同一厩舎の多頭数出しで人気薄馬の成績を調べる

analyze.R

require(RSQLite)

drv <- dbDriver('SQLite')
conn <- dbConnect(drv, dbname = '..\\netkeiba-scraper\\race.db')
rs <- dbSendQuery(conn, '
  select * from race_result where cast(order_of_finish as int) <> 0
')

allData <- fetch(rs, n = -1)

dbClearResult(rs)
dbDisconnect(conn)

allData$stable <- factor(allData$stable)

allData <- split(allData, list(allData$race_id, allData$stable))

allData <- Filter(function(df) nrow(df) > 0, allData)

allData <- Map(function(df) {
  df <- df[order(df$popularity),]
  tail(df, 1)
}, allData)


allData <- do.call(rbind, allData)

mean(ifelse(allData$order_of_finish == 1, allData$odds, 0))
# => 0.4750358
#人気薄の方が回収率が低い