Your second guess: choosing a follow-up

Unless your opener solves the puzzle, its row gives you evidence for the next move. The second guess is your first chance to respond to evidence. It can investigate a different set of letters, narrow a position, or test a plausible solution. Choose it by reading what your first word revealed.

If your opener came back mostly empty

When almost every tile is ruled out, treat the second guess like a second opener: five fresh, common letters you have not tried. After CRANE comes back blank, a word like MOIST or DOILY tests ten different letters across the two guesses. That coverage is useful, but it does not establish how many candidates survive a particular pair of rows. This is the “two-word opening” from the starting-word comparison.

If you picked up a letter or two

A locked or floating letter is useful, but do not let it trap you into a timid guess. In normal play you are still allowed to spend the second turn gathering information: keep any locked letter in place if it is easy to, move a floating letter to a new spot, and fill the rest with letters you have not tested. Compare how the new word would separate the remaining candidates. A possible answer can do that too; a probe is not automatically the more informative choice.

Several locked letters can suggest a short list, but a shared ending may still leave many answers. Count the plausible alternatives before treating a nearly filled row as nearly solved.

Weigh “use it” against “learn more”

Every second guess is a trade-off between using what you know and learning what you do not. New letters help when the surviving candidates disagree about them. Repositioning known letters can be more useful when their arrangement is the main unresolved question. Later, once the list is short, using your clues to commit is the right call.

One caveat: on a self-imposed hard-mode style, every guess must respect your locked and floating letters, so the “play a fully fresh word” move is off the table — you trade some flexibility for the discipline.

Let the Solver check your shortlist

Not sure which follow-up splits the field best? Feed guess one’s result into the Word Solver and it lists the accepted words that still fit, so you can pick a second guess that tells the different candidates apart. Still stuck after that, see breaking a stuck board. Ready to play? Start today’s word or practise in Unlimited.

Follow-ups after an all-gray STARE

From the 1,402-answer pool checked on October 17, 2026, an all-gray STARE leaves 121 candidates; the calculations below give each survivor equal weight. They do not apply to STARE with a correct or misplaced tile. “Average remaining” measures the compatible group after the next feedback; “worst group” is the largest such group. Both include the answer itself, even when the guess would solve the game.

CLOUD: average remaining 4.1; worst group 8.
BUILD: average remaining 4.3; worst group 9.
NOISY: average remaining 7.2; worst group 14.
MOIST: average remaining 14.2; worst group 22.

CLOUD and BUILD provide smaller average groups than MOIST in this specific comparison. MOIST reuses S and T, which the empty STARE row already excluded. But counting fresh letters is not a complete explanation: the remaining letters must also divide the surviving words usefully. NOISY reuses S yet still gives a different result from MOIST. The calculation examines full feedback patterns, including positions, rather than adding independent letter frequencies. These are selected follow-ups, not a search proving an optimal word across the entire accepted dictionary.

When the first row is not empty

Consider a fictional example with PLANT as the hidden answer and STARE as the opener. The row is − ↔ ✓ − −. A is fixed in the middle, T belongs outside the second column, and S, R and E have no match. The all-gray comparison above is now irrelevant. Its candidate group excludes A and T, although this board requires them. Applying that comparison anyway would mean solving the wrong filtering problem.

A follow-up such as CLAMP keeps A, tests C, L, M and P, and leaves the T placement question for later. In this fictional game it returns − ✓ ✓ − ↔. A different choice might include T and learn its position sooner. The useful comparison is between the remaining candidates each choice distinguishes, not between how attractive the words sound. Puzzico permits either approach because it does not enforce hard mode.

Plan the decision after the follow-up

Before submitting, ask what you would do if the new letters all missed. If the answer is simply “try another word with the same ending,” you may be entering a large family without a way to distinguish its members. Look for a guess that tests several unresolved letters, or choose a candidate that gives useful feedback even when it is wrong. Keep enough turns to act on the information; a probe that identifies a solution after your final turn has no practical value.

After the row arrives, discard your planned third word if it conflicts with the evidence. Recheck earlier arrows and duplicate-letter counts before narrowing the list. The feedback examples explain those counts, while the word lists can help when you know the pattern but cannot recall a spelling. Neither a prepared pair nor a solver result replaces checking the actual board.

Read the follow-up without losing the opener

Keep a compact record of fixed positions, excluded positions and minimum copies. The newest row adds constraints; it does not replace the opener. A letter missing from the second guess can still be required by the first. This matters after a free probe, because the probe may deliberately omit letters already known to be present. Return to the combined evidence before deciding that a newly suggested word fits.

Enter the two-row example in the Solver

For STARE followed by CLAMP against PLANT, put L in the second green box and A in the third. Put T in the second yellow box, because STARE ruled out that column for T, and P in the fifth yellow box, where CLAMP found it misplaced. Enter S, R, E, C and M in the gray field. Leave the other position boxes empty, then select “Find words”.

The green and yellow boxes in the same column describe different constraints: L belongs there, while T belongs somewhere else. Do not move T to a guessed destination in the yellow row; that row records the rejected location. P remains required even though it was absent from your opener. PLANT satisfies these combined entries, but inspect every suggestion against any additional rows you have, since the form does not import a game.

For a completed-game review, use the checklist in six mistakes to check.

Keep exploring

Play today’s Puzzico or browse all guides & strategy.