swiss-pairing

swiss-pairing is a tiny swiss pairing library with basic deterministic functionality

Installation

• npm install --save swiss-pairing
• require('swiss-pairing')(options)
  • options contains the following variables:
  • maxPointsPerRound - the number of points a participant can score in a round - usually 1 or 2
    • default: 1
  • rematchWeight - rematch penalty weight
    • default: 100
  • standingPower - the power to which standings differences should be raised
    • default: 2
  • seedMultiplier - the deterministic PRNG seed multiplier, ideally a prime of at least 4 digits
    • default: 6781
  • See test/test.js for usage example

Usage

swiss-pairing exposes the following three methods:

getMatchups(round, participants, matches)

See participants and matches formats below.

Determines the matchups for the given round by pairing participants:

• who have not faced each other
• in order of:
  • highest standing to lowest
  • highest modified median score to lowest
  • lowest seed to highest
• where the highest ranked team in a pair is given the home side

When byes are needed (in the case of an odd number of participants), they bubble up from the lowest to the highest ranking, (starting with the lowest seed when no match history is available). participants cannot have more than one bye.

Matchups returned are in the form:

[
  {
    'home': home_participant_id,
    'away': away_participant_id
  },
  ...
]

getStandings(round, participants, matches)

See participants and matches formats below.

Determines the standings for a given round by accumulating won points and lost points and calculating modified median scores. Standings are ordered by wins, modified median score, and (inverse) seed in that order.

Standings returned are in the form:

[
  {
    'id': participant_id,
    'seed': seed,
    'wins': won_points,
    'losses': lost_points,
    'tiebreaker': modified_median_score
  },
  ...
]

getModifiedMedianScores(round, participants, matches)

See participants and matches formats below.

Caculates modified median scores based on the given participants and paired match history.

Scores returned are in the form:

{
  <participant_id>: modified_median_score,
  ...
}

participants argument

The participants argument expects an array in the form:

[
  {
    'id': participant_id,
    'seed': participant_seed
  }
]

• participant_id may be any value that exposes a toString method (and can therefore be used as a key on a javascript object)
• participant_seed may be any directly sortable unique value, although numeric values (1..N) are suggested for reliability

matches argument

The matches argument expects an array in the form:

[
  {
    'round': match_round,
    'home': {
      'id': home_participant_id,
      'points': home_won_points
    },
    'away': {
      'id': away_participant_id,
      'points': away_won_points
    }
  }
]

• match_round must be a value sortable against the given currentRound argument - all matches will be limited to those where the round is less than the currentRound
• home_participant_id and away_participant_id may be any values that javascript is capable of using as an object key and must exist in the given teams argument
• home_points_won and away_points_won should be numeric values, but currently can be anything that can be accumulated against 0 + ... and compared against numeric values - there are also no checks against whether or not these values overrun maxPerRound and any data that does overrun maxPerRound will likely result in strange or erroneous results from every function in this library

Contributors

• @dambrisco
• @philkjacobs
• @kabobrocks