DnD human names
Human is the culture with the widest opening table on this site — twelve first letters against six vowels — and the smallest clan vocabulary in proportion to how often it is drawn. Both show up in the numbers below: 10,000 draws, every clan counted, every ending counted, and the two endings that hit the ceiling the generator enforces.
The answer in numbers
| Measure | Value |
|---|---|
| Distinct full names | 9,964 / 10,000 (0.36% repeats) |
| Distinct given names | 8,393 |
| Clan names | 64 (8 heads x 8 tails), each used 156 or 157 times |
| Mean given-name length | 6.98 letters (median 7) |
| Length spread | p10 = 5, p90 = 8, longest = 9 letters |
| Busiest endings | -%ric and -%ley, 1,250 each |
| Quietest endings | -%ay, -%on, -%er, -%an at 937 |
Twelve names from that run:
-
%s
Where the 10,000 draws went, clan by clan
Each cell is how many times that clan name came up. With 64 clans and 10,000 draws, an even split is 156.25 — so a perfectly balanced pool would put every cell at 156 or 157. That is exactly what happened: the busiest cell is 157, the quietest is 156, and the whole grid sits inside one draw of even.
| Head | -ford | -moor | -cott | -holt | -fen | -gate | -mayne | -wick |
|---|---|---|---|---|---|---|---|---|
| Ash | 156 | 156 | 156 | 156 | 156 | 156 | 156 | 157 |
| Black | 156 | 156 | 156 | 156 | 156 | 156 | 156 | 156 |
| Dray | 156 | 157 | 156 | 156 | 157 | 156 | 156 | 156 |
| Fair | 156 | 156 | 156 | 156 | 157 | 156 | 156 | 156 |
| Grey | 156 | 157 | 157 | 156 | 157 | 157 | 157 | 156 |
| Harrow | 156 | 156 | 156 | 156 | 156 | 157 | 156 | 156 |
| Marrow | 157 | 156 | 157 | 156 | 156 | 157 | 156 | 156 |
| Red | 156 | 156 | 156 | 157 | 157 | 157 | 156 | 156 |
The two endings that hit the ceiling
The generator caps any single ending at ceil(n / 10 x 1.25) names per batch — at 10,000 draws that ceiling is 1,250. Two endings reach it exactly, and the other eight share what is left. The row of numbers below adds up to 10,000, which is the check that nothing was hand-edited:
| Ending | Names | Share |
|---|---|---|
| -ric | 1,250 | 12.50% |
| -ley | 1,250 | 12.50% |
| -ic | 938 | 9.38% |
| -ard | 938 | 9.38% |
| -ell | 938 | 9.38% |
| -ins | 938 | 9.38% |
| -ay | 937 | 9.37% |
| -on | 937 | 9.37% |
| -er | 937 | 9.37% |
| -an | 937 | 9.37% |
If two endings taking 12.5% each looks like a bias, it is the opposite: it is the ceiling doing its job. Without it, a random draw would let one ending run to 13–14% and starve the rest; with it, no ending can pass 1,250 in a batch. If you want the flattest possible spread across endings, draw in batches under about 800 names, where the ceiling never binds.
How long human names run
Given names run 5 to 9 letters. Half of them are 7 letters or shorter, 90% are 8 or shorter, and nothing in 10,000 draws reached 10. That is a narrower band than elf (mean 7.88 letters on the same site) — human names are shorter and cluster harder around the middle.
How this was measured
Pools. Given names: 12 openings x 6 vowels x 10 endings. Clan names: 8 heads (Ash, Black, Dray, Fair, Grey, Harrow, Marrow, Red) x 8 tails (ford, moor, cott, holt, fen, gate, mayne, wick).
Sample. 10,000 human draws at seed 20261028, gender Neutral, five styles cycled in order, one tracker for the whole run.
Counting. Ending = the closing the generator returns; length = letters in the given name; percentiles are taken from the sorted length list, so p50 is the median and p90 is the 90th percentile.
Reproduce it
var s=document.createElement('script'); s.src='/assets/human-stats.js'; document.body.appendChild(s);
Open the generator, paste that line in the console, and it prints the same clan grid and ending table for seed 20261028.
Generate human names · Dragonborn names, 64 clans · Surnames and clan names