DEV Community

Cover image for The ENABLE Lexicon Study — Letter Frequency vs Scrabble Tile Values
Lucian (LKB)
Lucian (LKB)

Posted on • Edited on • Originally published at lkforge.com

The ENABLE Lexicon Study — Letter Frequency vs Scrabble Tile Values

This was originally published on the LK Forge blog, where the charts are interactive.

Scrabble's tile values were set in the 1930s by counting letters on a newspaper page. We counted every letter across all 172,835 words in the ENABLE open word list — not the TWL or CSW/SOWPODS Scrabble dictionaries, just the open list — to see how well those 90-year-old values still line up with usage. They don't, in eleven places.

C is the 10th-most-common letter in English word-stock, yet Scrabble scores it 3 points — more than the 1-point U it outnumbers. Eleven letters carry a tile value that doesn't line up with how often they actually appear.

The numbers

ENABLE words analyzed 172,835
Tile value / frequency inversions 11
Letters in the longest isograms 15
More words start with S than X 139×

How often each letter actually appears

Frequency here is occurrences per 1,000 letters: every letter of every word counted, divided by the total letter count, times a thousand. E leads at 115.1 per 1,000, S is second at 95.2, I third at 90.1. Q and J tie for last at 1.6.

Bar chart of letter frequency per 1,000 letters across all 172,835 ENABLE words, ranked from E (115.1) down to the rare Q and J (1.6), bars colored by Scrabble tile value. Red arcs mark all 11 tile-value inversions — each a higher-value letter that is more common than a lower-value one it links to, e.g. C outranks D, G and U

Where value and frequency diverge

Butts set higher values for rarer letters, so by design tile value roughly tracks how uncommon a letter is — a rough stand-in for how hard it is to play. Across the full ENABLE list that alignment holds at the extremes (E is common and worth 1; Q and Z are rare and worth 10) but breaks down in the middle. The harness finds all 11 inversions — every case where a higher-value letter is actually more common than a lower-value one. Here is the complete list:

More common (tile value) Rarer, but lower-value (tile value) Frequency (per 1,000 letters)
H (4 pt) B (3 pt) H 23.2 vs B 18.4
C (3 pt) D (2 pt) C 40.9 vs D 33.8
C (3 pt) G (2 pt) C 40.9 vs G 27.0
C (3 pt) U (1 pt) C 40.9 vs U 32.7
D (2 pt) U (1 pt) D 33.8 vs U 32.7
M (3 pt) G (2 pt) M 28.4 vs G 27.0
P (3 pt) G (2 pt) P 29.4 vs G 27.0
K (5 pt) W (4 pt) K 8.5 vs W 7.4
Q (10 pt) J (8 pt) Q vs J — 2,536 vs 2,497 uses*
Z (10 pt) J (8 pt) Z 4.7 vs J 1.6
Z (10 pt) X (8 pt) Z 4.7 vs X 2.9

* Q and J both round to 1.6 per 1,000 letters, so the ranking uses exact counts: Q edges J on raw usage (2,536 vs 2,497) across the 172,835 words.

The single cleanest example is H against B: H is worth a full point more, but shows up more often across the lexicon. If tile values were re-derived from ENABLE frequencies today, several 3- and 4-point consonants — C, H, M, P — would come out lower.

The longest words with no repeated letter

An isogram is a word in which no letter repeats. The longest in ENABLE are exactly 15 letters — and there are only two of them: dermatoglyphics and uncopyrightable.

The most anagram-rich letters

Group every word by its sorted letters and one set wins: the five letters A, E, P, R, S rearrange into 12 different ENABLE words — apers, apres, asper, pares, parse, pears, prase, presa, rapes, reaps, spare, spear.

Letters Words The words
A E P R S 12 apers, apres, asper, pares, parse, pears, prase, presa, rapes, reaps, spare, spear
A E L R S T 11 alerts, alters, artels, estral, laster, ratels, salter, slater, staler, stelar, talers
A E L S T 10 least, setal, slate, stale, steal, stela, taels, tales, teals, tesla
A E I N R S T 9 anestri, antsier, nastier, ratines, retains, retinas, retsina, stainer, stearin

Which letter starts the most words

S runs away with it: 18,988 ENABLE words start with S. At the other end, only 136 start with X — a spread of roughly 139 to 1.

Bar chart of how many ENABLE words start with each letter, ranked from S with 18,988 down to X with 136, S and X highlighted

About the Scrabble reference

SCRABBLE is referenced here only descriptively, as the source of the standard English tile-value set — a public reference point for comparison, not an official Scrabble analysis. SCRABBLE is a trademark of its respective owners; this post is not affiliated with, endorsed by, or sponsored by the trademark holder.

Reproduce it yourself

The harness is public and deterministic — one script against the shipped ENABLE list (172,835 words), printing frequency per 1,000 letters, tile-value inversions, isograms, anagram-richest sets, and starting-letter counts.

ENABLE lexicon harness (public gist)

Every number on this page is printed by that one script. Run it yourself to check any figure above to the exact digit.


The interactive charts are on the LK Forge blog.

Top comments (0)