How this works

The rating

Every name starts on 1500. When you pick one name over another, the winner takes points from the loser — more if it was an upset, fewer if it was expected. This is the Elo system used for chess ratings.

The amount at stake shrinks as a name accumulates head-to-heads. A name with 20 of them moves fast; a name with 500 barely budges. That means a burst of votes can introduce a new name quickly but cannot drag an established one from the bottom to the top overnight.

Which names you get shown

Not at random. A quarter of matchups pull from the current top 40, so the leaders keep being tested against each other. Half anchor on whichever name has fought least, so coverage stays even. The rest are random. The opponent is always close in rating, because a close fight tells us far more than a blowout.

Filters, and what they mean

You can narrow the field before you start: girls or boys, US or UK, and a style band — Top 100, Popular, Rising, Fading, Revival or Rare. Each band is defined from official figures, not taste. Rising means a name climbed 50 or more places in the US last year; Revival means it peaked before 1970 and is back in the top 1,000; Rare means it is outside both current charts but was still given to at least 40 babies in its best year.

Every chip shows how many names it holds, and a combination too small for a real contest is greyed out rather than serving you the same pair repeatedly.

With no filter set, matchups are weighted by how often babies actually get each name, so about four cards in five are names you would recognise. Choose a band and that weighting is dropped — you asked for that slice, so you get it evenly.

UK movement carries a floor. ONS ranks every name down to three babies, and all of those share a single tie block — so a name going from three births to sixteen would appear to climb four thousand places. We only count UK movement for names given to at least fifty babies, where a rank change reflects real change in use rather than reshuffling at the bottom of the list.

Filtered rankings are ranked within that filter, not against the full field — the page says so when a filter is active.

Where the data comes from

Popularity figures are official. US ranks come from the Social Security Administration, which publishes the top 1,000 names per sex every year. UK figures come from the Office for National Statistics. Historic trend lines run back to 1880.

The ranking on this site is not official and does not measure how many babies are given a name. It measures which name people prefer when forced to choose one. Those are different questions and they often disagree.

What the numbers can't do

The order near the top is less certain than it looks. The leading names are close enough in appeal that separating third from seventh would take far more votes than any site is likely to collect. Treat the top twenty as a group, not a queue.

Names need five head-to-heads before they appear in the rankings at all, and win rates are shrunk toward 50% until a name has a real record — otherwise a name that went 3–0 on its first day would sit above one that has won 340 of 500.

Vote integrity

Each matchup is issued with a signed token. A vote is only counted if it comes back with a valid token for the exact pair we showed, arrives no faster than a human could click, and hasn't been submitted before. Votes are rate-limited per visitor. None of this is unbeatable, but it makes casual ballot-stuffing more effort than it's worth.