openskill package

Subpackages

• openskill.models package
  • Submodules
  • openskill.models.bradley_terry_full module
    • BradleyTerryFull
      • BradleyTerryFull.calculate()
  • openskill.models.bradley_terry_part module
    • BradleyTerryPart
      • BradleyTerryPart.calculate()
  • openskill.models.plackett_luce module
    • PlackettLuce
      • PlackettLuce.calculate()
  • openskill.models.thurstone_mosteller_full module
    • ThurstoneMostellerFull
      • ThurstoneMostellerFull.calculate()
  • openskill.models.thurstone_mosteller_part module
    • ThurstoneMostellerPart
      • ThurstoneMostellerPart.calculate()
  • Module contents

Submodules

openskill.constants module

class openskill.constants.Constants(**options)

Bases: object

openskill.constants.beta(**options) → float

openskill.constants.beta_squared(**options) → float

openskill.constants.epsilon(**options) → float

openskill.constants.mu(**options) → float

openskill.constants.sigma(**options) → float

openskill.constants.z(**options) → float

openskill.rate module

class openskill.rate.Rating(mu: Optional[float] = None, sigma: Optional[float] = None, **options)

Bases: object

The storehouse of the skill and confidence the system has in an agent. Stores the mu and sigma values of an agent.

Parameters

• mu – The mean skill of the agent.

• sigma – How confident the system is the skill of an agent.

• options – Pass in a set of custom values for constants defined in the Weng-Lin paper.

openskill.rate.create_rating(rating_list: List[Union[int, float]]) → Rating

Create a Rating object from a list of mu and sigma values.

Parameters

rating_list – A list of two values where the first value is the mu and the second value is the sigma.

Returns

A Rating object created from the list passed in.

openskill.rate.ordinal(agent: Union[Rating, list, tuple], **options) → float

Convert mu and sigma into a single value for sorting purposes.

Parameters

• agent – A Rating object or a list or tuple of float objects.

• options – Pass in a set of custom values for constants defined in the Weng-Lin paper.

Returns

A float object that represents a 1 dimensional value for a rating.

openskill.rate.predict_draw(teams: List[List[Rating]], **options) → Union[int, float]

Predict how likely a match up against teams of one or more agents will draw. This algorithm has a time complexity of O(n!/(n - 2)!) where ‘n’ is the number of teams.

Parameters

teams – A list of two or more teams, where teams are lists of Rating objects.

Returns

A float that represents the probability of the teams drawings.

openskill.rate.predict_win(teams: List[List[Rating]], **options) → List[Union[int, float]]

Predict how likely a match up against teams of one or more agents will go. This algorithm has a time complexity of O(n!/(n - 2)!) where ‘n’ is the number of teams.

Parameters

teams – A list of two or more teams, where teams are lists of Rating objects.

Returns

A list of probabilities of each team winning.

openskill.rate.rate(teams: List[List[Rating]], **options) → List[List[Rating]]

Rate multiple teams consisting of one of more agents. Order of teams determines rank.

Parameters

• teams – A list of teams, where teams are lists of Rating objects.

• rank – A list of int where the lower values represent the winners.

• score – A list of int where higher values represent the winners.

• tau – A float that modifies the additive dynamics factor.

• prevent_sigma_increase – A bool that prevents sigma from ever increasing.

• options – Pass in a set of custom values for constants defined in the Weng-Lin paper.

Returns

Returns a list of Rating objects.

Note

tau will default to 25/300 in the next major version update.

openskill.rate.team_rating(game: List[List[Rating]], **options) → List[List[Union[int, float]]]

Get the whole rating of a list of teams.

Parameters

• game – A list of teams, where teams are lists of Rating objects.

• options – Pass in a set of custom values for constants defined in the Weng-Lin paper.

Returns

Returns a list of lists containing mu, sigma, the team object and the current rank of the team.

openskill.statistics module

openskill.statistics.phi_major(x)

openskill.statistics.phi_major_inverse(x)

openskill.statistics.phi_minor(x)

openskill.statistics.v(x, t)

openskill.statistics.vt(x, t)

openskill.statistics.w(x, t)

openskill.statistics.wt(x, t)

openskill.util module

openskill.util.gamma(**options)

openskill.util.ladder_pairs(ranks: List[int])

openskill.util.rankings(teams, rank: Optional[List[int]] = None)

openskill.util.score(q, i) → float

openskill.util.transpose(xs)

openskill.util.unwind(ranks, teams) → Tuple[List, List[int]]

openskill.util.util_a(team_ratings)

openskill.util.util_c(team_ratings, **options)

openskill.util.util_sum_q(team_ratings, c)

Module contents