H-1B lottery odds depended on how many times you were registered
The H-1B lottery's yearly odds are well documented: about a 25% selection rate in FY2024, a number USCIS publishes and every law-firm tracker and odds calculator repeats. It's also widely understood that, since the FY2025 beneficiary-centric rule, registering the same person through many employers no longer helps — everyone now gets one entry. (The market design of these lottery reforms is studied in NBER w26767.)
What that single per-year number hides is the thing that actually set your odds before FY2025: how many times you were registered. Condition on that and the one "lottery rate" splits in two — roughly 25% for a single registration vs ~70% for the average multi-registered beneficiary, a 2.8× advantage the rule then erased. The conditioning below is built entirely from USCIS's own published totals; the arithmetic is ours, the inputs are theirs.
Layer 1 — the per-registration selection rate
Selections ÷ eligible registrations, from USCIS:
| Cap FY | Eligible registrations | Selected | Per-registration rate |
|---|---|---|---|
| 2021 | 269,424 | 124,415 | 46.2% |
| 2022 | 301,447 | 131,924 | 43.8% |
| 2023 | 474,421 | 127,600 | 26.9% |
| 2024 | 758,994 | 188,400 | 24.8% |
| 2025 | 470,342 | ~135,137 | ~28.7% per reg (28.9% per beneficiary) |
| 2026 | 343,981 | 120,141 | ~34.9% |
Caveat that rides this table: FY2021–22 rates are inflated because USCIS ran multiple selection rounds when many selectees never filed. "Selected" is not "got an H-1B."
Layer 2 — odds per beneficiary, by registration count (pre-FY25)
Selection was uniform across registrations, so a beneficiary with k registrations had roughly 1 − (1 − p)k odds:
| Cap FY | p (per reg) | k = 1 | k = 2 | k = 3 | k = 5 |
|---|---|---|---|---|---|
| 2023 | 26.9% | 26.9% | 46.6% | 61.0% | 79.2% |
| 2024 | 24.8% | 24.8% | 43.5% | 57.5% | 76.0% |
FY2024 concretely: the average multi-registered beneficiary held ~4.3 registrations (408,891 multi-reg registrations ÷ ~95,897 multi-reg beneficiaries, using USCIS's ~446,000 unique-beneficiary figure), for ~70% odds vs 24.8% for a single registration — a 2.8× advantage. That asymmetry is exactly what the FY2025 rule removed: every beneficiary now has identical odds regardless of employer count — 28.9% in FY25, ~35% in FY26. The odds rose because the pool got smaller, but that shrinkage isn't all the counting rule: the FY2026 registration fee jumped from $10 to $215 and the tech-hiring market cooled, both of which cut registrations independently of the beneficiary-centric change.
Layer 3 — corroboration from the FOIA microdata
Our selected-and-filed petitions show the multi-registration share rising 8.2% → 25.5% (FY21→24) while the registration-pool share rose 10.4% → 53.9% — demonstrating both the over-selection advantage and the lower filing propensity of speculative multi-registrations. (See the multi-registration story.)
How to read these numbers
- These are odds of selection, never "odds of a visa" — selected registrations still must be filed and approved.
- FY21/22 per-registration rates are inflated by multiple non-filing rounds.
- The k-conditional table uses the binomial identity 1 − (1 − p)k (an independence approximation; the finite-population correction is negligible at k ≤ 10 out of 759k). The only estimated input — the ~4.3 average registrations — traces to USCIS's own published ~446k unique-beneficiary count.
- A bare "your odds were X% in FY2024" is misleading without the single-vs-multi conditioning; the unconditional number was practically meaningless pre-FY25, which is the whole point.
Explore the data yourself
The per-year selection rates above, the cap-season calendar and the history of additional selection rounds are kept current on the H-1B lottery pages. The multi-registration behavior behind these odds is broken down in the multi-registration story, and the pay side in the H-1B pay decomposition. For the workers who were selected and filed, the wage search shows certified pay by employer, role and state.
Source & method
The petition-level figures come from the I-129 microdata released under the Freedom of Information Act — "sourced from USCIS, obtained by Bloomberg" (Bloomberg v. DHS), the same dataset behind Borjas's H-1B wage-gap paper (NBER w34793). Registration-pool figures (eligible registrations, selections, the FY2025 rule effects) are from USCIS's own published H-1B Electronic Registration statistics and its FY2025 selection-rule rulemaking.
Universe: fiscal years 2021–2024, cap-subject H-1B lottery petitions that were selected and filed (372,841 rows) — not the full registration pool. A frozen FOIA snapshot; per-beneficiary identifiers are redacted, so figures are aggregates only. Every number on this page was checked against the live database on 2026-07-13.
Custom data slices for research or reporting. If you're a researcher, academic or journalist and want a specific cut of this data — a breakdown by employer, occupation, country of birth, fiscal year, or any combination not shown here — get in touch and I'll put it together. Attribution to visa-bulletin.us is appreciated.