pred.simulation.K: Predictive Simulation of Streakiness Assuming Consistent...

In bayesball/BayesTestStreak: Testing Streakiness in Binary Sequences by Bayes Factors

Predictive Simulation of Streakiness Assuming Consistent Model

Description

Implements a predictive simulation of the streakiness behavior for a group of binary sequences under a consistent model. Underlying probabilities of success are estimated using an exchangeable model, binary sequences are simulated using these constant probability values, and the log Bayes factors in support of streakiness are computed for all sequences.

Usage

pred.simulation.K(data, log.K)

Arguments

data	matrix with 2 columns – first column are the numbers of successes, and the second column contain the corresponding sample sizes
log.K	value of log K defining the streaky model for the Bayes test comparing the consistent and streaky models

Value

vector of values of the log Bayes factor in support of streakiness

Author(s)

Jim Albert

References

Albert, J. (2013), Looking at Spacings to Access Streakiness, Journal of Quantitative Analysis of Sports, Vol 9, Issue 2, 151-163.

Examples

y <- rbinom(10, size=500, 
     prob=c(rep(0.1, 3), rep(0.3, 4), rep(0.5, 3)))
data <- cbind(y, 500)
pred.simulation.K(data, 5)

bayesball/BayesTestStreak documentation built on April 28, 2024, 6:28 a.m.