What Is a Round Robin? Everyone Plays Everyone, and the Problems Start There

The two basic tournament shapes make opposite trades. A knockout eliminates half the field each round, so it needs very few matches, produces steadily higher stakes and reaches a single winner quickly, and it is a poor measure of overall strength since one bad day removes the best competitor and the bracket determines who meets whom. A round robin needs a number of matches that grows with the square of the field, so it is far more expensive, and it measures strength well since every competitor faces the same opposition and a single poor result does not eliminate anyone. That makes round robins standard where accuracy matters and the field is small, including league seasons, group stages and chess tournaments, and knockouts standard where the field is large or the schedule short. Hybrids are the usual compromise, with group stages ranking a field into a knockout, which is the structure of most major international tournaments.

The format has known weaknesses and each has produced rule changes:

• •Dead rubbers, where a competitor has already qualified or been eliminated and the final matches no longer matter to them, which affects the opponents who still need a result
• •Collusion, where a particular scoreline suits both sides, which produced the notorious 1982 match between West Germany and Austria and led directly to final group fixtures being played simultaneously
• •Deliberate losing to secure an easier knockout draw, documented in football and in badminton at the 2012 Olympics where eight players were disqualified
• •Scheduling asymmetry, since facing a strong opponent early or late, or immediately after they have played a hard match, is not the same thing
• •Tie-breaking, which is unavoidable and always arbitrary to some degree, using goal difference, head-to-head results, points scored or drawing lots
• •Incomplete round robins, where withdrawal or abandonment leaves competitors having played different opponents, which invalidates the comparison the format exists to provide

Producing a fixture list where everyone plays everyone once, with nobody playing twice in a round and with home and away balanced, is a genuine combinatorial problem with a neat classical solution. The circle method fixes one competitor in place and rotates the others around them, generating a valid schedule for any even number of entrants, with an odd number handled by adding a bye. Real leagues add constraints that make it considerably harder, including television requirements, stadium availability, policing considerations that prevent certain fixtures coinciding, travel minimisation, and rules preventing long runs of home or away matches, and modern league schedules are produced by constraint solvers rather than by hand. The resulting schedules are argued about every season, since no arrangement is neutral, and studies have found measurable effects from fixture ordering on final standings, which means the scheduler makes decisions that affect outcomes however carefully they try not to.

The structure appears wherever a set of options must be compared pairwise. Voting systems of the Condorcet family conduct a round robin among candidates, comparing each pair and looking for one that beats all others, which is an appealing criterion that sometimes has no winner because the pairwise results form a cycle, with A beating B, B beating C and C beating A. That possibility, identified in the eighteenth century, is a fundamental result about collective preference and it appears in sporting round robins too, where three competitors beat each other in a loop and any ranking between them is arbitrary. Round robin scheduling also names a processor allocation strategy in computing, where each task receives a fixed time slice in turn, chosen for fairness rather than efficiency. The recurrence of the same structure and the same cycle problem across voting, sport and computing reflects that they are all solving the problem of turning pairwise comparisons into a single order.

Everyone meeting everyone measures strength far better than a knockout and costs a number of matches growing with the square of the field. Dead rubbers and collusion are the structural weaknesses, and simultaneous final fixtures exist because of a 1982 match both sides were content to fix. Pairwise results can form a cycle in which any ranking is arbitrary, which is the same problem that appears in voting theory.