Author: slavkoputar

120 players enter a chess tournament. Within a week, there must be a winner — in a way that every player accepts as fair. After sitting in hundreds of tournament halls over the past decade, I can tell you: that is harder to solve than it sounds. Chess has developed three distinct answers to this problem, and each one has its place.

Contents

1. Why tournament format matters
2. Round robin — the fairest format in chess
3. Why round robin does not scale
4. Knockout — the most dramatic format
5. The Swiss system — how most tournaments work
6. Which format fits which event

Why tournament format matters

A chess tournament is not just a collection of games. It is a system designed to find the best player in the field within a fixed amount of time. The format — the rules that determine who plays whom and when — decides whether the result feels legitimate. Get it wrong and players leave feeling cheated. Get it right and the winner is the person who deserved it.

Whichever event you walk into — a club night, a weekend open, an invitational — the format was chosen deliberately, for reasons specific to that event’s size and purpose. A ten-player club championship, a hundred-player weekend open and an eight-player elite invitational cannot use the same rules and produce a fair result — the right format for one would be unworkable for the other two. The three formats used in chess are round robin, knockout, and the Swiss system. They solve the same underlying problem in completely different ways, and each one works best in a specific, fairly narrow context.

Round robin — the fairest format in chess

The most obvious solution is also the most fair. In a round robin, every player faces every other player exactly once. Whoever accumulates the most points wins. There is no pairing algorithm deciding who you face and no possibility of a soft schedule — the winner has beaten, or drawn with, the entire field. That is why round robin results carry more weight than any other format: nobody can argue they got an easy draw.

There is one complication: White moves first and has a proven statistical advantage. In a standard round robin, not every player gets the same number of White games. The solution is the double round robin — everyone plays everyone twice, once as White and once as Black. The Candidates Tournament, which produces the World Championship challenger, uses exactly this format: eight players, fourteen rounds, every colour played equally. Eight players are enough for a meaningful field, and few enough that a double cycle stays manageable.

Why round robin does not scale

Round robin has one fatal problem: mathematics.

Number of Rounds = Number of Players − 1

Eight players need seven rounds. Thirty players need twenty-nine. One hundred players need ninety-nine. A weekend tournament plays five to nine rounds in two or three days — a hundred-player round robin at that pace would take months, not a weekend, which is exactly why nobody uses it at that scale.

Round robin works wherever the field is small and fixed: chess leagues, elite invitationals, norm events of ten players, and Quads — where a large group is divided into pools of four and everyone plays everyone within their pool. All of these share the same trait: the organiser knows the exact field size in advance, and can set the number of rounds accordingly, well before anyone sits down to play. But for open tournaments with dozens or hundreds of participants, entering right up until the last moment, round robin is simply not viable — there is no way to fix the round count for a field that is still growing.

Knockout — the most dramatic format

Knockout is conceptually simple: lose and you are out. Win and you advance. Every game carries maximum stakes. As a spectator format, nothing beats it — the drama is real, the upsets are real, and tension compounds with each round. The Chess World Cup uses knockout for exactly this reason: a bracket of well over a hundred players narrowed down to a single champion.

Matches are seeded so the strongest players cannot meet in round one — top seed against bottom seed, working inward — which keeps the biggest clashes for later rounds, when the stakes are highest for everyone left in the draw.

The real cost of knockout is what it means for players who lose early. In a 64-player field, 32 players go home after round one. At an elite event, that is perhaps two years of qualification effort ended in a single round. At amateur level, it means paying an entry fee, travelling to the venue, losing on day one, and driving home. Would you come back to play that tournament the next year?

The Swiss system — how most tournaments work

Round robin does not scale. Knockout throws players out after one bad round. Swiss solves both problems at once: it runs at any size, and every player plays every round, win or lose.

There is no upper limit on field size. A ten-player club night and a thousand-player open both use the same system — the number of rounds does not need to grow with the number of players, the way it does in round robin. And nobody goes home early: the format guarantees a fixed number of games for everyone, which is exactly what knockout cannot offer.

The trade-off is precision. Round robin proves the winner beat everyone. Knockout proves the winner survived everyone. Swiss proves the winner performed best across a limited sample of opponents — and if that sample is too small, the result gets fuzzy. A large field played over only five or six rounds will often finish with several players tied on the same score, simply because there were not enough rounds to separate them. More rounds means more separation; fewer rounds means more ties. Organisers weigh this deliberately, balancing the number of rounds against the size of the field and the time available.

Read more about the Swiss System.

Which format fits which event

Round robin belongs wherever accuracy matters more than practicality: the Candidates Tournament, elite invitationals, chess leagues, norm events of ten players. Every result there is earned head-to-head, with nothing left to chance in the pairings.

Knockout belongs where spectacle and stakes matter as much as the result: the World Cup, and any qualification event built around the drama of elimination. It produces heroes and upsets in equal measure — and a long list of strong players who lost early and went home.

Swiss is the practical, everyday format — the one that makes large open tournaments possible without asking anyone to play a hundred games or leave after a single loss. It is the format you will encounter at almost every regional, national or open event you enter, from a Tuesday club night to a national championship. If this will be your first time, Your First Chess Tournament covers what to expect. If you have ever wondered why you ended up playing a particular opponent at a big weekend open, you were inside the Swiss system.

If you found this useful, join the Chess Dad list — chess in plain English, direct to your inbox.

How the Swiss system works in chess comes down to three rules that run almost every open chess tournament in the world — yet very few players who compete in one understand how the pairings actually happen behind the scenes. It was designed in 1895 by a teacher named Julius Müller for a tournament in Zurich, and it solved a problem that had no clean solution before: how to run a large event where every player plays every round, nobody goes home early, and the best player still rises to the top. Here is exactly how it works, round by round.

Contents

1. The three rules of the Swiss System
2. Round one — how the Swiss system works in chess
3. Round two — score groups take over
4. Colour assignment — who gets White?
5. Floaters — what happens when score groups are uneven
6. Later rounds — how the system develops
7. Why Swiss is not perfect — and why it is still used

The three rules of the Swiss System

The Swiss system runs on three rules, and everything else in this guide is simply those three rules working themselves out round by round.

Rule one: no rematches. Two players can only meet once in the same tournament, no matter how the scores fall. This is the rule the pairing software checks first, before anything else — it can override colour balance if it has to, but it can never be broken.

Rule two: players with equal scores play each other. After every round, you face someone who has the same result as you — winners play winners, players on half a point face other half-point players. This is what makes Swiss self-correcting: lose round one and your next opponent gets easier, not harder.

Rule three: colour alternation. White and Black are balanced across rounds as evenly as possible. If you had White last round, you are due Black this round, and the pairing software tracks this for every player throughout the event.

Three rules, in that order of priority. Every pairing you will ever see in a Swiss tournament is those three rules resolving themselves against a real field of players.

Round one — how the Swiss system works in chess

Before round one, every player is sorted by rating — strongest at the top, weakest at the bottom. Let me take ten players entering a Swiss tournament as an example. The field is split in half, and the top half plays the bottom half: seed one meets seed six, seed two meets seed seven, and so on down the list. Round one is the only round where rating alone determines the pairing — from round two, score takes over.

Colours alternate by seeding: if the top seed has White, the second seed has Black, the third has White, and so on. Here is how that plays out for a ten-player field:

Board	Pairing	White / alternate
1	Seed 1 vs Seed 6	Seed 1
2	Seed 7 vs Seed 2	Seed 7
3	Seed 3 vs Seed 8	Seed 3
4	Seed 9 vs Seed 4	Seed 9
5	Seed 5 vs Seed 10	seed 5

Pairings are set purely by rating, top to bottom. The only thing arbitrary is who gets White on Board 1. Sometimes this is settled with an actual draw of lots, an arbiter concealing a white pawn in one hand and a black pawn in the other for the top seed to pick. Just as often today, the pairing software decides it automatically the moment the field is set, with no ceremony at all.

Round two — score groups take over

After round one, players are regrouped by score, not rating — this is the shift that defines every round from here on. In a ten-player event where two games were decisive and three ended in draws, the field splits into three groups: two players on one point, six players on half a point, two players on zero.

Each group pairs internally using the exact same method as round one: sort by rating, split the group in half, and pair the top half against the bottom half. The size of the group does not change the method — a group of two pairs the same way as a group of sixty, just with fewer players to sort:

Score group	Players	Pairing logic
1 point	2	The only pair in the group — one game
½ point	6	Split by rating: top 3 play bottom 3 within the score group
0 points	2	The only pair in the group — one game

From here on, score decides who you can face — rating only decides where you sit within that group.

Colour assignment — who gets White?

One principle governs colour assignment from round two onwards: the higher-rated player’s colour history takes priority whenever there is a genuine conflict. If two players in the same score group both had White in round one, both are technically due Black — but only one of them can have it. The higher-rated player’s preference wins: they get Black, and the lower-rated player gets White instead, even though the rule says they were both due the same colour.

The same logic runs in reverse. If both players had Black in round one, both are due White, and again the higher-rated player’s preference is granted first.

In the half-point group from our example, with six players involved, this gets applied player by player. Where a pairing naturally works out — one player due White, one due Black — the software uses it without needing to break any ties. Where both players in a pairing are due the same colour, the higher-rated player’s preference takes priority, and their opponent simply absorbs the mismatch instead.

Floaters — what happens when score groups are uneven

After two rounds, only one player is on two points in our example — having won both games. No other player shares that score, so there is nobody left in their score group to pair against. This is the moment a floater happens.

The floating player drops down and plays the highest-rated player in the one-and-a-half-point group, provided the two of them have not already played each other earlier in the tournament. If that particular pairing would create a rematch, the algorithm skips to the next available player in the group below, and keeps checking until it finds someone who satisfies both the score-group rule and the no-rematch rule.

Floaters happen constantly in Swiss tournaments — every uneven score group produces one. If you are ever paired against someone with a different score to you, this is not a mistake or a glitch in the software. It is the system doing exactly what it is designed to do: resolving an imbalance that is unavoidable.

Later rounds — how the system develops

In rounds three, four and five, the same three rules apply, but the score groups become steadily more refined as results pile up. Players near the top of the standings have faced increasingly strong opposition, because winning repeatedly pushes you into score groups with fewer and fewer players left in them. Players in the middle of the field are meeting opponents with genuinely similar records — which is usually where the closest, hardest-fought games of the tournament happen. The bottom of the draw consists mostly of players who have lost several games in a row and are now facing each other exclusively.

By the final rounds, the leader and the second-place player will usually need to play each other directly — assuming they have not already met earlier — and that single game often decides the entire tournament. This is not a coincidence. The Swiss system is built so that the strongest performers rise through the score groups and collide with each other before the event is over, even though nobody planned that specific pairing in advance.

Why Swiss is not perfect — and why it is still used

The Swiss system has known limitations, and it is worth naming them plainly rather than pretending the format is flawless.

The first is separation: with fewer rounds than players, several people will inevitably finish tied on the same score, because there were not enough games to tell them apart — which is exactly what tiebreak systems exist to fix. The second is complexity at scale: in very large fields, the pairing software occasionally cannot satisfy every rule for every player at once, and has to fall back on a fixed priority order instead.

Limitation	Why it happens	How it is resolved
Rounds too few to separate all players	Fewer rounds than players means ties on score are common	Tiebreak systems rank players after the event
Pairing rules conflict in very large fields	Software cannot satisfy all three rules at once	Priority order: avoiding rematches first, then score group, then colour

None of these limitations have ever been solved better by another format at this scale — which is why Swiss remains the standard worldwide despite them.

If you found this useful, join the Chess Dad list — chess in plain English, direct to your inbox.