EloRatingModel

Remarks

Right for 1v1, and right for any game whose players will ask. Doc 28 names transparency as the reason it ships alongside a better model rather than being replaced by one — a rating a player can compute themselves is a rating they will argue with rather than about.

⚠ A team's rating is the mean of its players', not the sum. A sum makes a five-player team five times as strong as one player and matches it against nobody; the mean is what the expectation formula's four-hundred-point scale is calibrated for.

Fields and properties (3)

• public double K

  How far one result may move a rating.

• public string Name

  What it is called, in a queue definition and in a report.

• public Rating Starting

  What somebody starts on.

Methods (5)

• public EloRatingModel(double k = 32, double start = 1500)

  Makes one.

• public static double Expected(double rating, double opponent)

  What one rating is expected to score against another, from zero to one.

• public double AfterPeriod(double rating, IReadOnlyList<double> opponents, double score)

  What a rating becomes after a set of games treated as one period.

• public IReadOnlyList<Rating[]> Update(IReadOnlyList<Rating[]> teams, IReadOnlyList<int> ranks)

  Updates every team's ratings after a result.

• public double Quality(IReadOnlyList<Rating[]> teams)

  How even a match would be, from zero to one.

Used by (2)

• DuelsVixen.Live.Matchmaking.Tests
• EloTestsVixen.Live.Matchmaking.Tests