Love calculator

Only the five letters in LOVES matter; every other letter of the names is ignored. The tester counts each of the five across both names and sets the counts side by side, in the order L, O, V, E, S. Each digit is then added to its right-hand neighbour, which gives a new row. The adding repeats, row by row, until only two remain, and they are read as a percentage.

The LOVES game was set out in full by Iain G. Johnston of the University of Birmingham in his 2016 paper "Endless love: On the termination of a playground number game". The University's news item says it is played "in playgrounds and online across the world"; all it needs is pen and paper. Write the two names, then make a letter count: how often each letter of LOVES appears in the names together. The word LOVES itself is not counted. Next, add each pair of adjacent digits to make the next row. Johnston's rule repeats this "until we reach a string containing fewer than three digits". The two digits left form a two-digit number, which is the score. In his own example, Alice and Bob hold L 1, O 1, V 0, E 1, S 0. Their rows run 11010 → 2111 → 322 → 54, so the pair scores 54%.

A sum of 10 or more is written as two digits, as in Johnston's examples, so a row does not always get shorter. Oliver and Lily show it: in the row 643, the first two digits make 10. The next row is therefore 107, and one more step gives 17%. A few pairs never reach two digits at all, which Johnston calls endless love.

Olivia and Muhammad were the most given girls' and boys' names of 2025 in England and Wales, according to the Office for National Statistics (ONS), and together they score 74%.

Between them, the two first names hold L 1, O 1, V 1, E 0, S 0, so the first row reads 11100. Adding neighbours gives 2210 and 431, and the last row, 74, is the result.

The other top pairs from the same lists come out differently, with Lily and Noah on 51%. The ONS lists for 2025 are headed by Olivia, Lily and Amelia for girls and by Muhammad, Noah and Leo for boys.

The tester's known answers include Johnston's Alice and Bob, the TRUE LOVE example from Her Campus and the top two pairs.

Yes, for a few pairs. Johnston calls it endless love: either a row comes back, so the adding runs in a loop, or the rows keep getting longer. His example Steve-O and Esmie never finishes.

Their rows run 01142 → 1256 → 3711 → 1082 → 1810 → 991 → 1810. Row 5 comes back as row 7, so the pair would cycle through 1810 and 991 for ever. The tester stops as soon as a row repeats and shows "Endless love" with the row that came back.

The other kind of endless game happens when big digits keep making two-digit sums, so each row comes out longer than the one before. The tester gives up once a row passes 120 digits; that limit is this site's own choice.

Endless pairs are rare. Running every starting row in which no letter of LOVES is counted more than 3 times, 1,024 rows in all, this site finds 41 that never end, about 4%.

Johnston's paper, also published in Recreational Mathematics Magazine, works out each name's expected score: its average result against every other name on a list. It finds that no pair among the 50 most popular UK names loops.

TRUE LOVE gives a percentage from two totals instead of a row of five counts. Her Campus describes it as the game Keisha and Justin play in Netflix's series Forever, where it gives them 31%.

First count how often T, R, U and E appear in both names and add them up. Then do the same for L, O, V and E. Write the TRUE total and the LOVE total side by side. Keisha and Justin make TRUE 3, LOVE 1, which reads as 31%.

E belongs to both words, so every E in the names counts once in each total. The top pair of 2025 scores 13% this way, a long way from its LOVES result.

No source says what to do when a total reaches 10 or more, which makes the row longer than two digits. The tester then adds it down with the LOVES rule; that is this site's house rule. Genevieve and Oliver make TRUE 6, LOVE 10, so the row 610 adds down to 71%.

This version is often searched for as a true love calculator. Either total can be zero, so a score under 10 is possible.

FLAMES stands for Friends, Lovers, Affectionate, Marriage, Enemies and Siblings, and it ends on one of those words instead of a percentage. Johnston mentions it as a game "sometimes played" beside LOVES. It works as a counting-out game, like a rhyme that picks who is "it".

1. Write both names and cross out the shared letters, one for one: each letter of the first name cancels one matching letter in the second.
2. Count the letters left in both names together.
3. Count that number round the six words, starting at Friends, and strike out the word the count ends on.
4. Start the next count on the word after the one struck out, skip words already gone, and repeat until one word is left.

Take Rohan and Lenna. Crossing out A and N leaves 6 letters, and counting 6 round the ring strikes out Siblings, Friends, Affectionate, Lovers and Enemies in turn. Marriage is left.

Versions differ, and a flames calculator can follow any of them. Some players cross out every copy of a shared letter, and some read the word after a single count without striking anything out. This tester crosses one for one and strikes out until one word remains.

Famous couples get no special treatment: Romeo and Juliet end on Enemies, although the LOVES count gives them 98%. When two names share every letter, nothing is left to count and no word is picked.

No. The score comes from counting a handful of letters and adding digits, so it measures nothing about the two people. It is a playground game played for fun, and it is not scientific.

Names do turn up in research on attraction, but not like this. In 2004, Jones, Pelham, Carvallo and Mirenberg published a study in the Journal of Personality and Social Psychology. It found people "disproportionately likely to marry others whose first or last names resemble their own". The authors read it as implicit egotism, a quiet liking for things linked to oneself.

Uri Simonsohn's 2011 paper in the same journal replicated those findings. It traced them most likely to cohort, geographic and ethnic confounds, together with reverse causality. Neither paper looks at a letter count such as LOVES.

So the percentage is no measure of compatibility. A low score only shows how the digits happened to add up, and a high score says just as little.

Same names, same result: nothing in the count is random, so a pair scores the same every time and for everyone. The order makes no difference either, because each count covers both names together. Muhammad and Olivia give 74% whichever name comes first.

Spelling does matter, because every letter typed is counted. Nicknames change the letters, so they change the score: Mo and Olivia give 15%. Full names add more letters again, and with surnames Olivia Smith and Muhammad Khan give 75%.

• Accents count as the plain letter, so Zoë and Chloé give the same 98% as Zoe and Chloe.
• Capital and small letters count the same.
• Spaces, hyphens, apostrophes and full stops are skipped, so a double-barrelled name counts as its letters.
• Digits and symbols are refused, and a name may run to 40 characters.

The tester works for any two names, so it suits a party game as well as a quiet look at a crush. One name can also be tested against up to 12 others at once, and the results come back as one list, scores sorted highest first.

Copy result puts a line with both names, the score and a share link on the clipboard, ready to paste into WhatsApp or any other chat. Whoever opens the link sees the same names and the same steps.

The names travel after the # sign in that link. RFC 3986, the standard for web addresses, says in section 3.5 that this part is separated before the page is fetched, so the names never reach a server. Nothing is kept in the browser either.

For a game in which one person guesses a word from the others' clues, the forehead guessing game also runs from a single phone.