Globle Capitals: How to Play the Capital City Map Game

01What Globle Capitals actually is

Globle Capitals is the capital-city edition of the country-guessing map game. The board looks familiar: a world map, a search box, and a colour that gets warmer as you close in. The difference is the answer. In classic Globle the secret is a country, so guessing France and seeing red tells you something useful about France. In the capitals edition the secret is a capital city, and the map colours according to the distance between the capital you guessed and the capital you are trying to find.

That single change rewrites the whole game. Country borders stop being the unit of thought. You are no longer asking "which country is near the last one?" You are asking "which city is near the last city?" Those are different questions, and on some continents they give very different answers. A country can be enormous and still have its capital tucked into one corner of it. A capital can sit a hundred kilometres from a border while the rest of its country stretches away in the opposite direction. Once you start thinking in cities instead of territories, the map becomes more precise and considerably less forgiving.

Because it is a distance game rather than a border game, Globle Capitals rewards a specific kind of geographical knowledge. You need to know where capitals are, not just which continent a country belongs to. You need a rough mental atlas of capital coordinates, and you need the discipline to distinguish between the city everyone has heard of and the city the government actually sits in. Most of the difficulty in this game is not the map. It is that gap between fame and fact.

If you have never played the country version, start with the free Globle game for a few rounds to internalise the colour system, then come back. The feedback language is shared. Only the target type changes.

02How it differs from classic Globle

Classic Globle treats a country as a single point on the map — usually a centre point — and colours the world by how far your guessed country is from the secret country. Because countries are large, a "close" answer can still be several hundred kilometres off. The game is designed to be a process of continental elimination: you guess something big and central, read the heat, and squeeze inward.

Globle Capitals tightens the geography. Capitals are points, not polygons. The distance between two capitals is unambiguous, and that makes every guess a sharper instrument. A warm colour in the capitals game means you are genuinely near the target city, not merely in the same country or the same region. The flip side is that the game punishes vague thinking. Guessing a country name that is not a capital, or guessing a large famous city that is not the seat of government, gives you either an error or a misleadingly cold result.

Three practical differences follow from this:

• The answer set is a fixed list of capitals, not every country. Many capitals are cities most players cannot place without help. That makes the first guess less informative than it is in classic Globle and makes the endgame more about recall.
• Distance is measured city to city. A guess in a neighbouring country can be red while a guess much further away in the same country is colder, simply because the target capital sits near one border and far from the interior.
• Naming matters. Alternate spellings, former names, and "the biggest city" are the three most common ways to lose a round that you had actually solved in your head.

The strategic consequence is simple: in classic Globle you spread guesses out to triangulate a country; in Globle Capitals you should spread guesses out to triangulate a point. The second is more like navigating than like guessing.

03How the distance and colour feedback works

The core loop is the same as the country game. You type a capital city. The game measures the great-circle distance between that city and the secret capital, converts the distance into a colour band, and paints the guessed location — and often the surrounding landmass — in that colour. Cold colours mean far away. Warm colours mean close. The exact palette and the exact thresholds vary between implementations, so the honest advice is to read the legend on the page you are playing rather than assume a fixed scale.

What stays constant across versions is the shape of the information. A distance colour tells you one number and one direction at a time: it tells you how far, and the map tells you where from. Two guesses give you two circles. Where those circles overlap is where the answer lives. That is the entire geometry of the game, and it is worth stating plainly because most players skip it.

Think of each guess as drawing a ring around a city. If the colour says "about eight hundred kilometres", the target is somewhere on a ring of that radius. The ring is a circle, not a blob, which means it probably crosses several countries. Your next guess should not be chosen because it is interesting. It should be chosen because it will cut that ring into a smaller arc.

There is a second piece of information that is easy to miss: the colour of the guessed city's own country. In many builds, the whole country tints, which lets you see whether you are close to the target capital or merely close to the border of a country whose capital is elsewhere. When the country tints warm but the city marker does not, you are near a border, not near the answer. That distinction is the single most useful reading skill in the game.

04Reading the map like a data layer

Most players look at the map. Better players read it. The difference is whether you treat the colour as decoration or as a measurement with an error bar. Distance bands are coarse by design — a band might span several hundred kilometres — so treat each result as a ring with thickness rather than a precise circle. That thickness is why a single very warm guess is usually not enough, and why two medium guesses placed far apart often beat one perfect-sounding guess placed in the middle of nowhere.

Here is a reading routine that works in any build:

• Note the band, not the shade. Internally label each result "very far", "far", "mid", "near", "very near". Five buckets are enough to reason with and far more stable than trying to remember exact hex colours.
• Note which landmass changed colour. If a whole region warms, the target is near a shared border or a strait. If only a narrow strip warms, you are close to a coastline or a mountain corridor.
• Note what did not change. A guess that leaves an entire continent cold has eliminated that continent more reliably than any list of countries you can recite.
• Write down your rings. On paper or in a notes app: city, band, approximate radius. Two lines of notes beat a heroic memory every time.

The most common reading error is anchoring. You guess a city, get "near", and then spend four guesses circling that city because it feels like the answer must be adjacent. But "near" can mean four hundred kilometres, and four hundred kilometres from Lisbon is not Lisbon's suburbs — it is a different country. Anchor on the ring, not the city.

The second most common error is treating a cold result as useless. A cold result is an elimination. If you guess a capital in Southeast Asia and the map goes deep blue across the whole region, you have just removed a few dozen candidates in one move. Cold guesses are cheap information when they are placed far from your existing rings. They are waste when they are placed next to a guess you already knew was cold.

05A strategy for narrowing a continent

The efficient opening in Globle Capitals is not a single magic city. It is a spread. You want your first two or three guesses to be far apart in a way that divides the world into large, non-overlapping regions. Because capitals cluster on coastlines and plains far more than they cluster in the middle of continents, the useful openers are usually well-known capitals with excellent "distance geometry" — cities from which most of the world is either clearly near or clearly far.

A workable three-guess opening looks like this:

1. One European capital. Europe is dense with capitals, so a European guess immediately classifies a large, high-probability region. It also tends to produce mid-range readings for Africa and western Asia, which is exactly the kind of ambiguous-but-useful signal you want early.
2. One East Asian or Southeast Asian capital. This pulls the opposite end of Eurasia into focus and separates "somewhere in Asia" from "somewhere in Europe" with a single number.
3. One American capital — North or South depending on which you find easier to reason about. This tells you whether the answer is in the Americas at all, and if it is, roughly where.

After three guesses you should be able to name the continent with reasonable confidence. If you cannot, your guesses were probably too close together. Three capitals inside Europe tell you a lot about Europe and almost nothing about the rest of the world.

Once the continent is known, switch from spread to bisection. Pick two capitals at opposite ends of the candidate region and guess the one whose ring will cut the region most evenly. This is the same logic as binary search: each guess should roughly halve the remaining area, not nibble at the edge. If the region is Africa and your two candidates are Cairo and Cape Town, guessing Nairobi or Lusaka splits the continent better than guessing Alexandria, which sits next to a guess you have already made.

In the endgame, stop bisecting and start listing. When your ring is a few hundred kilometres wide, the remaining candidates are a short list of capitals you can enumerate. At that point the game stops being geography and becomes recall. Enumerate the capitals of every country the ring touches, then order them by how well they fit the last two distances. Guessing a country name here is a wasted move. Guessing the largest city is usually a wasted move too — which is the subject of the next section.

If you want a slower, more instructive version of the same discipline, the Globle tips and strategy guide breaks the colour system down in more detail, and the country-game habits transfer directly.

06Thinking about capitals properly

The hardest part of Globle Capitals is not the map. It is the mental model of what a capital is. Most of us learn capitals as trivia pairs — France–Paris, Japan–Tokyo — and those pairs are the easy 30 percent. The remaining 70 percent is where the game lives, and it is full of patterns that trivia lists do not teach.

Planned capitals

Some capitals were built on purpose, from nothing, to be capitals. Brasília, Canberra, Islamabad, Abuja, Naypyidaw, Astana's modern expansion, and Ngerulmud all belong to this family. Planned capitals are usually inland, usually near the geometric centre of the country, and usually not the largest city. That combination is a gift in this game: if the map pulls you inland and away from a coast, and the country is large, a purpose-built capital is a strong hypothesis.

Relocated capitals

Other capitals moved. Kazakhstan moved from Almaty to Astana in the late 1990s. Myanmar moved from Yangon to Naypyidaw in the mid-2000s. Nigeria moved from Lagos to Abuja. Tanzania moved its official seat from Dar es Salaam to Dodoma. Côte d'Ivoire designated Yamoussoukro. Burundi moved to Gitega. In each case, older maps, older textbooks, and older habits still point at the former capital. If a country has a famous coastal metropolis and the map keeps pulling you inland, relocation is a live explanation.

Capitals that are not the largest city

This is the single most productive pattern in the game. Turkey: Ankara, not Istanbul. Australia: Canberra, not Sydney. Brazil: Brasília, not Rio de Janeiro. Canada: Ottawa, not Toronto. Switzerland: Bern, not Zurich. Morocco: Rabat, not Casablanca. Myanmar: Naypyidaw, not Yangon. The United States: Washington, not New York. Pakistan: Islamabad, not Karachi. Vietnam: Hanoi, not Ho Chi Minh City. Ecuador: Quito, not Guayaquil. Cameroon: Yaoundé, not Douala. Belize: Belmopan, not Belize City. New Zealand: Wellington, not Auckland. India: New Delhi, not Mumbai. China: Beijing, not Shanghai. The list goes on, and every entry on it is a trap that has cost someone a streak.

Cities that are also countries

Singapore, Monaco, Vatican City, San Marino, and Luxembourg are places where the capital and the country share a name. In a capitals game these are dangerous in a specific way: they are easy to guess, easy to overlook, and they sit inside the smallest possible distance rings. If your map is warm over a tiny patch of Europe and every normal-sized guess feels wrong, a microstate is worth considering. Just remember that some of these places are so small that the distance feedback has almost no resolution.

Countries with more than one capital

South Africa splits its capital functions across Pretoria, Cape Town, and Bloemfontein. Bolivia has Sucre as the constitutional capital and La Paz as the seat of government. The Netherlands has Amsterdam as the constitutional capital and The Hague as the seat of government. Sri Lanka has Sri Jayawardenepura Kotte as the legislative capital and Colombo as the commercial centre. Malaysia has Kuala Lumpur as the official capital and Putrajaya as the administrative centre. Chile's congress sits in Valparaíso while Santiago is the capital. In each case, the game's answer key picks one, and different builds pick differently. If you are certain you are in the right country but cannot close the last ring, try the other capital before you doubt your geography.

Capitals named after people, or renamed

Nur-Sultan became Astana again. Salisbury became Harare. Lourenço Marques became Maputo. Léopoldville became Kinshasa. Batavia became Jakarta. Constantinople became Istanbul, though it stopped being the capital long before the name changed. Games usually accept the current name, sometimes accept an older one, and almost never accept a name change that happened after the game's dictionary was frozen. If a name is rejected and you are confident it is right, try the modern spelling before you conclude the answer is elsewhere.

07Common capital confusions, in table form

The table below lists pairs that account for a large share of wrong guesses in any capital game. The left column is the country, the middle column is the actual capital, and the right column is the city most often guessed instead. None of these are obscure facts — they are simply facts that fame outranks.

Two observations about this table. First, the pattern is not random: the "wrong" city is almost always the colonial-era port, the industrial powerhouse, or the media capital. Capitals are frequently chosen for defensibility, neutrality between rival regions, or pure administrative convenience, none of which produce the biggest city. Second, the pattern is not universal. Plenty of countries do have their largest city as their capital — Paris, London, Tokyo, Cairo, Buenos Aires, Bangkok, Moscow. Assuming "the famous city is wrong" is just as lazy as assuming it is right. The useful habit is to ask, for each country, whether the capital was chosen for reasons that would have pushed it away from the coast.

08A capital-shaped atlas, continent by continent

You do not need to memorise every capital to play well. You need enough anchors that a warm colour lands on a plausible region instead of a blank space. Here is a compact set of anchors worth having ready.

Europe

The dense western cluster is easy: Lisbon, Madrid, Paris, London, Dublin, Brussels, Amsterdam, Berlin, Bern, Vienna, Rome. The pattern that trips people is the east, where capitals are smaller and more numerous than expected — Ljubljana, Zagreb, Sarajevo, Podgorica, Skopje, Tirana, Pristina, Sofia, Bucharest, Chișinău, Kyiv, Minsk, Vilnius, Riga, Tallinn, Helsinki, Stockholm, Oslo, Copenhagen. If a warm ring lands in the Balkans, the candidate list is long and the distances are short, so precision matters more than spread.

Africa

Africa is where the "largest city" trap does the most damage. Dakar, Accra, Lagos, Nairobi, Addis Ababa, and Cairo are both capitals and giants, which builds a false sense of pattern. Then you meet Yaoundé, Rabat, Dodoma, Gitega, Yamoussoukro, Abuja, and Kigali, and the pattern collapses. A useful mental line runs from Rabat through Algiers, Tunis, Tripoli, Cairo, Khartoum, Addis Ababa, Nairobi, Dodoma, Lusaka, Harare, Gaborone, Windhoek, and down to Pretoria — most African capitals sit closer to that spine than to the coast.

Asia

Asia splits into a dense western band and a sparse northern one. In the west, capitals are close together and often mountainous: Ankara, Tbilisi, Yerevan, Baku, Tehran, Baghdad, Amman, Beirut, Damascus, Riyadh, Kuwait City, Doha, Abu Dhabi, Muscat. In the east, the distances are enormous: Beijing, Seoul, Tokyo, Ulaanbaatar, Bangkok, Hanoi, Phnom Penh, Vientiane, Kuala Lumpur, Jakarta, Manila. A warm result in East Asia narrows slowly because the gaps are so large; a warm result in the Levant narrows almost instantly because the countries are small.

The Americas

South America has a helpful rhythm: the capitals mostly hug the northern and western edges — Caracas, Bogotá, Quito, Lima, La Paz, Santiago, Buenos Aires, Montevideo, Asunción, Brasília, Georgetown, Paramaribo, Cayenne. The interior is empty, which means an inland warm signal in South America is unusually informative. Central America and the Caribbean are the opposite: capitals are coastal, close together, and often share names with their countries, so the ring has to be read carefully. In North America the anchors are Washington, Ottawa, Mexico City, and the Central American chain down to Panama City.

Oceania

Oceania is mostly water, which makes distance feedback dramatic. Canberra and Wellington are the two big anchors. Beyond them, Port Moresby, Honiara, Port Vila, Suva, Apia, Nuku'alofa, Palikir, Majuro, and Ngerulmud are spread across an ocean so large that a mid-range colour can still mean two thousand kilometres. In this region, the practical rule is to trust the colour less and the country list more.

09The spoiler-free hint approach

Half the fun of a daily map game is not being told the answer. But there is a real difference between a spoiler and a hint, and this game is unusually good at supporting the second. A spoiler gives you the capital. A hint gives you a constraint you could have derived yourself but did not have time to.

A good escalating hint ladder looks like this:

1. Region. "The capital is in West Africa." This is a constraint, not an answer, and it usually unlocks two or three plausible rings.
2. Coastal or inland. "The capital is inland." For a continent full of coastal capitals, this single bit eliminates most of the list.
3. Relative size. "The capital is not the country's largest city." This quietly encodes the entire trap table above and is often enough to finish.
4. First letter or syllable count. Use only when you are genuinely stuck, because this converts the game from geography into anagram solving.
5. Country. At this point you have stopped playing the map game, but you still have a capital to recall, so the round retains some value.

The discipline that makes hints work is taking them one at a time. Ask for the smallest constraint that could plausibly change your next guess, then guess again before asking for more. Most players who feel stuck are not missing information; they are refusing to act on the information they have. A hint that confirms something you already suspected is cheap. A hint that arrives before you have used your last two rings is expensive.

There is also a social version of this. If you play with other people, agree in advance on what counts as a hint and what counts as a spoiler, and never announce a country after someone has already guessed it. The daily format is shared, and the etiquette is part of the experience.

10Habits that separate good players from lucky ones

After enough rounds, the players who consistently solve in a handful of guesses are not the ones with the best memories. They are the ones with the best process. Five habits show up again and again.

They guess capitals, never countries. This sounds trivial and is not. The habit of typing the country you are thinking about is strong, especially when the country is famous and the capital is not. Every rejected country name is a lost turn.

They alternate hemispheres early. A spread of three guesses across different parts of the world produces three genuinely independent measurements. Three guesses in Europe produce one measurement repeated three times.

They keep a written ring list. Two lines of notes turn a fuzzy memory of colours into a solvable geometry problem. This matters most in the middle of a round, when the map is full of tints and you cannot remember which guess produced which band.

They separate "near the capital" from "near the country". Border proximity and capital proximity are different things, and the tinted country layer is where the distinction becomes visible. Players who read the country tint as if it were a capital distance tend to overshoot into neighbouring countries.

They review the miss. When a round ends, the useful question is not "what was the answer?" but "which guess added the least information?" That is the guess to replace next time. Over a month, this habit improves more than memorising a capital list.

If you want to put the process into practice without waiting for tomorrow's puzzle, the unlimited Globle practice mode runs the same map logic on demand. Play three rounds with the three-guess spread, then three rounds with pure bisection, and notice which style suits you. You can also check the daily Globle answer page after you have finished, but only after — reading it first removes the only part of the game that teaches you anything.