Methodology · OFLC wage year 2026-27

Show me the math

For the curious and the skeptical. Every number on H1B Math is computed from public data with the formulas on this page. None of them is generated.

The short version

Your job, work location and salary set your OFLC wage level. The level gives you 1 to 4 entries in the lottery. We compute the chance that one of your entries is drawn when drawn tickets are not put back, in two rounds, for three possible sizes of the FY2028 pool.

1Your wage level

OFLC publishes four wage thresholds for each job code (SOC) and area. Your level is the highest threshold your salary equals or exceeds. Hourly thresholds become annual at 2,080 hours a year. With a salary range, the lower end counts. With several worksites, the lowest level counts.

Level III, Software Developers, Dallas-Fort Worth-Arlington, TX$66.63 × 2,080 = $138,590.40 → starts at $138,591

We always round up: a dollar short is a level short.

ZIP code to wage area

A ZIP code is matched to Census 2020 ZIP code tabulation areas, then to counties, then to OFLC wage areas. When a ZIP code spans several wage areas and none of them covers at least 90% of its land area, the calculator asks you to choose. Land area is only a proxy for where people work. PO box and single-organization ZIP codes have no tabulation area and are not covered.

Data quality notes

When OEWS has no estimate for an area, OFLC falls back to a wider geography. The calculator marks thresholds built from the area together with adjacent areas, from statewide data, or from national data.

Top-coded occupations

For rows OFLC labels "High Wage", mostly physicians and surgeons, OEWS cannot estimate the wage and OFLC issues $115.00 an hour ($239,200 a year). The calculator treats every level as starting at that amount: at or above it the salary is Level IV, below it the salary is below Level I. This reading of the OFLC technical notes is an assumption and is being checked against the FLAG Wage Search.

2Entries

Under the weighted selection rule (90 FR 60864), each level puts a different number of entries into the draw.

Lottery entries by wage level
Level	Entries
Level I	1
Level II	2
Level III	3
Level IV	4
Below Level I, with a legitimate alternative wage source	1

3The draw

USCIS first draws for the regular cap, then draws again among US master's registrants who were not picked. A drawn ticket is not put back, so each draw changes the odds of the next one. That is the same as a race of exponential clocks, one per registration, ticking at a rate equal to its entries.

round 1, everyoneΣ c · (1 − e^(−w·τ₁)) = K₁ K₁ = m × 65,000round 2, US master's not yet pickedΣ c · e^(−w·τ₁) · (1 − e^(−w·τ₂)) = K₂ K₂ = m × 20,000your chancewithout US master's P = 1 − e^(−w·τ₁)with US master's P = 1 − e^(−w·(τ₁ + τ₂))

w is entries (1-4), c is the number of registrations with that weight, m is the selection multiplier: how many registrations USCIS selects per available number. If a round has fewer registrations than slots, everyone in it is selected. Four entries behave like four independent Level I entries, P4 = 1 − (1 − P1)^4, not like four times the Level I chance. The model is checked against a Monte Carlo simulation of the draw.

4FY2028 assumptions

Nobody knows how many people will register for FY2028. So we show a base value and a range, and every assumption can be changed in the full calculator.

Default FY2028 assumptions
Assumption	Default
Registrations	210,000 base · 170,000 high odds · 260,000 low odds
Selection multiplier	1.45 · 1.41-1.59 since FY2025
Share with a US master's	63%
Registrations by level	30 / 36 / 23 / 11% for I / II / III / IV

Defaults are calibrated on published FY2027 selection data.

5Odds by level, base scenario

Estimated FY2028 selection odds by wage level, base scenario
Level	No US master's	With US master's
Level I	≈26%	≈44%
Level II	≈45%	≈69%
Level III	≈59%	≈82%
Level IV	≈69%	≈90%

210,000 registrations. The calculator's range uses 170,000 and 260,000.

6Why not 61%

The DHS table that gets passed around gives 15 / 31 / 46 / 61% for Levels I-IV. It treats a drawn ticket as if it went back into the drum. Without putting tickets back, on the same inputs, the numbers are 17 / 31 / 43 / 52%.

Inputs from the DHS final rule (90 FR 60864): 320,711 beneficiaries, 94,900 selections, level mix 28 / 55 / 12 / 5%.

7Sources

This site includes information from the O*NET 31.0 Database by the U.S. Department of Labor, Employment and Training Administration (USDOL/ETA). Used under the CC BY 4.0 license. O*NET® is a trademark of USDOL/ETA. This site has modified all or some of this information. USDOL/ETA has not approved, endorsed, or tested these modifications. CC BY 4.0, O*NET license.

Source files and checksums

OFLC_Wages_2026-27.zip sha256 edc01f64e2f805efe577337935e5c664adbd93d3135750aff8015fefe805f08f
census/national_county2020.txt sha256 9f6e5f6eb6ac2f5e9a36d5fd01dec77991bddc75118f748a069441a4782970d6
census/tab20_zcta520_county20_natl.txt sha256 3ed41278d637dc249e0323306f68be8a6c234e3090f4de88ef328dee71aeaaaf
census/tab20_zcta520_cousub20_natl.txt sha256 406d2f1b11692a185e930e53f63a68951e5b64dbb7b0cf201a934a8e54aee27b
census/ct_cou_to_cousub_crosswalk.xlsx sha256 f70c3ce6bac6e1e667ee29c150735224a9fb19d66edbdef17822ddc8006c238a
db_31_0_text.zip sha256 6883548adf5fde64cf6f801b35d15519c9225f2732c3cab0e281c652d16b23a9

OFLC wage year 2026-27 (BLS OEWS May 2025) · file released 2026-07-01 · page built 2026-10-28 · Rule: 90 FR 60864

8Changelog

2026-10-14: First draft with OFLC wage year 2026-27.

This is not a substitute for the advice of an attorney.