5-Letter Words That Start and End With the Same Letter

691 five-letter words start and end with the same letter

Start with the shape of the constraint, because it explains the size of the list. A five-letter word that starts and ends with the same letter has the form A B C D A. Only two of the five positions are locked together; the three in the middle are free. That gives 26 × 26 × 26 × 26 = 456,976 possible letter strings, and 691 of them happen to be English words. The hit rate is 0.151%.

Compare that with the same arithmetic for a palindrome, which is A B C B A and locks two pairs instead of one. That shape leaves 26³ = 17,576 possibilities and yields 35 real words. And compare with a word with no structural constraint at all, where the 12,478 real words are found among 26⁵ = 11,881,376 strings, a hit rate of 0.105%.

Put differently, the same-letter list is about twenty times larger than the palindrome list and about twelve times smaller than the list of words with five different letters. It sits in the middle, and that is exactly why it is the most useful of the three for filling a slot you have already half-solved.

One fact about the whole set is worth holding on to, because it explains a lot about the words themselves. Every single one of the 691 repeats a letter. If the first and last letters are the same, that letter appears twice, and a five-letter word with a repeated letter is a double-letter word by definition. So this list is a subset of the 4,465 double-letter words on the double-letter page, and it shares no member at all with the 8,013 words that use five different letters.

Why a crossword builder lives on this list

This is the practical reason the page exists, so it is worth being concrete about the mechanics.

In a crossword grid, an across answer in a five-letter slot gets its first letter from one crossing and its last letter from another. Those two crossings are fixed before you fill the entry, and they are frequently the same letter, because a grid is a web of interlocked constraints and repeats are what keep the interlocking working. The moment you know that the first and last letters must be the same, you have left a list of five-letter words with a specific letter at both ends, and this page is that list, sorted by letter.

That is the whole use. Instead of scanning 12,478 words looking for one that starts and ends with N, you click the N chip and get a short list. For N = E you get 47 words, for N = S you get 387, for N = W you get exactly one.

The second letter is the next most valuable column, because it is the first position that is still open after the crossings are fixed. Across the 691 words the second letter is A in 87 cases, O in 85, E in 67, I in 66, L in 46, U in 46, R in 33, T in 33 and H in 33. That is a much flatter distribution than the shared-letter column, and it is the column to work from when you have a crossing in the second square.

There is a second, less obvious reason this list is useful in a grid. Because 68 of the 691 carry two repeated letters and 32 carry a letter three or more times, this list includes a ready supply of the heavily doubled shapes that a grid needs when every crossing letter is already used up. LEVEL, CIVIC, KAYAK and RADAR are all here, and so are MINIM, MALAM, MADAM and REFER, which are palindromes as well.

Grid work rewards having the fill table open, not having it in your head. The crossword puzzle maker will generate a grid, and the crossword solving tips page covers the technique of working a grid from its crossings — this list is the tool that sits beside both of them.

S takes more than half the list

Here is the single most striking fact on the page. Of the 691 five-letter words that start and end with the same letter, 387 begin and end with S. That is 56.01% of the list, from one letter.

Put it against the baseline. S starts 1,521 of the 12,478 five-letter words, which is 12.19% of all first letters. So the same-letter list is roughly four and a half times more S-heavy than the dictionary as a whole. Something about S is doing double duty.

The reason is that S is both the commonest initial and the commonest final letter in English, and the two meet in a very specific way: English marks plurals and third-person verbs with an -S, and a great many of those stems already begin with S. That produces a family of words where an S-initial stem takes an -S ending — and the S appears at the front and the back at once.

Look at the everyday words on this list and the pattern is unmistakable. SALES, SAVES, SEATS, SEEDS, SEEKS, SEEMS, SELLS, SENDS, SHIPS, SHOES, SHOPS, SHOTS, SHOWS, SIDES, SIGNS, SITES, SIZES, SKINS, SLOTS, SOCKS, SONGS, SORTS, SOULS, SPECS, SPIES, SPOTS, STARS, STAYS, STEPS, STOPS and SUITS. Almost every one is a plural or a third-person verb form on an S-initial stem. English simply has a lot of S-initial nouns, and English marks them with an S, and the S-initial ones collide with the marking.

Only fourteen of the 691 have an S in the fourth position as well, and not one has an S in the second position, because the outer S is already spoken for. The 387-word S group is essentially one long column of -S forms.

For a grid builder the consequence is straightforward. An S-to-S fill is easy and has 387 candidates; an R-to-R fill has 35; a W-to-W fill has one. When you are stuck on a slot, the letter you are hoping for is very unlikely to be W, X, B, U, F or G.

The letters that are nearly dead

At the other end of the table the counts collapse. Four letters cannot appear at both ends at all: J, Q, V and Z have no entry in this list whatsoever. Five more have almost nothing.

Two of those rows are trivia gold and worth knowing by heart. The only five-letter word that starts and ends with W is WIDOW. The only one that starts and ends with X is XEROX, which also happens to be the only five-letter word in the language that doubles an X.

Why do J, Q, V and Z manage nothing at all? The reason is the same reason the rare letters are rare in the first place, compounded. A word that starts and ends with the same letter needs a stem that both begins and returns to its own opening letter. English builds such words with inflections, suffixes and borrowings, and it builds them around common letters. A J-initial word is usually a proper noun or a borrowing, and proper nouns do not take a plural -S; a Q-word is nearly always a short unit, and none of them takes an ending that restores the Q. V is common enough in the middle of words and rare enough at the start and the end. Z starts 98 five-letter words and ends 31, but never both.

There is also a compounding effect from the repetition itself. Every word here repeats a letter, and English is reluctant to double its rarer letters. WIDOW and XEROX are the only two entries in the whole table whose repeated letter is a rare one. XEROX repeats its X and is the only five-letter word in the language that doubles an X at all; WIDOW repeats its W, which is unique to it among the 691 because it is the only W-to-W word there is.

Every word on this list repeats a letter, and most repeat only one

Because the first and last letters match, each of these 691 words has at least one repeated letter. But how many repeats, and where, varies in a way that matters for a grid.

591 of the 691 have exactly one repeated letter, and it is the outer pair. Words like ALPHA, DRIED, EAGLE, LOCAL and TRAIT fit that description: the first and last letters are the same and the three in the middle are all different. 68 have two repeated pairs — LEVEL, CIVIC, MINIM, MALAM, RADAR and REFER — and 32 carry a letter three or more times, such as GLOGG, FLUFF and FEOFF.

The breakdown into the three cases is worth memorising because it tells you what the middle three squares are likely to look like: usually three different letters, occasionally a second pair, rarely a triple.

And the 35 palindromes are all inside this list, because a palindrome obviously begins and ends with the same letter. The five-letter palindrome list covers the 35 words where the second and fourth letters match as well, which is the tighter constraint on the same theme.

The middle of these words is a vowel more than half the time

Three letters in the middle of a five-letter word is a very tight frame, and the frame needs a vowel to anchor the syllable. The third letter of these words is a vowel in 358 of the 691 cases, which is 51.81%. Across the whole dictionary, the third letter is a vowel in only 37.45% of words. The second letter is a vowel in 351 cases, or 50.80%.

So this list has a vowel in the middle of the frame roughly one time in two, when the ordinary dictionary manages a bit less than two times in five. That is not a coincidence; it is the same reason the palindrome list has three vowels in a third of its entries. When you force five letters into A B C D A, you have left room for exactly one syllable nucleus, and English puts it in the middle.

The most common middle letters are A in 93 words, I in 78, O in 66, U in 64 and E in 57. The most common fourth letters are E in 139 words, T in 63, A in 41, G in 37 and P in 37. So if you are filling a slot and the middle is open, A is the best bet, and the fourth letter is very likely to be E.

Thirteen shapes cover the whole list

Write a C for a consonant and a V for a vowel and the 691 words fall into thirteen patterns. That is far fewer than the thirty patterns the full dictionary uses, because the outer letters have to match.

CCVCC and CVCVC together account for 380 of the 691 words, which is 55% of the list between them. Both are the classic alternating frame, and both are near the top of the general five-letter dictionary too, where CVCVC leads with 3,054 words and CCVCC is third with 1,375. So the constraint did not distort the shape distribution much — it only removed 17 of the 30 patterns that occur outright.

Two shapes are worth a second look because they are so rare: VVCVV has only two entries, AECIA and EERIE, and CVVVC has exactly one, KAIAK. If you need a five-letter word with a matching first and last letter and three adjacent vowels, there is precisely one answer in 12,478 words.

How many vowels these words have

The vowel counts across the 691 are unusually flat, and that is another consequence of the tight frame. 296 words have exactly one vowel, 294 have exactly two, 93 have three, 6 have none at all and 2 have four.

Compare that with the whole dictionary, where two vowels is the clear default and one vowel is the second place. Here the two groups are within two words of each other. A single-vowel word of this shape has to be something like CIVIC or KNOCK, where the one vowel carries a whole syllable on its own, and there are more of those than you would expect.

The six words with no A, E, I, O or U at all are SKYFS, SKYRS, SYNCS, SYNDS, SYPHS and TRYST. Five of them are Scrabble inflections, and the sixth, TRYST, is an ordinary English word that has been a published answer in the daily five-letter game. The two words with four vowels are AECIA and EERIE.

What these words are worth in Scrabble

The average Scrabble value across the 691 is 8.07 points, against 9.17 for the whole 12,478-word dictionary and 8.98 for the double-letter words generally. So a same-letter word is a little below average, and for a structural reason: the matching letter is usually a common one, and a doubled common letter is cheap.

The expensive end of the list is the Y group. Because Y is worth 4 points, the fourteen Y-to-Y words are the best scores in the set: YUKKY and XEROX are worth 19, KYACK 18, YUCKY 17, and YECHY, MAXIM, KAYAK, DOZED and DAZED are worth 16. YAPPY, YIPPY, YEUKY, YOLKY, YUPPY and YUMMY are all worth 15.

That Y group is worth knowing for a specific reason. It is a set of fourteen words that all start and end with a four-point letter, and every one of them is hard to play unless you happen to be holding a Y. If you are holding a Y and a Y, this list is the answer, and it is the highest-scoring corner of the entire 691-word set.

For everything else, the Scrabble solver is the better tool, because it takes the rack and the board and works out the value. This list is for the stage before that, when you know which letters you have and you want the legal words that use them.

How this compares with the published daily answers

Of the 1,994 published answers in the daily five-letter game, 77 start and end with the same letter. That is 3.86% of the answer set, against 5.54% of the full dictionary, so the answer pool slightly under-indexes on same-letter words.

The 77 are worth reading, because they are the everyday end of this list: ALOHA, ALPHA, AORTA, AROMA, ATRIA, BLURB, CIVIC, COLIC, CONIC, CUBIC, CYNIC, DREAD, EAGLE, EERIE, ELATE, ELITE, ELOPE, ELUDE, EMCEE, ENSUE, ERASE, ERODE, ETUDE, EVADE, EVOKE, EXILE, FLUFF, GOING, HARSH, HATCH, HEATH, HITCH, HUMPH, HUNCH, HUTCH, KAYAK, KIOSK, KNACK, KNOCK, LABEL, LAPEL, LEGAL, LEVEL, LIBEL, LOCAL, LOYAL, MADAM, MAXIM, MODEM, NYLON, OUTDO, PLUMP, POLYP, PRIMP, RACER, RECUR, REFER, RIDER, RIPER, RISER, RIVER, ROVER, ROWER, RUDER, RULER, TACIT, TAUNT, THEFT, TOAST, TRACT, TRAIT, TREAT, TROUT, TRUST, TRYST, TWEET and TWIST.

Notice what is missing from that list of 77: not one of them is a Scrabble inflection. The published answers are all ordinary English, and the answer pool includes only words a solver would actually think of.

Five of the 77 are also palindromes — CIVIC, KAYAK, LEVEL, MADAM and REFER — which is the same five the palindrome page lists. That is the whole overlap: those five are palindromes, and the other 72 are simply words that happen to begin and end with the same letter. MAXIM, for example, is on this list because its first and last letters match, but it is not a palindrome, because M-A-X-I-M read backwards is M-I-X-A-M.

The everyday words among the 691

The dictionary behind this page is a word-game lexicon, and word-game lexicons are stuffed with forms no speaker would produce. This site keeps a separate list of 1,170 everyday five-letter words for exactly that reason, and 65 of the 691 same-letter words are in it.

Those 65 are the ones worth learning by heart, because they are the ones a solver or a setter would actually reach for. They break into two clear groups. Thirty-three end in S, and almost every one of those is a plural or a third-person verb form on an S-initial stem: SALES, SAVES, SEATS, SEEDS, SEEKS, SEEMS, SELLS, SENDS, SHIPS, SHOES, SHOPS, SHOTS, SHOWS, SIDES, SIGNS, SITES, SIZES, SKINS, SLOTS, SOCKS, SONGS, SORTS, SOULS, SPECS, SPIES, SPOTS, STARS, STATS, STAYS, STEPS, STOPS, SUITS and SWISS. The other 32 do not end in S at all, and they are plain nouns, adjectives and verbs: ALPHA, ARENA, CIVIC, COMIC, CUBIC, DATED, DRIED, EAGLE, ELITE, GOING, HEATH, KNOCK, LABEL, LEGAL, LEVEL, LOCAL, MODEM, NYLON, RADAR, REFER, RIDER, RIVER, ROGER, ROVER, THEFT, TIGHT, TRACT, TREAT, TROUT, TRUST, TWIST and XEROX.

That is the practical answer to "which of these 691 would a person actually say?" — about one in ten. The other 626 are real dictionary entries, but they are Scrabble and crossword material.

How to use the list by letter

Here is a fill routine that works.

Step one, fix the shared letter. Read the crossings at the two ends of the slot. If they match, click that letter's chip above. If they do not match, this list does not apply and you need the full 12,478-word five-letter list instead.

Step two, fix the second letter. If you have a crossing in the second square, type it into the "Must include these letters" box together with the shared letter. That turns a 387-word column into a handful, usually under ten words.

Step three, kill the vowels you cannot use. If a crossing has already ruled out a vowel, add it to the include box as a letter you do not want — the box only requires letters to be present, so in practice it is easier to sort Z–A and read past the misses, or to work down the A–Z list and cross off the words containing a letter you have already placed elsewhere in the grid.

Step four, sort and copy. Sort Z–A when the letters you have are late in the alphabet, A–Z when they are early. Copy the result and paste it into whatever you are working in.

One caveat worth keeping in mind: the middle three letters of a same-letter word are the part you still have to solve, and the list gives you no help with those. What it gives you is a guaranteed-legal starting point for a slot that already has its ends fixed. That is worth more than it sounds, because the alternative is guessing at a five-letter word from nothing.

For broader pattern work, the site's five-letter word generator works from a pattern such as _E__E, and the word finder handles anagrams and pattern matches across the same dictionary.

What this list is, and what it is not

Two limits, stated plainly.

First, this is the site's own 12,478-word five-letter dictionary, and it is built for word games. Words like BEROB, UHURU, URUBU, FEOFF, OVOLO, LYSOL and HAITH are in it because game lexicons accept them, not because English uses them. If you want everyday five-letter words grouped by the letter they start and end with, the everyday pool is the list to read instead, and the site's words-by-length page is where the shorter lengths are broken out.

Second, it is not an answer list for any daily game and it is not a grid. It tells you which words exist for a given structural constraint. Whether one of them is the right one depends on the crossings around it, and whether it will ever be a published answer depends on an editorial judgement you cannot derive from a list.