness of the applicant-proposing algorithm — exact for an individual's list; the couples match below is a special case. NRMP's own instruction is simply to "rank programs in order of your true preference and not where you think you will match.")

The one rule

Two corollaries follow from the same logic:

Rank every program you would genuinely attend — a match is a binding commitment, so the list is also a promise. NRMP: "Rank only those programs where you are able and would be happy to train."
Never rank a program you would not attend, even at the bottom "just in case." If the algorithm reaches it, you are committed.
Don't rank a program where you didn't interview — it won't rank you back, so no match can occur there.

Building the list

Turn your interview record into an ordered list:

Start from your scored notes — the checklists exist so this step is a sort, not a memory test. Weight what actually matters to you: how residents seemed, case mix and volume, the program's track record placing graduates (and, for you, its history with IMGs and visa sponsorship), location and support, and gut fit.
Confirm each program's eligibility before you rank it — visa sponsorship, year-of-graduation limits, USCE and Step requirements. A program that won't sponsor your visa is not a program you can rank, however much you liked it.
Include a spread — rank your reaches, your solid fits, and at least one program you're confident about. NRMP advises including "at least one 'safety' program."
Order by preference, then stop second-guessing. Once it reflects your honest ranking, the list is done.

How long should the list be?

Longer, for almost every IMG. There is no penalty for a long list — in NRMP's words, "a longer ROL in no way affects the chances of matching to choices higher on the ROL." And the pattern in the data is stark:

Matched IMGs rank far more programs

Two different NRMP measures point the same way. In NRMP's 2026 Charting Outcomes (now published as separate U.S.-citizen and non-U.S.-citizen IMG editions), the pattern is stark for Internal Medicine, the largest IMG destination: U.S.-citizen IMGs who matched into IM had ranked a median of 8 contiguous IM programs (mean ~9) against a median of just 1 for those who didn't; non-U.S.-citizen IMGs, a median of 5 (mean ~6) against 2. And by total list length in the 2025 Match, matched U.S. IMGs averaged about 10 ranks against about 3 for the unmatched.

Read this correctly: it is an association, not a lever. More ranks don't cause a match — applicants with more to rank tend to be more competitive and to have completed more interviews (NRMP attributes list length to "the competitiveness of applicants… and the number of interviews completed"). You cannot rank your way in. But given the interviews you have, ranking all the acceptable ones — and no fewer — is the version of your list most likely to match.

What programs say — and the Code of Conduct

Interview season ends with a flurry of warm messages. Read them through the lens of the NRMP Match Codes of Conduct — there are separate codes for programs and for applicants — which bind both sides:

A program may tell you they're enthusiastic. A program may not ask where you'll rank them, or what other programs you're considering — they "must not request an applicant disclose ranking preferences, ranking intentions, or the specialty or locations of other programs."
Programs "must not solicit or require post-interview communication for the purposes of influencing applicants' ranking preferences," and are told to "take great care not to promote misleading communication… about ranking intentions."
You have the mirror-image obligations: don't ask a program how it will rank you, keep your communications honest, and build your list "based on your true preferences."

Letters of intent & thank-yous

These are advising conventions, not NRMP mechanics — and because the algorithm is applicant-proposing, none of them change your own match directly. Their only possible effect is on how a program ranks you.

Letter of intent (LOI): a note telling one program they are your top choice. Send at most one, to your genuine #1, and only if it is true — telling two programs each that they're your "number one" is dishonest and small worlds talk. It is optional, and a strong list doesn't need one.
Thank-you notes: fine to send, entirely optional, and non-binding; many programs state they don't factor them into ranking. Keep them brief and sincere, not strategic.
Second looks: optional. A program may not require one or imply it affects your placement. Go if it will genuinely inform your preference — not to be seen.

The couples match

If you and a partner are matching together, you link your lists and rank pairs of programs — the algorithm considers each pair together and matches you to the most preferred pair where both of you are offered a position. Two practical rules from NRMP: partners must submit the same number of ranks, and you can pair a program with a "No Match" code (placed toward the bottom) so that one partner matches while the other stays unmatched rather than forcing a bad pair. If a couple doesn't match as a couple, the algorithm does not fall back to matching each of you individually — so build the paired list thoughtfully.

Dates & certifying

Registration opens: September 15, 2026; standard deadline January 29, 2027.
Rank order list opens: February 1, 2027.
Rank Order List Certification Deadline: March 3, 2027, 9:00 p.m. ET — the same deadline by which an IMG's ECFMG verification requirements must be met. No changes are possible after this time.
Match Week: March 15-19, 2027. On March 15, 2027 you learn whether you matched; on March 19, 2027 (Match Day) you learn where.

Place the whole sequence against the rest of your season on the IMG application timeline.

Common mistakes

Ranking too few programs. The single clearest divide between matched and unmatched IMGs. Rank every acceptable program you interviewed at.
Ranking by "where I think I'll match" instead of preference. The algorithm protects a true-preference list; a strategically re-sorted one only costs you.
Letting a program's warmth (or silence) reorder your list. Post-interview signals are non-binding.
Ranking a program you wouldn't actually attend — the match is binding.
Missing the certification deadline, or never certifying. An uncertified list is not in the Match.

And if it doesn't fully come together on Match Monday, the supplemental round begins immediately — see The SOAP Playbook, and, for the longer view, Going Unmatched.