Last Row
The board arrives finished except the last row. Read the colors, deduce the only word that fits, commit.
Loading today's board…
Your record
Results are stored only in this browser.
What pure endgame deduction trains
Last Row isolates the endgame: the part of Wordle where nothing new needs to be guessed, only read. Every clue you will ever get is already on the board, and exactly one word in the site's reconstructed accepted-guess list agrees with all of it. That makes the deduction score luck-free by construction — there is no opener variance to blame and no hidden quirk of the answer list; if a commit is wrong, the evidence was misread, not unlucky. The live consistency check trains the reading itself: type a word and the page names the exact row and clue it contradicts, which is the fastest way to notice the rules you tend to skip. Its sibling Solver Says trains the opposite muscle — choosing the guess that creates the best evidence — and Probe sits between the two. Play the three as a set and the middle game stops feeling like guessing; How Sure? applies the same honesty about evidence to facts instead of words.
How the game and the score work
The given rows are realistic play: a popular or strong opening word, the solver's recommended follow-up for that reply, and — on three-row days — the kind of third guess a real player makes. The puzzle generator keeps a board only when exactly one word in the site's reconstructed accepted-guess list (14,855 words) reproduces every visible color, so the deduction always has one right end. A wrong first commit is not fatal: its colors join the board as new evidence and you get a second attempt.
| Result | Deduction score | What it means |
|---|---|---|
| Solved in 1 | 100 | You extracted everything the board said before committing — the habit this game exists to build. |
| Solved in 2 | 40 | The first commit missed, but you used the extra row it revealed. Real progress, weaker read. |
| Failed | 0 | The visible rows were sufficient on their own — the technique note names the rule that was missed. |
The four technique badges
After you finish, the page names the deduction rule the day's board could not be solved without — found by re-solving the puzzle with each rule relaxed in turn:
- gray sweep
- The greens plus the eliminated gray letters already leave one word — no yellow reasoning needed.
- position elimination
- A yellow tile's "in the word, but not in this slot" ban is what separates the answer from its lookalikes.
- duplicate-letter rule
- Only the strict reading of repeated letters — a gray copy beside a colored copy caps the count — pins the word.
- cross-row synthesis
- No single row decides; each one still allows several words, and only their intersection leaves one.
Answers come from the site's release-pinned Daily answer pool (2,332 words) and consistency is checked against the site's reconstructed accepted-guess list — neither is a claim about what the NYT will choose. There is deliberately no timer and no speed bonus: timed deduction punishes screen-reader, switch and slow-typing players and rewards typing speed, which is not deduction. The score depends only on how many commits the deduction took, so it is the same measure for every player.