WTC 2017 Analysis

nationalityRating.csv

wtc_predictions_2017.R

# master script to make predictions

library(dplyr)
library(jsonlite)
library(WTCTools)

players2017 <- read_json("World_Team_Championship_2017.json")
# matrix is efficient for data extraction
players17 <- matrix(NA, nrow = length(players2017) * 5,
    ncol = 7, 
    dimnames = list(NULL, 
        c("team", "player", "faction", "caster1", "caster2", "theme1", "theme2")))

for (i in seq_along(players2017)) {
    team <- matrix(NA, nrow = 5, ncol = 7)
    dat <- players2017[[i]]
    team[, 1] <- dat$name
    # extract and transpose
    team[, -1] <- t(vapply(dat$players[1:5], 
        FUN = function(x) c(x$name, x$faction, x$list1$list[[1]], x$list2$list[[1]], 
            ifelse(is.null(x$list1$theme), yes = "", no = x$list1$theme), 
            ifelse(is.null(x$list2$theme), yes = "", no = x$list2$theme)), 
        FUN.VALUE = character(6)))
    players17[seq.int(((i - 1) * 5) + 1, i * 5), ] <- team
}

# to data.frame
players17 <- as.data.frame(players17, stringsAsFactors = FALSE)

# check names
players17 <- mutate(players17, 
    caster1 = harmonizeCasters(caster1), 
    caster2 = harmonizeCasters(caster2))

# remove nicknames
grep(" \\\".+\\\" ", x = players17$player, value = TRUE)
players17$player <- gsub(" \\\".+\\\" ", x = players17$player, replacement = " ")

players17$player[players17$player == "Anthony O Reilly "] <- "Anthony O'Reilly"
players17$player[players17$player == "Dmtrii Tochenov"] <- "Dmitrii Tochenov"
players17$player[grep(".*Mader", x = players17$player)]

write.csv(players17, file = "players2017.csv", row.names = FALSE)
players17 <- read.csv(file = "players2017.csv", stringsAsFactors = FALSE)

###############################################################################

# get ratings

###############################################################################

# moved to wtc_predictions_2017_player_reg.R
#source("wtc_predictions_2017_update_rating.R")
load("rating2015-reg16.rda")

wtc$allrounds <- wtc$round + (wtc$year - 2013) * 6 

###########################################################

# ratings based on pairings with selected caster pairings
rating2016 <- steph(x = select(filter(wtc, year == 2016), 
        allrounds, player1, player2, TP), 
    status = rating2015$ratings)

save(rating2016, file = "rating2016.rda")

rankings <- rating2016$ratings
write.csv(rankings, file = "rankings2016.csv", row.names = FALSE)

###########################################################

# register players
source("toPlain.R")
players17$player <- toPlain(players17$player)
players17$player <- gsub(pattern = " \\\".+\\\"", 
    x = players17$player, replacement = "")
players17$player <- gsub(pattern = " '.+'$", 
    x = players17$player, replacement = "")
players17$player <- gsub(pattern = " '[A-Za-z ]+' ", 
    x = players17$player, replacement = " ")
players17$player <- gsub(pattern = "\\\"[A-Za-z ]+(''|\\\" )", 
    x = players17$player, replacement = " ")
players17$player <- gsub(pattern = " \\(.+\\)$", 
    x = players17$player, replacement = "")

# remove all caps
ind <- grepl("^[A-Z -]+$", x = players17$player)
players17$player[ind] <- tools::toTitleCase(casefold(players17$player[ind]))

newPlayers <- players17$player[!players17$player %in% 
        rating2015$ratings$Player]
newPlayers

# replace names in players17 with correct similar names
pNames <- showSimilarNames(c(newPlayers, rating2015$ratings$Player))
# keep matched name pairs
pNames <- pNames[c(10, 11, 13, 14, 15, 16, 17, 20, 21, 32, 33, 34, 35), ]
# align replacements
bool <- pNames[, 2] %in% rating2015$ratings$Player
for (i in seq_along(bool)) if (bool[i]) pNames[i, ] <- rev(pNames[i, ])
# "Georg Buscha" %in% rating2015$ratings$Player
# replace values in second column with values in first 
for (i in seq_len(nrow(pNames))) {
    players17$player[players17$player == pNames[i, 2]] <- pNames[i, 1]
}

players17$player[players17$player == "kerazoglou  nikos"] <- "Kerazoglou Nikolaos"
players17$player[players17$player == "Victor Pons"] <- "Victor Pons Tec"

write.csv(players17, file = "players2017.csv", row.names = FALSE)
players17 <- read.csv(file = "players2017.csv", stringsAsFactors = FALSE)

###########################################################

# check players in rating
p17 <- unique(players17$player)

knownp <- p17[p17 %in% rating2015$ratings$Player]
length(knownp)
unknownp <- p17[!p17 %in% rating2015$ratings$Player]
length(unknownp)

i = 3
pp17 <- grep(pattern = paste0("^", LETTERS[i]), x = unknownp, 
    value = TRUE)
pp15 <- grep(pattern = paste0("^", LETTERS[i]), 
    x = rating2015$ratings$Player[!rating2015$ratings$Player %in% knownp], 
    value = TRUE)
nm <- matrix(NA, nrow = max(length(pp17), length(pp15)), ncol = 3)
nm[seq_along(pp17), 1] <- sort(pp17)
nm[seq_along(pp15), 2] <- sort(pp15)
rownames(nm) <- sort(pp15)
nm[rating2015$ratings$Player[rating2015$ratings$Player %in% pp15], 3] <- rating2015$ratings$country[rating2015$ratings$Player %in% pp15]
rownames(nm) <- NULL
nm

players17$country <- gsub(pattern = "Team ", replacement = "", 
    x = players17$team)
players17$country <- gsub(pattern = "Epic ", replacement = "", 
    x = players17$country)
players17$country <- gsub(pattern = "Prime ", replacement = "", 
    x = players17$country)
players17$country <- gsub(pattern = "Northern ", replacement = "Northern.", 
    x = players17$country)
players17$country <- gsub(pattern = "United ", replacement = "United.", 
    x = players17$country)
players17$country <- gsub(pattern = "Middle ", replacement = "Middle.", 
    x = players17$country)
players17$country <- gsub(pattern = "^.*Czech.*$", replacement = "Czech.Republic", 
    x = players17$country)
players17$country <- sapply(strsplit(x = players17$country, split = " "), function(x) x[1])
players17$country[players17$country == "United.States"] <- "USA"


sort(unique(players17$country))

library(dplyr)
nationalityRating <- as.data.frame(
    summarise(group_by(rating2015$ratings, country), 
        Rating = quantile(Rating, probs = 0.35)))

write.csv(nationalityRating, 
    file = "nationalityRating.csv", 
    quote = FALSE, row.names = FALSE)

players17 <- left_join(x = players17, 
    y = nationalityRating)
players17 <- left_join(x = players17, 
    y = transmute(rating2015$ratings, 
        player = Player, Rating1 = Rating, country = country))
dim(players17)
# set unrated players to have average score
players17$Rating1[is.na(players17$Rating1)] <- players17$Rating[is.na(players17$Rating1)]
sum(is.na(players17$Rating1))
players17$Rating1[is.na(players17$Rating1)] <- 1800
#help(package = "data.table")

# complete ratings object

new <- transmute(filter(players17,
    !players17$player %in% rating2015$ratings$Player),
    Player = player, Rating = Rating1, Deviation = 200, Games = 0, 
    Win = 0, Draw = 0, Loss = 0, Lag = 0, country = country)
new <- slice(group_by(new, Player), 1)

# add records to object
rating2015$ratings <- arrange(as.data.frame(rbind(rating2015$ratings, as.data.frame(new))), 
    Player)


players17 <- gather(players17, key = listno, value = List, 
    caster1, caster2)
players17 <- rename(players17, Team = team, Player = player)

#uniCasters <- sort(unique(players17$List))
#pairlookup16 <- matrix(0, nrow = length(casters), 
#    ncol = length(casters), dimnames = list(casters, casters))

# pairlookup16 <- setMatrixVal(x = pairlookup16, 
#     list1 = rep("Wurmwood1", times = length(uniCasters[!uniCasters %in% "Irusk2"])), list2 = uniCasters[!uniCasters %in% "Irusk2"], val = 2)
# pairlookup16 <- setMatrixVal(x = pairlookup16, 
#     list1 = rep("Madrak2", times = length(uniCasters[!uniCasters %in% "Haley2"])), list2 = uniCasters[!uniCasters %in% "Haley2"], val = 2)

players17 <- arrange(players17, Team, Player)

#rating2015$ratings$Player[!rating2015$ratings$Player %in% players17$Player]

players17[players17$Team == "Austria Mozart", ]
players17[players17$Team == "Czech Eagles", ]
players17[players17$Team == "Germany Loreley", ]
# remove duplicates
players17 <- slice(players17, -c(14, 337, 76, 399, 140, 463))

###############################################################################
#
# Simulation
#
###############################################################################

load("rating2016.rda")

nteams <- length(unique(players17$Team))
nn <- 10000L
out <- matrix(NA_character_, nrow = nteams, ncol = nn)

library(foreach)
library(doSNOW)
cl <- makeCluster(2L)

registerDoSNOW(cl)

set.seed(287934)
sseed <- as.integer(runif(n = nn, min = 1, max = nn) * 1000)
sseed[duplicated(sseed)] <- as.integer(
    runif(n = sum(duplicated(sseed)), 
        min = nn, max = nn * 10) * 1000)
sum(duplicated(sseed))
#[1] 0
system.time(
    #out <- foreach(ii = seq_len(nn)) %do% {
    out <- foreach(ii = seq_len(nn), .packages = "WTCTools") %dopar% {
        #for (ii in seq_len(nn)) {
        set.seed(sseed[ii])
        thisRound <- doRound(data = players17, results = NULL, 
            ratings = rating2016, 
            pairlookup = mark3lookup)
        
        for (jj in 8:12) { 
            thisRound <- doRound(data = players17, 
                round = jj, 
                results = thisRound, 
                ratings = rating2016, 
                pairlookup = mark3lookup)
        }
        tmp <- rownames(thisRound)
        tmp[order(thisRound$Matches.Won, 
            thisRound$Total.Games.Won, 
            decreasing = TRUE)]
        
    })

stopCluster(cl)

save(out, file = "out21092017.RData")


###############################################################################
#
# Summarise Simulation Results
#
###############################################################################


load("out21092017.RData")
nn <- length(out)

out <- t(do.call(what = "rbind", out))
# order for each simulation
tab <- apply(X = out, MARGIN = 1L, FUN = function(x) {
    sort(table(x), decreasing = TRUE) })

(tab[[1]] / (nn/100))[1:6]

res <- matrix(NA, nrow = length(tab), ncol = length(tab))
teamNames <- out[, 1]
rownames(res) <- teamNames
for (i in seq_along(tab)) {
    perc <- tab[[i]] / (nn/100)
    perc[teamNames[!teamNames %in% names(perc)]] <- 0
    res[teamNames, i] <- perc[teamNames]
}


###############################################################################
#
# Summarize Results
#
###############################################################################


moment <- function(x) {
    mean(x * seq_along(x))
}

mmax <- apply(X = res, MARGIN = 1L, FUN = moment)
reso <- res[order(mmax), ]
mmax <- mmax[order(mmax)]

mmax

dat <- data.frame(Rank = seq_along(mmax), Team = names(mmax), "Score Moment" = round(mmax, 3))
rownames(dat) <- NULL
library(xtable)
print(xtable(dat), type = "html")

###############################################################################
#
# Expert Opinion Bootstrap
#
###############################################################################

klaw <- matrix(c(
    "Sascha Maisel",     "Australia Drop Bear", "Canada Beaver", "England Lions", "Germany Loreley", "Italy Michelangelo", "Poland Hussar", "Sweden Götaland",
    "Chris Davies",      "Australia Drop Bear", "Canada Beaver", "England Lions", "Germany Loreley",  "Norway Hugin",    "Poland Hussar", "Sweden Norrland",
    "Bubba Dalton",      "Australia Drop Bear", "Canada Beaver", "England Lions", "Germany Loreley", "Poland Hussar",   "Sweden Götaland", "USA Texas",
    "Tom Guan",          "Australia Drop Bear", "Canada Beaver", "England Lions", "Germany Loreley", "Italy Michelangelo", "Poland Hussar", "USA Texas",
    "Ryan Evans",        "Australia Drop Bear", "Canada Moose",  "England Lions", "Germany Loreley", "Norway Hugin",    "Poland Hussar", "Sweden Götaland",
    "Konrad Sosnowski",  "Australia Drop Bear",  "Canada Moose",  "Germany Loreley", "Sweden Götaland", "USA Texas",     "Poland Hussar", "Poland Knights",
    "Johan Dyrlind",     "Australia Drop Bear",  "Canada Beaver", "Germany Loreley", "Poland Hussar", "Sweden Götaland", "Sweden Norrland", "Sweden Svealand",
    "Norbert Brunhuber", "Australia Drop Bear",  "Canada Beaver", "England Lions", "Germany Loreley", "Poland Hussar",   "Sweden Götaland", "USA Texas"), 
    nrow = 8, byrow = TRUE)
klaw <- as.data.frame(klaw, stringsAsFactors = FALSE)
klaw <- rename(klaw, Player = V1)
klaw <- gather(klaw, key = Rank, value = Team, -Player)

klaw <- mutate(klaw, Rank = (1:7)[factor(Rank)])
klaw <- arrange(klaw, Player, Rank)
head(klaw)
unique(klaw$Team[!klaw$Team %in% teamNames])
# showSimilarNames(c(klaw$Team, teamNames))
# borda count
klaw1 <- arrange(summarize(group_by(klaw, Team), 
    Borda = sum(7 - Rank, na.rm = TRUE)), desc(Borda))
print(xtable(klaw1), type = "html")

###############################################################################
#
# Simulation Plot
#
###############################################################################

library(ggplot2)
png("Australia_Drop_Bear_Prob_Density_2017.png", height = 600, width = 800, res = 120)
qplot(data = cbind.data.frame(Rank = c(0.9, seq_len(nrow(res))), 
    `Probability Density` = c(0, res["Australia Drop Bear",])), 
    x = Rank, y = `Probability Density`, 
    geom = "polygon", fill = I("gold")) +
    ggtitle("Likelihood of Rank of Australia Drop Bear") +
    theme_bw()
dev.off()

wtc_predictions_2017_update_rating.R

# cleaning of wtc data and rating data

library(WTCTools)
library(tidyr)

wtc$country <- gsub(pattern = "Team ", replacement = "", 
    x = wtc$team1)
wtc$country <- gsub(pattern = "Epic ", replacement = "", 
    x = wtc$country)
wtc$country <- gsub(pattern = "Prime ", replacement = "", 
    x = wtc$country)
wtc$country <- gsub(pattern = "Northern ", replacement = "Northern.", 
    x = wtc$country)
wtc$country <- gsub(pattern = "United ", replacement = "United.", 
    x = wtc$country)
wtc$country <- gsub(pattern = "Middle ", replacement = "Middle.", 
    x = wtc$country)
wtc$country <- gsub(pattern = "Czech ", replacement = "Czech.", 
    x = wtc$country)
wtc$country <- sapply(strsplit(x = wtc$country, split = " "), function(x) x[1])
wtc$country[wtc$country == "United.States"] <- "USA"

sort(unique(wtc$country))

nationalities <- wtc[!duplicated(wtc$player1), c("player1", "country")]
wtc1 <- gather(wtc, key = player, value = name, player1, player2)
gameCounts <- summarize(group_by(wtc1, name), n())
pNames <- showSimilarNames(gameCounts$name)
# remove non-matched names
pNames <- pNames[-c(1, 2, 3, 8, 12, 23, 29, 30), ]
# align replacements
index <- c(5, 7, 8, 9, 18, 19, 23)
for (i in index) pNames[i, ] <- rev(pNames[i, ])
pNames
for (i in seq_len(nrow(pNames))) {
    wtc1$name[wtc1$name == pNames[i, 2]] <- pNames[i, 1]
}

# manually fix
wtc1$name[wtc1$name == "PA¤r-Ola Nilsson"] <- "Per-Ola Nilsson"
wtc1$name[wtc1$name == "BenoA®t CA©rA©"] <- "Benoit Cere"
wtc1$name[wtc1$name == "HA¥vard Kongsrud"] <- "Haovard Kongsrud"
wtc1$name[wtc1$name == "AⱲic Grux"] <- "Eric Grux"
wtc1$name[wtc1$name == "KA©vin Maeder"] <- "Kevin Maeder"
wtc1$name[wtc1$name == "SA¸ren Kongsgaard"] <- "Soeren Kongsgaard"
wtc1$name[wtc1$name == "Matej TkA¡c"] <- "Matej Tkac"
wtc1$name[wtc1$name == "Mathias LA¼tzkendorf"] <- "Mathias Luetzkendorf"
wtc1$name[wtc1$name == "Torjus EftestA¸l"] <- "Torjus Eftestel"
wtc1$name[wtc1$name == "SA¸ren Husum"] <- "Soeren Husum"
wtc1$name[grep(wtc1$name, pattern = "^.+ric Grux$")] <- "Eric Grux"
wtc1$name[grep(wtc1$name, pattern = "^.+s Leite$")] <- "Luis Leite"
wtc1$name[wtc1$name == "SA¸ren Laust Jacobsen"] <- "Soeren Laust Jacobsen"
wtc1$name[wtc1$name == "Benjamin FA¼tzi"] <- "Benjamin Futzi"
wtc1$name[wtc1$name == "LorA¡nd SA¶vA©nyhA¡zi"] <- "Lorand Sovenyhazi"
wtc1$name[wtc1$name == "PA©ter MolnA¡r"] <- "Peter Molnar"
wtc1$name[wtc1$name == "AndrA¡s PorkolA¡b"] <- "Andras Porkolab"
wtc1$name[wtc1$name == "JA¡nos Kendi"] <- "Janos Kendi"
wtc1$name[wtc1$name == "KA©vin Aouragh"] <- "Kevin Aouragh"
wtc1$name[wtc1$name == "AurA©lien Guiramand"] <- "Aurelien Guiramand"
wtc1$name[wtc1$name == "LuboA¡ Hribik"] <- "Luboa Hribik"
wtc1$name[wtc1$name == "Gerard O' Brien"] <- "Gerard O'Brien"
wtc1$name[wtc1$name == "Gerard O'brien"] <- "Gerard O'Brien"
wtc1$name[wtc1$name == "Kevin Aourragh"] <- "Kevin Aouragh"




wtc1$name[grep(wtc1$name, pattern = "^Peter Cornelious Dyhre.*Poulsen$")] <- "Peter Cornelious Dyhre-Poulsen"
wtc1$name[wtc1$name == "Konrad KrA³licki"] <- "Konrad Krolicki"
sort(unique(wtc1$name[!grepl("^[A-Za-z '\\.-]+$", wtc1$name)]))

wtc2 <- spread(wtc1, key = player, value = name)

load("rating2015-reg.RData")

# check special character
z <- rating2015$ratings$Player
source("toPlain.R")
z <- toPlain(z)
z[z == "(C) Daniel Bogdanoski"] <- "Daniel Bogdanoski"
z[z == "Dominik de Cassan: Axis+Iron Mother"] <- "Dominik de Cassan"
z[grep(z, pattern = "^P.*r-Ola Nilsson$")] <- "Per-Ola Nilsson"
z[grep(z, pattern = "^Beno.*t C.*r.*$")] <- "Benoit Cere"
z[grep(z, pattern = "^H.*vard Kongsrud$")] <- "Haovard Kongsrud"
z[z == "KA©vin Maeder"] <- "Kevin Maeder"
z[grep(z, pattern = "^S.*ren Kongsgaard$")] <- "Soeren Kongsgaard"
z[grep(z, pattern = "^Matej Tk.*$")] <- "Matej Tkac"
z[grep(z, pattern = "^Mathias L.*tzkendorf$")] <- "Mathias Luetzkendorf"
z[grep(z, pattern = "^Torjus Eftest.*l$")] <- "Torjus Eftestel"
z[grep(z, pattern = "^S.*ren Husum$")] <- "Soeren Husum"
z[grep(z, pattern = "^.+ric Grux$")] <- "Eric Grux"
z[grep(z, pattern = "^.+s Leite$")] <- "Luis Leite"
z[grep(z, pattern = "^S.*ren Laust Jacobsen$")] <- "Soeren Laust Jacobsen"
z[z == "Benjamin FA¼tzi"] <- "Benjamin Futzi"
z[z == "LorA¡nd SA¶vA©nyhA¡zi"] <- "Lorand Sovenyhazi"
z[grep(z, pattern = "^P.*ter Moln.*r$")] <- "Peter Molnar"
z[grep(z, pattern = "^Andr.*s Porkol.*b$")] <- "Andras Porkolab"
z[grep(z, pattern = "^J.*nos Kendi$")] <- "Janos Kendi"
z[z == "KA©vin Aouragh"] <- "Kevin Aouragh"
z[grep(z, pattern = "^Aur.*lien Guiramand$")] <- "Aurelien Guiramand"
z[grep(z, pattern = "^Lubo.* Hribik$")] <- "Luboa Hribik"
z[grep(z, pattern = "^Konrad Kr.*licki$")] <- "Konrad Krolicki"
z[grep(z, pattern = "^Peter Cornelious Dyhre.*Poulsen$")] <- "Peter Cornelious Dyhre-Poulsen"
#z[grep(z, pattern = "^M.*t.* T.*ri$")]
#z[z == "Gerard O' Brien"] <- "Gerard O'Brien"
#z[z == "Gerard O'brien"] <- "Gerard O'Brien"
#z[z == "Kevin Aourragh"] <- "Kevin Aouragh"

z1 <- sort(z[!grepl("^[A-Za-z '\\.-]+$", z)])
z1

rating2015$ratings$Player <- z
save(rating2015, file = "rating2015-reg16.rda")

# merge duplicated names
dat15 <- rating2015$ratings
showSimilarNames(dat15$Player)

# remove nicknames
dat15$Player[grep("Snot", dat15$Player)] <- "Daniel Pawelka"
dat15$Player[grep("Raghnath", dat15$Player)] <- "Martin Florian"


# filter(dat15, grepl(pattern = "Kasd.*vasilis", x = Player))
# missed Dimitris
dat15 <- filter(dat15, !grepl(pattern = "Dimitris Kasdiovasilis", x = Player))
dat15 <- mutate(dat15, 
    Player = replace(Player, 
        grepl(pattern = "Dimitrios Kasdovasilis", x = Player), 
        "Dimitris Kasdiovasilis"))
# Frederik Punt
filter(dat15, grepl(pattern = "Fre.*k Punt", x = Player))
row <- filter(dat15, grepl(pattern = "Fre.*k Punt", x = Player))
row$Rating <- mean(row$Rating)
row$Games <- sum(row$Games)
row$Win <- sum(row$Win)
row$Loss <- sum(row$Loss)
dat15 <- filter(dat15, !grepl(pattern = "Freek Punt", x = Player))
dat15[dat15$Player == "Frederik Punt", ] <- row[1, ]

# filter(dat15, grepl(pattern = "Pawelka", x = Player))
dat15 <- slice(dat15, -grep(pattern = "Daniel Pawelka", x = Player)[2])
filter(dat15, grepl(pattern = "Florian$", x = Player))
dat15 <- slice(dat15, -grep(pattern = "Daniel Pawelka", x = Player)[1])
# "Jason Mcleod"
filter(dat15, grepl(pattern = "Ja.* Mcleod", x = Player))
dat15 <- filter(dat15, !grepl(pattern = "Jason Mcleod", x = Player))
dat15 <- mutate(dat15, 
    Player = replace(Player, 
        grepl(pattern = "Jay Mcleod", x = Player), 
        "Jason Mcleod"))
# "Josh Bates"  
filter(dat15, grepl(pattern = "Josh.* Bates", x = Player))
dat15 <- filter(dat15, !grepl(pattern = "Josh Bates", x = Player))

# "Michal Nakonieczny"
row <- filter(dat15, grepl(pattern = "Michal .+onieczny", x = Player))
row$Player <- "Michal Nakonieczny"
row$Games <- sum(row$Games)
row$Win <- sum(row$Win)
row$Loss <- sum(row$Loss)
dat15 <- filter(dat15, !grepl(pattern = "Michal Konieczny", x = Player))
dat15[dat15$Player == "Michal Nakonieczny", ] <- row[1, ]
  
# "Nikolaos Kerazoglou"
filter(dat15, grepl(pattern = "Niko.*s Kerazoglou", x = Player))
dat15 <- filter(dat15, !grepl(pattern = "Nikos Kerazoglou", x = Player))
   
# "Panagiotis Demiris"
filter(dat15, grepl(pattern = "Panagiotis .+emiris", x = Player))
dat15 <- filter(dat15, !grepl(pattern = "Panagiotis Demiris", x = Player))
dat15 <- mutate(dat15, 
    Player = replace(Player, 
        grepl(pattern = "Panagiotis Ntemiris", x = Player), 
        "Panagiotis Demiris"))
 


# dat15$Player[grep(".*Mader", x = dat15$Player)]
# grep(" \\\".+\\\" ", x = dat15$Player, value = TRUE)

showSimilarNames(dat15$Player)
rating2015$ratings <- dat15


showSimilarNames(c(rating2015$ratings$Player, wtc$player1, wtc$player2))

showSimilarNames(c(dat15$Player, players17$player))



save(rating2015, file = "rating2015-reg16.rda")

# rating2015$ratings <- merge(x = rating2015$ratings, 
#     y = nationalities, all.x = TRUE, by.x = "Player", by.y = "player1")


sum(is.na(rating2015$ratings$country))