The 90% Question: How Many Interviews Is "Enough"?
Meet Maya. She is a U.S. MD senior, Step 2 in the low 250s, solid derm research, and by mid-December she is sitting on seven dermatology interviews. Down the hall, her classmate applying to orthopedic surgery has twelve invitations and has stopped checking his email entirely. Maya is still refreshing hers twice an hour, wondering whether she should keep grinding for an eighth, ninth, tenth invite, or whether seven is quietly enough.
Let me break this down specifically, because Maya's question is the wrong one.
The metric that predicts your Match probability is not interview invitations. It is not even interviews attended. It is contiguous ranks, the number of programs you actually place on your certified rank list within a single specialty. The NRMP's probability curves are built on that number. Invitations are raw material. Your rank list is the finished product, and only the finished product gets charted.
Here is the punchline, and I will defend it with data in the next section: for U.S. MD seniors, dermatology, orthopedic surgery, and otolaryngology all converge on roughly the same answer. Ten to twelve contiguous ranks puts you in the 90% match probability zone. Derm sits at the top of that range, ortho and ENT at the bottom, and the difference between the three is far smaller than the anxiety on interview trail suggests.
Everything below draws on NRMP Charting Outcomes data for U.S. MD seniors, and I will flag where the curves shift for DO and IMG applicants, because they do shift, and pretending otherwise does nobody any favors.
The Raw Data: Where the 90% Threshold Actually Sits
The NRMP publishes match probability as a function of contiguous ranks, specialty by specialty. If you have never stared at these curves, they are the single most important charts in your application cycle. More important than Step percentiles. More important than how many aways you squeezed in.
Here is how the curves actually behave across all three specialties:
- At 4-5 contiguous ranks: probability hovers in the 55-65% range. Coin flip with slightly weighted odds.
- At 7-8 ranks: you climb into the high 70s to mid-80s. This is where Maya sits, and it is why her anxiety is rational.
- At 9-11 ranks: the curve crosses 90%. This is the practical threshold zone.
- At 12-14 ranks: the curve flattens into the mid-90s and stops rewarding you meaningfully.
The specialty-specific thresholds are approximately 11 for dermatology, 10 for orthopedic surgery, and 10 for otolaryngology. Read that again. The three most feared applications in the Match are separated by roughly one rank list entry. The terror differential between derm and ENT is manufactured almost entirely upstream of interviews, at the screening stage, not at the rank list stage.
Structural context explains why. Dermatology has roughly 140 programs and 500 positions. Orthopedic surgery runs about 170 programs with 850 positions. Otolaryngology is the smallest at around 120 programs and 330 positions. Program count determines how many total interview slots exist in the ecosystem, which determines how hard it is to get the interviews. But once you hold the invitations, the probability math per contiguous rank is strikingly similar across all three.
Now the part students consistently get wrong: invite-to-rank attrition. If you attend ten interviews but certify only seven programs, because two were malignant, one had a resident exodus you discovered at the pre-interview dinner, and one was in a city you realized you cannot survive, then you are operating at seven ranks. Not ten. I have watched applicants do this arithmetic wrong in February and learn what it cost them in March. Every program you strike from your list moves you left on the probability curve. The chart does not care why you deleted it.
This is why the working target is twelve interviews attended: because not every interview survives contact with reality and becomes a rank.
For a deeper look at how program directors evaluate applicants, see how program directors rank applicants.
Why Your Number Is Not the Same as Your Classmate's: Variables That Shift the Curve
The 10-12 figure is the curve for the median U.S. MD senior. Your personal curve slides left or right depending on four variables, and you should calibrate honestly.
Step 2 and AOA. An applicant with a 260+ and AOA membership can live near the low end of the range, 10 ranks behaves more like 11 or 12 for them, because programs rank them higher and their per-rank conversion is better. An applicant in the 240-250 band without AOA should push toward 12 or beyond. Same curve, different position on it. This is not unfair; it is simply how rank-order lists aggregate.
DO and IMG status. This is where I stop being diplomatic. The published 90% thresholds are U.S. MD senior numbers. DO seniors and international medical graduates face lower match probabilities at every rank count in these three specialties, and visa needs compress it further. If that is you, add two to four ranks to the target and give serious thought to a parallel application in a less competitive specialty. Fourteen ranks as a DO applicant in ortho is not overkill. It is arithmetic.
Geographic restriction, the silent killer. I have seen this torpedo more applicants than any board score. "I am only ranking programs on the West Coast" sounds like a lifestyle choice. Mathematically, it is a rank count reduction. If three of your twelve interviews are outside your acceptable geography, your effective list is nine. Applicants routinely believe they hold twelve ranks when their actual, livable list is eight. Audit the list you would genuinely certify, not the one you attended.
Post-interview communication. Letters of intent and thank-you notes do not replace rank count, and anyone selling you a "one strong LOI locked my match" story is describing survivorship bias. But at the margins, a well-timed, honest letter to a borderline program can move you from their rank list's dead zone into the range where they actually reach you. Think of communication as improving your conversion rate per rank, not as adding ranks.
Interview Strategy by Specialty: Where to Spend Your Extra Invites
The conversion pipeline looks like this, and every specialty stresses a different joint in it:
Dermatology. Rank broadly, research-heavy academic departments and lesser-known community programs alike. Derm's threshold sits at ~11 because the field is small and every program's list behaves a little idiosyncratically. If your Step 2 is below 260, that twelfth interview at the program nobody on Reddit is discussing can be the one that converts. Breadth is the strategy. Prestige-chasing is not.
Orthopedic surgery. Ortho runs on away rotations and formal preference signals. Your audition grades effectively functioned as your first two interviews. If you have reached ten or more genuine interviews, stop collecting and redirect energy into performance and rank list positioning, an eleventh invite you attend unprepared adds less than converting the ten you have. Ortho applicants hoard interviews like airline miles. It is a waste of preparation bandwidth.
Otolaryngology. The smallest field of the three, which makes every interview high-stakes, there are fewer total slots to replace a blown one. Nine to eleven ranks can still reach 90%, but if any part of your file is borderline, target twelve and treat every audition and interview day as decisive.
Run every invitation through the checklist: fit screen → signal priority → interview performance → contiguous rank. An interview that fails the fit screen is a likely deletion in February. Either prepare hard enough to convert it, or decline it early and let the waitlist move.
Signals, Tokens, and the "True" Interview Count
Dermatology and orthopedic surgery now use formal preference signals through ERAS, and the correct use of them is badly misunderstood. Spend tokens on programs where you have genuine geographic, research, or mission fit, places that will look at your file and see a plausible match. Burning signals on dream reaches where your file is a standard deviation below their median is ceremonial. It feels good and converts nothing.
Otolaryngology's signaling culture has historically been less standardized; away rotations and letters of interest have functioned as the informal signal layer. Policies shift between cycles, so verify the current ERAS rules for your application year rather than trusting a blog post, including, potentially, this one.
Two closing principles on count management:
- Diminishing returns are real. After 12-14 ranked programs, each additional interview adds almost no probability and costs preparation cycles that degrade your performance at the programs that matter. Fifteen interviews is not a flex. It is usually a scheduling error.
- Deleting is allowed, but count first. Declining to rank a toxic program is a legitimate, sometimes wise decision. Just know your exact resulting rank count before you hit delete. Dropping from twelve to eleven is a rounding error. Dropping from nine to eight is stepping off the 90% curve, and you should make that trade with open eyes.
Summary: The Niche Breakdown for Your Rank List
The bottom line, condensed:
- For U.S. MD seniors in dermatology, orthopedic surgery, and otolaryngology, 10-12 contiguous ranks lands you in the ~90% match probability zone, derm at the top of the range, ortho and ENT at the bottom.
- Interviews are not ranks. Because invitations decay through attrition, bad fits, and February deletions, target at least twelve interview experiences to protect a double-digit final list.
- Know your multiplier. Lower Step 2, no AOA, geographic restriction, DO/IMG status, or visa needs each push your safe target toward 13-15 ranks, or justify a parallel plan.
- After fourteen ranks, marginal returns flatten. Redirect that energy into signals, audition performance, and rank list precision.
Maya, from the opening, should keep grinding for interviews eight through ten. Her ortho classmate should stop collecting and start preparing. Both of them should build their final lists from NRMP Charting Outcomes data for their specific applicant type, not from group chat folklore. Interview FOMO is expensive. Rank list precision is free.