Sources

Every figure on this site falls into one of three categories, and we would rather tell you which than let you assume. The third category is the important one: two things here have no source dataset behind them, and both say so on their own pages.

Computed

Derived from first principles and verified by enumeration.

Serial number pattern odds (34 patterns)

Every count is the number of serials satisfying the pattern definition across the 99,999,999 that can be issued in a run. Each is derived in a comment in src/lib/patterns.ts and re-derived independently in the test suite, and a slow test enumerates all 99,999,999 serials to confirm the detector agrees with its own figures.

Limits. Counts are definitional and overlap: a serial can match several patterns, and each count is for that pattern alone.

Value bands

Generated by a documented model — a premium added to face value, scaled sub-linearly by denomination and heavily by condition — not looked up from a price list.

Limits. The model is a judgement about how the market behaves. The inputs are informed estimates, which is why every output is a wide range and never a single number.

Primary source

Taken from a named public document, machine-parsed, and re-checked monthly.

Star note print runs

BEP Monthly Production Reports (public domain). 582 runs across 220 denomination and prefix combinations, series 1999–2025, parsed from 284 monthly production report PDFs. Every lookup result names the specific report file its run came from.

Limits. The BEP's online archive begins with fiscal year 2003. 74 further runs were parsed but withheld because the parser could not read them unambiguously.

Districts, series letters and production ceilings

Federal Reserve district assignments, the series-letter table for $5 and above, the exclusion of the letters O and Z, and the skip-numbering ceilings of 96,000,000 and 99,200,000 are all BEP and Federal Reserve practice.

Limits. The series-letter table is not chronological and is not meant to be. It reflects the order printing began per denomination.

No source dataset

Market observation. No public dataset exists, so these are the least reliable figures here and are labelled as such wherever they appear.

Printing error premiums

The bands on the misprinted bills page reflect commonly observed dealer and auction asking prices. There is no production or sales dataset for error notes.

Limits. Error values depend on how dramatic the error is and on grade, neither of which a band can capture. Treat them as a sense of scale only.

Silver certificate values (8 series)

Same basis: commonly observed dealer and auction ranges, split by denomination and condition.

Limits. No public price dataset exists for these notes. The ranges are wide on purpose and a real price comes from completed sales of the same series in the same grade.

Licensing and independence

Bureau of Engraving and Printing output is a US Government work and is in the public domain under 17 U.S.C. §105, so the print-run data is free to use and redistribute, commercially included.

SerialValue is not affiliated with the Bureau of Engraving and Printing, the Federal Reserve, Paper Money Guaranty, PCGS, or any dealer, auction house or grading service. We do not buy, sell, broker or appraise currency, and no figure on this site is influenced by anyone who does.

Third-party references we describe but do not reproduce

The 70-point grading scale and the EPQ designation described on the PMG grading page are that company's published standards. We explain what the grades mean and deliberately quote no fee figures, because those change and a stale price here would be worse than no price. Check current tiers with the grader directly.

Corrections

A correction with a citation is worth more to this site than any feature. If a derivation is wrong, a print run is misread, or a band is out of touch with what notes actually sell for, that is a bug and it gets fixed in the next build.