getLinearWeightsModel: getLinearWeightsModel

In frogman141/openWAR: Machinery for analyzing baseball data and computing WAR

Description

Retrieves a linear weights model object

Usage

getLinearWeightsModel(type)

Arguments

type	A string representing either 'xr', 'woba'

Details

Retrieves any of several linear weights formulae as lm objects

Value

A lm object

Examples

ds = getData()
event = data.frame(event = ds$event)
woba = getLinearWeightsModel(type = 'woba')
xr = getLinearWeightsModel(type = 'xr')
ds$woba = predict(woba, newdata=event)
ds$xr = predict(xr, newdata=event)

braa = getLinearWeightsModelfromData(ds, type = 'bp')
bp.events = levels(as.factor(braa$model$event))
event[which(!(event %in% bp.events))] <- NA
ds$braa = predict(braa, newdata=event)
leaders = ddply(ds, ~batterId, summarise, Name = batterName[1], PA = sum(isPA)
, WOBA = sum(woba) / sum(isPA == TRUE), XR27 = sum(xr) * 25.5 / sum(isPA & !isHit))
# The top 20
head(leaders[order(leaders$WOBA, decreasing = TRUE),], 20)
# meeting a plate appearance minimum
qualified = subset(leaders, PA > length(unique(ds$gameId)) * 3.1 / 15)
# WOBA leaders
head(qualified[order(qualified$WOBA, decreasing = TRUE),], 20)
# XR27 leaders
head(qualified[order(qualified$XR27, decreasing = TRUE),], 20)
# Compare the two metrics
xyplot(XR27 ~ WOBA, data = qualified)
with(qualified, cor(WOBA, XR27))

frogman141/openWAR documentation built on Dec. 20, 2021, 8:52 a.m.