Wordle Starting Word Calculator
Type any five-letter word and we'll rate it against every one of the 2,343 legal Wordle answers — not against English at large, which turns out to be a different thing entirely. Try your usual opener, then try ORATE.
Letters only — we rate the letters, so it doesn't have to be a dictionary word.
What the three scores mean
Wordle answers reached is the headline number: the percentage of all 2,343 answers that share at least one letter with your word. If your opener reaches 97% of answers, then on 97% of days your first guess is guaranteed to light up at least one tile — real information, every game. The average-letters figure underneath tells you how much information: an opener that averages two hits per answer narrows the field far faster than one that averages one.
Wordle letter share is the fraction of all 11,715 letters in the answer list that your five letters account for. It's the purest measure of whether you picked popular letters. English letter share is the same calculation against Oxford's letter-frequency analysis of the dictionary — and comparing the two columns is where it gets interesting, because they disagree more than you'd think.
What 515 million games actually open with
The New York Times has the one dataset nobody outside the game can build: what real players actually type first. Its analysis of 515 million games played from December 2022 to November 2023 produced a ranking of the twenty most popular openers, and ADIEU tops it. We ran all twenty through the same arithmetic this calculator uses, and added one column the tool above doesn't show — how often each word lands a letter in the right square, not merely somewhere in the answer.
| Opener | Answers reached | Right square | Avg hits | Can be the answer |
|---|---|---|---|---|
| 1ADIEU | 93.0% | 29.1% | 1.49 | never |
| 2STARE | 90.2% | 41.4% | 1.76 | yes |
| 3SLATE | 90.5% | 46.4% | 1.68 | yes |
| 4AUDIO | 91.8% | 23.9% | 1.33 | yes |
| 5RAISE | 92.7% | 43.8% | 1.76 | yes |
| 6CRANE | 88.6% | 43.7% | 1.64 | yes |
| 7ARISE | 92.7% | 40.3% | 1.76 | yes |
| 8IRATE | 91.6% | 39.7% | 1.77 | yes |
| 9TRAIN | 88.6% | 35.2% | 1.56 | yes |
| 10GREAT | 88.0% | 33.0% | 1.62 | yes |
| 11HEART | 88.7% | 35.9% | 1.66 | yes |
| 12AROSE | 92.1% | 39.4% | 1.77 | yes |
| 13HOUSE | 87.4% | 38.7% | 1.38 | yes |
| 14AISLE | 91.5% | 37.1% | 1.68 | yes |
| 15STEAM | 87.7% | 27.8% | 1.53 | yes |
| 16LEAST | 90.5% | 37.0% | 1.68 | yes |
| 17CRATE | 89.4% | 43.1% | 1.69 | yes |
| 18TEARS | 90.2% | 32.1% | 1.76 | never |
| 19SALET | 90.5% | 47.9% | 1.68 | never |
| 20DREAM | 85.1% | 27.0% | 1.50 | yes |
Three of the twenty can never be the day's solution: ADIEU, TEARS, and SALET — the least-played word on the list and the one exhaustive solvers keep landing on.
The ADIEU paradox
Here is the awkward part, and this page's own numbers cause it. ADIEU reaches 93.0% of all 2,343 answers — the best score of the twenty. Better than SLATE, better than CRANE, better than anything a solver would recommend. By our headline metric, the most popular opener in the world is also the best one.
The Times' play data says otherwise. Players who opened with ADIEU took about a third of a turn longer to finish than players who opened with SLATE, and failed to solve the puzzle 3.6% of the time against SLATE's 1.7% — more than double the loss rate, across hundreds of millions of games.
The right-square column is where those two facts meet. ADIEU puts a letter in its correct position in 29.1% of answers. SLATE manages 46.4%, and SALET 47.9%. AUDIO, the other vowel-stacker on the list, is worst of the twenty at 23.9%. That is the entire story of a vowel-heavy opener: stacking A, I, E and U into one word is an excellent way to guarantee that something lights up, because nearly every English word contains a vowel. It is a poor way to learn anything, because the tile that lights up is usually yellow — and a yellow vowel in the wrong slot is close to the least useful thing Wordle can tell you. SLATE gives up two and a half points of reach and buys seventeen points of position with them.
It is a useful reminder that a single headline percentage, ours included, is a compass and not a verdict. If you want to see what that difference costs in practice, we plotted it out in the guess distribution behind Wordle and Pairdle.
Why Wordle and English disagree
The Wordle answer list is not a sample of English — it's a curated, five-letter, no-plurals slice of it, and that bends the alphabet. N and I, dictionary superstars, get snubbed because the -ING and -TION words that made them famous can't fit in five letters. Y nearly doubles its frequency thanks to hundreds of -Y adjectives. Z shows up five times more often than the dictionary says it should. We wrote up the full analysis, with charts for every letter, in Why Does Wordle Hate the Letter N? — this calculator is that article's data, made interactive.
That divergence is why a starter chosen from an English frequency chart can be mis-tuned for the actual game. The five most common letters in Wordle answers (E, A, R, O, T) anagram into ORATE — a statistically ideal opener that, in a delightful twist, isn't in the answer list itself. English's top letters instead build IRATE and RATIO, which are. For the fuller strategy conversation — including why the SLATE/CRANE family keeps winning solver rankings — see our guide to the best Wordle starting words.
Why "the best starting word" keeps changing
If you have read three articles on this subject and come away with three different answers, the articles probably aren't disagreeing about the mathematics. They are quoting different versions of the same bot. WordleBot, the Times' own solver, originally recommended CRANE. In August 2022 it was rebuilt so that it no longer knows which words are on the answer list, and its pick moved to SLATE — LEAST in hard mode. In January 2024 the dictionary behind it was widened to around 4,500 relatively common words, and the pick moved again, to TRACE, with TROPE for hard mode. None of those revisions made the earlier answer wrong; each one quietly changed the question.
The spread at the top is narrow in any case. WordleBot scores CRANE, SLATE, TRACE, CRATE, CARET and CARTE all at 99 out of 100 for skill, which is a polite way of saying it cannot separate them. Run the six through the calculator above and you will see the same verdict from a different direction: under two percentage points of reach covers all of them. Choosing between words in that band is picking a favourite, not optimising.
The openers that can never be today's word
The calculator tells you whether your word is one of the 2,343 possible answers, which sounds like trivia and isn't. An opener that is on the list can win on the first guess. An opener that isn't is pure reconnaissance, however good its letters.
Most of the famous optimised starters fall on the wrong side of that line. SALET, CARET, CARTE, ROATE, SOARE, ORATE and TEARS are all fine letter-fishers, and not one of them can ever be the solution. That is precisely why Tom's Guide, working down the solver rankings, settles on STARE: the same five letters as TEARS and the same 90.2% reach, but STARE is on the list and TEARS is not. The upside is small — a one-in-2,343 shot at a one-guess game — but it is free, so where two openers score the same, take the one that can win.
It is worth auditing the lists you inherit, too. Vocabulary.com's widely shared set of "best starting words" shows how uneven curation gets: ATONE at 91.8% and ALERT at 91.5% are genuinely strong, but they sit in the same list as ODIUM at 74.9%, FRAUD at 80.0% and BLIND at 71.6% — three words that reach fewer answers than anything in the popular twenty above. Type them in and the gap shows up instantly. For the raw material behind all of this, see the most common five-letter words and the rarest answers Wordle has ever used.
A few things the score can't see
Letter coverage is the foundation of a good opener, but it isn't the whole game. Position matters — that is what the right-square column in the table above measures, and as ADIEU shows, it is the single biggest thing the three headline scores miss (Y is only valuable because it lives in slot five). So does the shape of what remains after the greens and yellows come back — solver rankings like WordleBot weigh both. Treat a high score here as "this word fishes in the right pond," not a guarantee of a three-guess week. And words with repeated letters score honestly here: a double letter means only four unique hooks in the water, which the calculator will warn you about. The flip side is also true: a word that scores modestly here can still be a fine opener if its letters sit in their most common positions, so use the score as a compass rather than a verdict.
If you enjoy this kind of daily word puzzle, we build one: Pairdle is our free daily game where you decode a six-letter word two letters at a time — and as the article shows, its letter distribution hews closer to real English than Wordle's does. More game guides live in our guides hub.
More on the same shelf: how to play Wordle if you are starting out, why Wordle is five letters long for where the constraint came from, and our five-letter word puzzles tool if you would rather play than measure.
Methodology: answer-list statistics — reach, average hits and the right-square rate — computed across all 2,343 legal Wordle answers; English baseline is Oxford's letter-frequency analysis of Concise Oxford English Dictionary headwords. Opener popularity order and the solve-rate figures come from The New York Times' analysis of 515 million games played from December 2022 to November 2023; published accounts of that study quote conflicting share percentages, so we cite the ranking and the outcomes rather than the shares. WordleBot version history via WordTips and Tom's Guide. All calculation for the tool above happens in your browser. Last updated 2026-09-04.