OxBoatRace {glarma}

Oxford-Cambridge Boat Race

Description

Results of the boat race between Oxford and Cambridge from 1829–2011.

Usage

data(OxBoatRace)

Format

A data frame containing the following columns:

[, 1]	Year	Year in which the race occurred. Some years are missing when the race was not run.
[, 2]	Intercept	A vector of ones, providing the intercept in the model.
[, 3]	Camwin	A binary response, zero for an Oxford win, one for a Cambridge win.
[, 4]	WinnerWeight	Weight of winning team's crew.
[, 5]	LoserWeight	Weight of losing team's crew.
[, 6]	Diff	Difference between winning team's weight and losing team's weight.

Source

Klingenberg, Bernhard (2008) Regression models for binary time series with gaps. Computational Statistics & Data Analysis, 52, 4076–4090.

Examples

### Example with Oxford-Cambridge Boat Race
data(OxBoatRace)

y1 <- OxBoatRace$Camwin
n1 <- rep(1, length(OxBoatRace$Year))
Y <- cbind(y1, n1 - y1)
X <- cbind(OxBoatRace$Intercept, OxBoatRace$Diff)
colnames(X) <- c("Intercept", "Weight Diff")

oxcamglm <- glm(Y ~ Diff + I(Diff^2),
                data = OxBoatRace,
                family = binomial(link = "logit"), x = TRUE)
summary(oxcamglm)

X <- oxcamglm$x

glarmamod <- glarma(Y, X, thetaLags = c(1, 2), type = "Bin", method = "NR",
                    residuals = "Pearson", maxit = 100, grad = 1e-6)

summary(glarmamod)
likTests(glarmamod)

## Plot Probability of Cambridge win versus Cambridge Weight advantage:
beta <- coef(glarmamod, "beta")
par(mfrow = c(1, 1))
plot(OxBoatRace$Diff, 1 / (1 + exp(-(beta[1] + beta[2] * OxBoatRace$Diff +
                                       beta[3] * OxBoatRace$Diff^2))),
     ylab = "Prob", xlab = "Weight Diff")
title("Probability of Cambridge win \n versus Cambridge weight advantage")

## Residuals and fit plots
par(mfrow=c(3, 2))
plot.glarma(glarmamod)