Get very rough estimates of the advantage/disadvantage matrix from match data

Combined with est_C_given_A(), this might be useful for warm starting mELO models.

est_A_from_data(data)

Arguments

data	a data.frame of matches. Four columns, time, agent 1, agent 2 and the outcome.

Value

a matrix.

Examples

est_A_from_data(rps_df)
#> Error in group_by(., agent_1, agent_2): could not find function "group_by"
est_A_from_data(rpssl_df)
#> Error in group_by(., agent_1, agent_2): could not find function "group_by"
est_A_from_data(rpsfw_df)
#> Error in group_by(., agent_1, agent_2): could not find function "group_by"
noisy_rpsfw_df <- rpsfw_df
noisy_rpsfw_df$outcome <- add_noise_to_outcomes(noisy_rpsfw_df$outcome, error_prob = 0.5)
est_A_from_data(noisy_rpsfw_df)
#> Error in group_by(., agent_1, agent_2): could not find function "group_by"