In real-world data — expenses, populations, invoices — a leading 1 shows up about 30% of the time and a 9 under 5%. Paste a column of numbers and this scanner compares their first digits with the log₁₀(1 + 1/d) curve, computes χ², the mean absolute deviation and a Nigrini-style conformity verdict used by fraud auditors.
| Leading digit | Count | Observed | Benford | Difference | χ² cell |
|---|
Benford's law holds when data spans several orders of magnitude and isn't bounded (invoice totals, town populations, river lengths). It fails for uniform or bounded things like ages or lottery numbers — check the data type before believing the verdict. A flag means “look closer”, never “guilty”.