TrueSkillThroughTime (version 1.0.0)

Game: Game

Description

Game class

Usage

Game(teams, result = vector(), p_draw = P_DRAW)
posteriors(g)
# S4 method for Game
posteriors(g)

Value

Game object

Arguments

teams

A list of Player objects. Each position represents a team, so it must contain a vector of Player objects.

result

A vector of numbers, with the score obtained by each team, or an empty vector. The default value is an empty vector. In this case, the outcome is defined by the order in which the teams list was initialized: the teams appearing firstly in the list defeat those appearing later (no ties). If the list is not empty, it must have the same length as the teams list. In this last case, the team with the highest score is the winner, and the teams with the same score are tied.

p_draw

A number, the probability of a draw. The default value is P_DRAW = 0. A rule of thumb states that the probability of a draw must be initialized with the observed frequency of draws. If in doubt, it is a candidate parameter to be optimized or integrated by the sum rule. It is used to compute the prior probability of the observed result, so its value may affect an eventual model selection task.

g

A game object

Examples

a1 = Player(Gaussian(mu=0, sigma=6), beta=1, gamma=0.03)
a2 = Player(); a3 = Player(); a4 = Player()
team_a = c(a1, a2)
team_b = c(a3, a4)
teams = list(team_a, team_b)

g = Game(teams) 
post = posteriors(g)
lhs = g@likelihoods
post[[1]][[1]] == lhs[[1]][[1]]*a1@prior 
ev = g@evidence
ev == 0.5

ta = c(a1)
tb = c(a2, a3)
tc = c(a4)
teams_3 = list(ta, tb, tc)
result = c(1, 0, 0)
g3 = Game(teams_3, result, p_draw=0.25)