About This Methodology
USMLE Step 1 transitioned to a Pass/Fail outcome on January 26, 2022 (USMLE official policy change). Since then, predicting Step 1 readiness has meant predicting probability of passing rather than estimating a 3-digit numeric score. Any tool that still reports a Step 1 3-digit score is using outdated calibration.
Our Step 1 predictor is anchored on the National Board of Medical Examiners' (NBME) officially published Comprehensive Basic Science Self-Assessment (CBSSA) pass-probability table, released in their July 2024 Score Interpretation Guidance. That table is calibrated on the full Step 1 examinee population (over 100,000 test-takers per year) using NBME's Item Response Theory (IRT) framework — the same methodology used to set the passing standard.
Why Anchor on NBME Directly
- Population calibration:NBME's table is built on the entire Step 1 examinee cohort (~100,000+/year), not a small self-reported sample.
- Standard methodology: Uses the same Item Response Theory and modified Angoff method that NBME applies to set the official passing score.
- Recency: The July 2024 update reflects current Step 1 content and difficulty (not pre-2022 calibration).
- Transparency: The anchor is a publicly published reference. Any examinee can cross-check the underlying calibration.
- No proprietary model claims:We don't claim to outperform NBME's 100,000-examinee calibration. We surface it cleanly.
Cross-Referenced Research
We validate our calibration against three independent published sources:
- NBME July 2024 CBSSA & CBSE Score Interpretation Guidance (PDF) — primary pass-probability anchor; sampled EPC values 50, 55, 60, 64, 68, 70.
- IAMSE 2024 (International Association of Medical Science Educators) — peer-reviewed study showing CBSE 2-digit score ≥53 corresponds to 96.4% Step 1 pass probability (n=172 examinees, 2022–2023 cohort).
- FSMB 2024 USMLE Performance Data — cohort first-attempt pass rates: US-MD 96%, US-DO 92%, IMG 74%. Used for cohort-specific advice text only; does not alter individual probability.
Why This Approach is Defensible
Many free Step 1 tools build proprietary regression models on small self-reported datasets. The problem with that approach: any dataset under ~50,000 examinees cannot reliably outperform NBME's own calibration, and selection bias in self-reported samples systematically distorts the pass-probability curve in the critical 60–75 EPC band.
By anchoring on NBME's published table and adding interpretation tiers, corroborator handling (Free 120 and UWorld), and cohort-specific guidance, we provide a Step 1 readiness tool that is calibrated correctly for the population actually taking the exam.
Practice Score Correlation with Step 1 Pass/Fail Outcome
The following correlations reflect each practice exam's value as a Step 1 readiness signal. NBME forms 29–33 carry the strongest correlation because they are calibrated on the same item bank as the operational Step 1 exam.
Primary Predictors
NBME Form 31
r = 0.74Notes: The current NBME-recommended form. Most reliable single signal for Step 1 readiness. Take within 1–2 weeks of your exam date for tightest calibration.
NBME Form 32 & 33
r = 0.73Notes: Newest forms, calibrated to current operational item bank. Recommended alongside Form 31 for trajectory tracking.
NBME Form 30 & 29
r = 0.71Notes: Strong predictors, slightly older calibration than 31–33. Useful for early-dedicated benchmarking. Take a newer form closer to exam day to confirm readiness.
Free 120
r = 0.62Notes: Uses retired real Step 1 items. Best taken within 1 week of exam day. Lower precision than NBME forms because of small item count (120 questions), but unique value as a final difficulty check.
UWorld
r = 0.55Notes: UWorld percent correct trends higher than NBME EPC because of repeated exposure and untimed conditions. Our predictor uses UWorld only as a secondary corroborator and to flag sharp divergence from NBME results.
NBME Form-Specific Notes
Current Forms (Recommended for Today's Step 1)
- NBME 33: Newest release. Best calibration to current operational item difficulty.
- NBME 32: Second-newest. Closely mirrors current Step 1 item style.
- NBME 31: Currently the most widely used. NBME recommends taking this 1–2 weeks before exam day.
- NBME 30: Strong predictor. May slightly underpredict for examinees with the most recent content updates.
- NBME 29: Reliable but oldest of the current-pool forms. Pair with a newer form for confirmation.
Legacy Forms (Not Accepted)
- NBME 25–28: Calibrated against pre-2022 Step 1 difficulty. These forms systematically overpredict pass probability by 2–4 EPC points relative to the current operational exam. Our predictor does not accept them to avoid misleading borderline examinees.
- CBSE / COMSAE:Different scale, different item pool. NBME's pass-probability table does not extend to these forms; we do not accept them as inputs.
How Our Step 1 Predictor Compares
| Source | Calibration Anchor | Sample Size | Updated |
|---|---|---|---|
| USMLEPredictor (this site) | NBME July 2024 official table | 100,000+ (NBME population) | June 2026 |
| AMBOSS | Proprietary Qbank model | Not published | Not published |
| Generic Reddit estimates | Self-reported spreadsheet | ~few hundred posts | Inconsistent |
| Q-bank built-in predictors | Internal Q-bank performance | Single-vendor | Varies |
We are not affiliated with the NBME, USMLE, FSMB, or UWorld. Comparison values reflect publicly stated methodology where available.
Step 1 for International Medical Graduates (IMGs)
FSMB 2024 USMLE Performance Data reports the following first-attempt Step 1 pass rates: US-MD 96%, US-DO 92%, IMG 74%. The 22-point cohort gap reflects differences in pre-clinical curriculum exposure, dedicated study window, and clinical context — not differences in individual examinee ability at a given NBME score.
How We Handle IMG Status
- Individual probability is unchanged:If an IMG examinee scores EPC 70 on NBME 31, NBME's table indicates ~98% pass probability — the same as a US-MD examinee at the same score. We do not penalize the individual probability based on cohort.
- Advice text shifts: For borderline tiers, IMG examinees see additional context noting that the IMG cohort has a 74% first-attempt pass rate and that an extra preparation cycle is often a defensible decision when the NBME score is uncertain.
- Selecting IMG is optional: The cohort field affects only advice wording. Leaving it blank produces neutral, cohort-agnostic guidance.
Readiness Tier System
Rather than presenting a raw decimal probability that students misread (a "90% pass" estimate sounds safe even though it implies meaningful fail risk), we classify each prediction into four readiness tiers calibrated to NBME's published pass-probability curve.
Safe to Test
NBME pass probability ≥ 97%. Most examinees at this level pass. Recommended action: confirm timing and continue practice. Trust the data.
Likely Pass
NBME pass probability 90–96%. Strong signal but with measurable downside risk. One additional NBME within 1–2 weeks of exam day builds margin.
Borderline
NBME pass probability 75–89%. Recommended: postpone or extend the dedicated study window. Re-test CBSSA in 2 weeks to confirm direction of progress.
High Risk
NBME pass probability below 75%. Strongly consider rescheduling. Focus on weakest content areas; a 4–6 week extension materially improves outcomes for examinees at this level.
Authoritative Sources & Research Basis
Our methodology and citations:
- NBME CBSSA & CBSE Score Interpretation Guidance — July 2024 (PDF) — primary pass-probability anchor.
- IAMSE 2024 — Predictive Value of NBME CBSE for Step 1 P/F — peer-reviewed external validation (n=172).
- FSMB — USMLE Performance Data — annual cohort pass-rate statistics.
- USMLE Bulletin of Information — official policy on Step 1 Pass/Fail transition (effective January 26, 2022).
- PubMed — Step 1 prediction research — supporting literature on NBME-to-Step 1 predictive validity.
Frequently Asked Questions
How accurate is the Step 1 pass-probability predictor?
The predictor is anchored on NBME's officially published CBSSA pass-probability table (July 2024 update), calibrated on the full Step 1 examinee population (over 100,000 examinees per year). This is the most reliable calibration available outside the NBME itself. Individual probabilities for examinees with recent NBME scores closely match NBME's published values.
Why does the predictor only accept NBME forms 29–33?
NBME forms 25–28 were calibrated before the January 2022 Pass/Fail transition and reflect pre-2022 exam difficulty. They systematically overpredict pass probability by 2–4 EPC points relative to the current operational Step 1 exam. Forms 29–33 are calibrated to the current item bank and provide reliable estimates. We exclude older forms to avoid misleading borderline examinees.
Will you provide a 3-digit score for Step 1?
No. Step 1 has been Pass/Fail since January 26, 2022 (per USMLE official policy). The NBME no longer produces a 3-digit numeric score for Step 1, and any third-party tool claiming to predict one is using outdated pre-2022 calibration. Our predictor reports pass probability, which is the only outcome NBME currently calibrates and publishes.
What is considered a "safe" NBME score for Step 1?
NBME's July 2024 published data indicates an Equated Percent Correct (EPC) of 68 or higher corresponds to approximately 97% pass probability. EPC 64–67 falls in the "likely pass" range (92–96%). Below EPC 56 the data suggests substantial fail risk and we recommend additional preparation before testing. These thresholds are calibrated to NBME forms 29–33; older forms run 2–4 points higher than equivalent EPC on current forms.
How are IMG candidates handled?
Individual pass probability at a given NBME score is similar across cohorts — if an IMG examinee scores EPC 70, NBME's table indicates ~98% pass probability for that individual. However, FSMB 2024 data shows the IMG first-attempt pass rate (~74%) is materially lower than US-MD (~96%), driven by cohort-level factors. When IMG status is selected, advice text reflects the cohort base rate without changing the individual NBME-anchored probability.
Does UWorld first-pass percentage matter?
UWorld percent correct is a useful corroborator but tends to overpredict pass probability by 3–8 points compared to NBME-anchored estimates, primarily because of repeated exposure and untimed conditions. Our predictor treats UWorld as a secondary signal that can lower a tier when it sharply diverges from your NBME (suggesting a knowledge gap), but cannot raise the tier above what NBME indicates.
How does the predictor handle multiple NBME scores?
When two or more NBME forms are entered, we use the most recent form (highest form number) as the primary anchor and compute a trend across the others. An improving trend (≥2 EPC points per cycle) is treated as a positive indicator; a declining trend triggers a recommendation to investigate what changed before test day.
What if my real Step 1 outcome differed from the prediction?
Pass-probability is exactly that — a probability, not a guarantee. At NBME EPC 72 the data shows ~98% pass, meaning roughly 2 in 100 examinees at that level still fail (often due to test-day variance, illness, or anxiety rather than content knowledge). If your outcome differed, your experience is still valuable — please consider submitting your real result through our contribution form to help refine future advice text.
Disclaimer
USMLEPredictor.com is an independent educational tool. We are not affiliated with the National Board of Medical Examiners (NBME), the Federation of State Medical Boards (FSMB), the United States Medical Licensing Examination (USMLE), or any Q-bank vendor. Pass probability is an estimate based on publicly-published NBME calibration data; individual outcomes vary based on test-day performance, content preparation depth, and other factors. Use this tool as one data point in your preparation strategy alongside official NBME self-assessments.