Letter Solver" />

Best Starting Words

Ranked by actual letter-frequency analysis across every valid 5-letter word -- not a generic recommendation.

Top 15 Starting Words

Each word below is scored by summing how often its letters (counted once each) appear across all 8,636 valid 5-letter words. A higher score means the word tests more common letters at once -- the classic strategy for narrowing down a puzzle fast.

#WordLetter Score
1AROSE18,514
2ARISE18,189
3RAISE18,189
4SERAI18,189
5ALOES18,051
6ARLES18,014
7EARLS18,014
8LARES18,014
9LASER18,014
10LEARS18,014
11RALES18,014
12REALS18,014
13SERAL18,014
14STOAE17,941
15ASTER17,904

Why letter frequency matters

The five most common letters across the whole word list are S, E, A, O, R (in that order). A starting word using several of these at once eliminates or confirms more possibilities on your very first guess than a word built from rare letters.

The S Problem, and Why the Top of This Table Is Misleading

The table above is honest arithmetic: it scores each word by how often its letters appear across all 8,636 valid five-letter words. But raw frequency across the full word list is not the same thing as frequency across the answers you will actually face, and the gap between those two is almost entirely one letter.

S appears in 4,001 of the 8,636 words, more than any other letter, and 2,695 of those words end in S. That is 31 percent of the entire list sitting in one grammatical bucket: plurals and third-person verbs. Daily five-letter puzzles overwhelmingly draw their answers from a curated list that strips most of those out, because ending a puzzle on a routine plural feels cheap.

So a starting word chosen purely on list frequency will over-invest in S. It is still a fine letter to test, but it is not the best letter to test, and a word that spends one of five slots on a trailing S is usually spending it badly. If you want a single practical adjustment to the table above, prefer the entries where the S is not in position five.

Frequency Is Not the Same as Information

The instinct behind a good opener is right: test common letters, learn a lot. But what you are actually trying to maximize is not how many common letters you touch, it is how evenly your guess splits the remaining candidates.

Think about it from the answer's side. A guess that comes back all grey is enormously informative if it eliminates 6,000 words. A guess that comes back with three yellows is informative in a different way: it tells you which letters are present but nothing about where. A perfect opener produces a result that leaves you with a small, evenly-sized pool no matter which colours come back.

That is why words like AROSE, ARISE, and RAISE sit at the top and stay there under almost any scoring method. They hit three of the five most common letters plus two more in the top ten, and crucially they spread those letters across positions rather than stacking them. ALOES and ARLES do the same job.

It is also why chasing an exotic "optimal" opener is not worth much. The difference between the best word and the fifteenth best is a fraction of a guess on average. The difference between a good opener and a bad one is much larger, and the difference between a good second guess and a bad one is larger still.

Where the Letters Actually Sit

Position matters more than most players account for, and the distributions are lopsided enough to act on.

Position one is led by S with 1,056 words, then C with 665, B with 589, P with 565, and T with 548. A leads the vowels at 497. Words very rarely start with a vowel: all five vowels together account for well under a quarter of first letters.

Position two is where the vowels live. A appears there in 1,479 words, O in 1,335, E in 1,063, I in 966, and U in 791. Adding those up, a vowel sits in position two in more than 60 percent of five-letter words. If you have no information at all, position two is the single best place to bet on a vowel.

Position three is the most evenly spread of the five, led by A at 850 and R at 791. It is the position that tells you least and therefore the one worth testing when you have a choice.

Position four is dominated by E at 1,705, which is more than twice the next letter. If a puzzle has confirmed an E somewhere and you do not know where, position four is the highest-probability guess by a wide margin.

Position five is S at 2,695, then E at 931, then Y at 906, then D at 582. With the plural caveat above applied, E and Y become the two endings worth testing first.

The Second Guess Matters More Than the First

Most players agonize over the opener and then improvise the second guess. That is backwards. The opener is a fixed decision you can make once and reuse forever. The second guess is where the actual skill is, because it is the first one that responds to information.

The key decision is whether to guess a candidate answer or to spend the turn purely on information. If your opener came back with one yellow and four greys, you probably have hundreds of candidates left, and playing one of them is a lottery ticket. A word that tests five completely new common letters will usually cut the field far more than a plausible guess would.

The reverse is true late. Once you are down to a handful of candidates, guessing one of them is correct, because the information you would gain from a non-candidate cannot outweigh the chance of simply being right.

A useful rule of thumb: if you can name fewer than about five remaining candidates, guess one. If you cannot, spend the turn on new letters.

Hard Mode Changes the Answer

In hard mode you must reuse every confirmed letter in every subsequent guess, which removes the information-only guess from your toolkit entirely. That has a specific consequence: openers that produce lots of yellows become liabilities, because a yellow locks a letter into every future guess without telling you where it goes.

Under hard mode, an opener that tends to produce clean greens or clean greys is worth more than one that maximizes letter coverage. It also means you should be more cautious about words with common letters clustered in unusual positions, since a misplaced yellow can trap you into a chain of forced guesses that all differ by one letter.

The classic hard-mode trap is a word family like BATCH, CATCH, HATCH, LATCH, MATCH, PATCH, and WATCH, where you have the last four letters green and are forced to burn guesses one letter at a time. If you see that shape forming and you are not in hard mode, spend a guess on a word containing several of the candidate first letters at once.

Using the Solver After Any Guess

Once you have colours on the board, stop guessing from memory. Put your greens, yellows, and greys into the solver and it will filter the full 8,636-word list against all of your constraints at once, which is a thing human recall does badly and software does perfectly.

What you do with the resulting list is still the interesting part. If it returns forty words, do not pick one at random. Look at what varies among them and choose a guess that splits that variation, even if the guess is not on the list. If it returns four, guess the most ordinary-looking one, because curated answer lists skew toward common vocabulary.

Try it on your puzzle

Use the solver above to narrow down results after any guess -- enter your letters and their colors to see every word that still fits.