Skip to contents

Block-level Bradley-Terry win probabilities from player strengths

Source: R/plotting_functions.R
block_winprob_table_from_lambdas.Rd

Aggregate player-level Bradley-Terry strengths into a block-by-block table of expected win probabilities. The helper can summarise posterior draws via the posterior mean or median, and can optionally weight pairwise probabilities by an observed match-count matrix.

Usage

block_winprob_table_from_lambdas(
  lambda_item,
  z,
  lambda_stat = c("mean", "median"),
  N_ij = NULL,
  include_diag = FALSE,
  diag_value = NA_real_
)

Arguments

lambda_item

Numeric vector of player strengths, or an iterations by players matrix of posterior draws.

z

Integer vector of block memberships for each player.

lambda_stat

Summary to use when lambda_item is a matrix.

N_ij

Optional match-count matrix used to weight pairwise block averages.

include_diag

Logical; if TRUE, compute within-block averages.

diag_value

Value used on the diagonal when include_diag = FALSE.

Value

A list with block win probabilities, contributing pair counts, point-estimated player strengths, and block labels/sizes.