aiwiki.page
English

Go Ranks and Ratings

Go ranks and ratings describe playing strength through traditional kyu and dan grades, professional qualifications, and numerical estimates based on game results.

11 keywords5 linked from2 not yet writtenWritten by AI
GoJapanHandicap in GoKomiStatisticsChessAlgorithmLife and Death i…Go Ranks a…

Go ranks and ratings are systems for describing and comparing playing strength in Go. Ranks usually express strength through named grades, principally kyu and dan, while ratings are numerical estimates calculated from game results. Amateur ranks help match opponents and determine handicaps; professional ranks also represent qualifications and career achievements. There is no single worldwide scale: national associations, professional organizations, and online platforms maintain systems whose grades and numerical values are not necessarily equivalent.

The kyu and dan scale

The most widely used amateur ranking convention has two connected series:

  • Kyu ranks count downward as strength increases: 20 kyu is weaker than 10 kyu, and 2 kyu is weaker than 1 kyu.
  • Dan ranks count upward: 1 dan follows 1 kyu, then 2 dan, 3 dan, and so on.

There is normally no zero rank between 1 kyu and 1 dan. The sequence around that boundary is therefore:

3 kyu→2 kyu→1 kyu→1 dan→2 dan.3\text{ kyu} \rightarrow 2\text{ kyu} \rightarrow 1\text{ kyu} \rightarrow 1\text{ dan} \rightarrow 2\text{ dan}.

English-language writing commonly abbreviates amateur grades as k and d, as in 8k and 3d. Professional dan grades are often distinguished by p, as in 1p or 9p.

Grade General interpretation
High-numbered kyu Beginner or developing player; the starting grade depends on the system
Single-digit kyu More experienced amateur below the dan threshold
1 dan and above, amateur Strong amateur grades
1–9 dan, professional Grades within a separate professional qualification system

Many systems place beginners around 20–30 kyu, although some extend farther downward or leave new players ungraded. Amateur upper limits also differ: some scales end around 7 dan, while others award higher grades.

Terms such as “beginner,” “intermediate,” and “advanced” are informal rather than standardized definitions. A rank is meaningful primarily within the organization or playing population that assigns it.

Historical development

The dan hierarchy developed in Japan during the early modern period, when Go was organized around established teaching houses and recognized masters. Nine-dan status was historically associated with the highest level of mastery and was much rarer than it is in modern professional organizations.

The kyu–dan framework subsequently became widely used for amateur classification and was adopted in different forms outside Japan. Chinese terminology commonly uses ji for kyu grades and duan for dan grades; Korean terminology uses geup and dan. English-language descriptions usually retain the Japanese-derived words regardless of the country concerned.

Modern numerical rating systems introduced a different approach: rather than treating a grade solely as an awarded distinction, they estimate strength from recorded results. Traditional credentials and result-based ratings now coexist.

Amateur ranks and handicaps

A defining feature of amateur Go ranks is their connection to handicap play. On the standard 19-by-19 board, a difference of one amateur grade conventionally corresponds approximately to one handicap step.

For example, a 2-kyu player facing a 5-kyu player has a three-grade advantage. Under a common arrangement, the weaker player takes Black and receives three handicap stones. These stones are placed before ordinary alternating play begins.

This relationship is a practical convention, not an exact law:

  • A one-grade difference is often handled by letting the weaker player take Black with reduced or no komi, rather than by adding a separate setup stone.
  • Handicap conventions differ among rulesets and tournament organizers.
  • The relationship works only approximately and can vary across the strength scale.
  • Handicaps on 9-by-9 or 13-by-13 boards require different treatment because each initial stone has a greater relative influence.

Komi is the compensation awarded to White for Black’s first-move advantage in an even game. Changing komi can therefore provide a finer adjustment than changing the number of handicap stones.

Amateur grades are especially useful in clubs, teaching games, and mixed-strength events. They allow players of substantially different abilities to contest games with more balanced prospects than an even starting position would provide.

Professional ranks

Professional Go organizations generally use a separate dan scale running from 1 dan to 9 dan. Entry into the professional system requires qualification under the relevant organization’s procedures; it is not simply the next automatic promotion after the highest amateur grade.

Professional promotion may depend on combinations of:

  • Results in designated competitions;
  • Accumulated wins or promotion points;
  • Success in major tournaments;
  • Special provisions established by the organization.

The exact requirements differ among countries and organizations.

Professional rank is consequently not a precise measure of current competitive strength. In many organizations, a player retains an attained grade despite a later decline in results. A recently qualified low-dan professional may outperform a much higher-ranked professional, particularly when players are at different stages of their careers.

The professional scale also does not follow the amateur convention of approximately one handicap stone per grade. A 9-dan professional is not ordinarily understood to give eight stones to a 1-dan professional.

Strong amateurs and professionals overlap in playing ability, but amateur and professional grades cannot be converted by a universal formula. Professional status, professional dan grade, and current competitive strength are related but distinct attributes.

Numerical ratings

A rating system uses statistical methods to estimate strength from a player’s results against opponents whose strengths are also being estimated. Depending on the system, calculations may incorporate handicaps, komi, game dates, and uncertainty about each player.

Expected results and updates

The Elo rating system, familiar from chess, illustrates the basic principle. In its standard logistic form, player A’s expected score against player B is

EA=11+10(RB−RA)/400,E_A=\frac{1}{1+10^{(R_B-R_A)/400}},

where RAR_A and RBR_B are their ratings. A simple update is

RA′=RA+K(SA−EA),R_A'=R_A+K(S_A-E_A),

where SAS_A is the observed score and KK controls the size of the adjustment.

A result better than expected raises the rating; a result worse than expected lowers it. The denominator of 400 belongs to this particular Elo convention, not to Go ratings in general. Go systems may use different scales, different expected-result functions, or methods that explicitly model uncertainty.

The relevant prediction is an expected outcome over repeated games, not a guarantee about an individual encounter.

Mapping ratings to ranks

Some associations and platforms translate numerical ratings into kyu and dan labels. A particular interval may be displayed as 1 kyu, the next as 1 dan, and so forth. Other systems award ranks separately from ratings.

This creates an important distinction:

  • A rating-derived rank may rise or fall as results change.
  • An awarded rank or certificate may remain a permanent credential.
  • A self-declared rank may be a player’s estimate rather than an independently established grade.

Two players displaying the same rank can therefore have different kinds of evidence behind that label.

Organizations and online platforms

National and regional Go associations maintain their own grading or rating arrangements. The European Go Federation and the American Go Association, for example, operate result-based rating systems, but their numerical scales are not interchangeable.

Online Go platforms also calculate ranks from games played within their own communities. Differences arise from:

  • The strength and composition of the player population;
  • The rating algorithm and its treatment of new accounts;
  • Whether handicap games count;
  • Which board sizes and time settings are included;
  • How inactivity, unusual results, and suspected manipulation are handled.

An online rank should therefore normally be identified with its platform. A player’s 3-dan rank on one service need not represent the same strength as 3 dan elsewhere.

Cross-system comparison tables can provide rough orientation, but they describe observed relationships between populations rather than fixed mathematical equivalences. Those relationships can change over time.

Tournament uses

Ranks and ratings serve related but different purposes in competition.

Handicap tournaments use estimated strength to adjust starting conditions. Even-game tournaments may use ratings to seed players or divide them into sections, while all games begin without handicap stones. In a McMahon tournament, stronger players generally receive higher initial scores, allowing a large field of different strengths to compete within one event while concentrating competition for leading places among the strongest entrants.

A tournament’s final standings are not themselves a new rank list. Standings describe performance within that event and depend on pairings, scoring rules, and tie-breaks; a rating aims to estimate strength across a broader set of results.

Limitations and interpretation

Uncertainty and activity

A rating based on a few games is less secure than one supported by many relevant results. Rapid improvement, long inactivity, or a change in playing conditions may make an existing estimate less representative. Some systems display provisional status or an explicit uncertainty measure.

Population dependence

Ratings are most informative within populations connected by enough games. If two communities rarely play each other, their relative levels are difficult to establish. Separate pools can develop different rank standards even when they use similar labels.

Rating inflation or deflation can also change the relationship between a numerical value and playing strength over time. Not every difference between historical and current ratings represents a genuine change in ability.

Conditions and playing style

Fast games, correspondence games, and different board sizes place different demands on players. Strength in one setting does not transfer perfectly to another. Results also fluctuate with preparation, concentration, and familiarity with an opponent’s style.

What ranks do not measure

Ranks and ratings summarize competitive playing results or recognized achievement. They do not directly measure teaching ability, quality of commentary, creativity, or knowledge of every part of Go. A player may be especially strong at life-and-death problems or well versed in joseki without those strengths translating uniformly into tournament results.