WE

Worth 5 points before any premium square, and the 12th highest-scoring of 96 — tied with 21 others.

Hooks

An opening rack can play it

7.5%

Spending a blank

15.2%

WE extends into 8 three-letter words, which is the 34th least extendable of 96 — tied with 7 others. That is the number worth memorising: a two-letter word is a place to put tiles, but a well-hooked one is a place to put them where the board stays open.

Every hook

In front — makes _WE

A→AWE
E→EWE
O→OWE

Behind — makes WE_

B→WEB
D→WED
E→WEE
N→WEN
T→WET

Not all of them are equally reachable, which a bare count of 8 hides. WE hooks on 7 distinct letters, and the ones you are likeliest to be holding are E (60% of opening racks), A (49% of opening racks) and O (45% of opening racks). The rarest is B, at 14%.

How often you can actually play it

Out of an opening rack, 7.48% of the time you hold the letters for WE outright, the 44th likeliest of 96 — tied with 8 others. Counting a blank as whichever letter is missing lifts that to 15.19%, though spending one of the two blanks on a two-letter word is rarely the play. Exact hypergeometrics over the 100-tile bag, not a simulation.

Read the hook count as a floor. Hooks are counted against ENABLE’s 972 three-letter words. Both official lists carry more three-letter words than that, so a word’s real hook count at a rated game is this number or higher, never lower. How much higher is not quantified here, and this page will not guess at it.Which list this is. ENABLE, which is public domain — the reason it can be published here at all. Its 96 two-letter words are a strict subset of NWL2023’s 107; the 11 it does not carry are DA, EW, FE, GI, KI, OI, PO, QI, TE, ZA, OK. Collins adds 20 more again — the letters those add back are set out on the index page.

Every two-letter word

Try next

All 96, ranked

Which ones hook, which ones are dead ends, and the four letters that have none.

The 972 three-letter words

Three quarters of them are a two-letter word plus one tile.