The Quest for the GOAT

ggplot2

flextable

sports

Solving the never-ending debate of the greatest basketball player of all time

Published

February 12, 2023

Who is the GOAT?

For decades armchair athletes have been debating, who is the the greatest basketball player of all time: LeBron James, Kobe Bryant or Michael Jordan.

But who really is the GOAT? Based on our analysis it is clear who the best player of all times is and we can finally put these debates to rest.

The Data

We based our findings off of various datasets from Kaggle.com.

Back to Basics

First we looked at some basic aggregates to see if a top player emerged.

Michael Jordan

15 seasons
1072 games
6 NBA championships
5x NBA MVP
14x NBA All-Star
32,292 career points
30.1 PPG

Kobe Bryant

20 seasons
1,346 games
5 NBA championships
1x NBA MVP
18x NBA All-Star
33,643 career points
25 PPG

LeBron James

20 seasons & counting
1,378 games
4 NBA championships
4x NBA MVP
18x NBA All-Star
37,371 career points
27.1 PPG

What does this mean? Still no clear winner based on these initial statistics.

More advanced statistics

Behind the scenes

goat.sub <- goat %>% 
  group_by(player) %>%
  select(player, playereffrating,trueshootingperc,freethrowrate,winshare,overallbpm,valueoverreplace)%>%
  summarise(across(everything(), mean))

goat.sub <- goat.sub %>%
  mutate(playereffrating=round(playereffrating,2)) %>%
  mutate(trueshootingperc=round(trueshootingperc*100,2))%>%
  mutate(freethrowrate=round(freethrowrate*100,2))%>%
  mutate(winshare=round(winshare,2))%>%
  mutate(overallbpm=round(overallbpm,2))%>%
  mutate(valueoverreplace=round(valueoverreplace,2))
         
ft <- flextable(goat.sub)
ft <- set_header_labels(ft,
  values = list(
    player = "Player",
    playereffrating = "Player Efficiency Rating",
    trueshootingperc = "True Shooting %",
    freethrowrate = "Free Throw Rate",
    winshare = "Win Share",
    overallbpm = "Overall BPM",
    valueoverreplace = "Value Over Replacement"
  )
)
ft <- theme_booktabs(ft)
ft <- align( ft, align = "center", part = "all" )
ft <- bold(ft, part = "header") # bold header
ft <- bg(ft, i = ~ playereffrating > 27.9 & trueshootingperc > .58 & freethrowrate > 43 & winshare > 9.5 & overallbpm > 10 & valueoverreplace > 5, 
        j = ~ playereffrating + trueshootingperc + freethrowrate + winshare + overallbpm + valueoverreplace, 
        bg = "wheat", part = "body")
ft

Player	Player Efficiency Rating	True Shooting %	Free Throw Rate	Win Share	Overall BPM	Value Over Replacement
Kobe Bryant	21.40	54.13	37.49	5.74	3.04	2.47
Lebron James	27.94	58.49	43.04	9.58	10.06	5.61
Michael Jordan	27.84	56.34	38.60	9.06	8.65	4.55

Let’s Get Advanced

We turned to various advanced statistics for more clarity and chose to look at each players Box Plus-Minus(BPM) throughout their careers.

BPM is a metric that:

is derived from a player’s own statistics (points, rebounds, and assists, etc.), along with their position, and overall team performance
estimates a player’s contribution to the team when that player is on the court in points above the NBA average, which is 0
Example: a BPM of +5 means that per game, the team is 5 points better with the player on the court than with the average NBA player

BPM: Individual Player vs. Team

How do they stack up against their teammates each season?

Behind the scenes

#remove all * from the Player column
fullnba$Player<- gsub("[*]", "", fullnba$Player)
#convert season to year
goat$season<-as.Date(goat$season, format="%Y")
goat$year<-year(goat$season)

#define custom colors
bullsred = "#CE1141"
lakeryellow = "#FDB927"
lakerpurple = "#552583"
cavswine = "#6f263d"
cavsgold = "#ffb81c"
miamigold = "#F9A01B"
miamired = "#98002E"

#filter dataset for specific players
justmj<-filter(goat, player == "Michael Jordan")%>%
  group_by(year) %>%
  summarise(mean = mean(overallbpm, na.rm=TRUE))
justkb<-filter(goat, player == "Kobe Bryant")%>%
  group_by(year) %>%
  summarise(mean = mean(overallbpm, na.rm=TRUE))
justlj<-filter(goat, player == "Lebron James", between(year, 2003, 2017))%>%
  group_by(year) %>%
  summarise(mean = mean(overallbpm, na.rm=TRUE))

wizbpm<-filter(fullnba, Player != "Michael Jordan", Tm == "WAS", between(Year, 2001, 2002))
cavbpm<-filter(fullnba, Player != "LeBron James", Tm == "CLE", between(Year, 2003, 2009)) 
miabpm<-filter(fullnba, Player != "LeBron James", Tm == "MIA", between(Year, 2010, 2013))
cav2bpm<-filter(fullnba, Player != "LeBron James", Tm == "CLE", between(Year, 2014, 2017))
lakbpm<-filter(fullnba, Player != "LeBron James", Tm == "LAL", between(Year, 2017, 2018))

#plots
#Michael Jordan
mj<-fullnba %>%
  filter(Player != "Michael Jordan", Tm == "CHI", between(Year, 1983, 2000)) %>%
  ggplot(aes(x=Year, y=BPM, fill="black"))+
    geom_point(alpha=.2, size=2)+
    geom_point(data=wizbpm, aes(x=Year, y=BPM), alpha=.2, size=2, fill="black")+
    geom_point(data=justmj, aes(x=year, y=mean, color=bullsred), size=3)+
    scale_colour_manual(values = bullsred)+
    scale_x_continuous(breaks=seq(1983, 2002, 1))+
    geom_hline(yintercept = 0, alpha=.1, size=2, linetype="solid")+
    geom_vline(xintercept = 1993, alpha=.05, size=8, linetype="solid")+
    geom_vline(xintercept = 1999, alpha=.05, size=40, linetype="solid")+
    geom_point(aes(x=1990, y=12.3),color='black', fill=bullsred, size=3, shape=21)+
    geom_point(aes(x=1991, y=9.1),color='black', fill=bullsred, size=3, shape=21)+
    geom_point(aes(x=1992, y=10.2),color='black', fill=bullsred, size=3, shape=21)+
    geom_point(aes(x=1995, y=8.35),color='black', fill=bullsred, size=3, shape=21)+
    geom_point(aes(x=1996, y=7.7),color='black', fill=bullsred, size=3, shape=21)+
    geom_point(aes(x=1997, y=5.4),color='black', fill=bullsred, size=3, shape=21)+
    annotate("text", x = 1993, y = 15, label = "RETIRED", family="sans", size=3)+
    annotate("text", x = 1999, y = 15, label = "RETIRED", family="sans", size=3)+
    theme_minimal()+
    theme(
      legend.position="none",
      panel.grid.minor = element_blank(),
      panel.grid.major = element_blank(),
    axis.title.y = element_text(margin = margin(t = 0, r = 20, b = 0, l = 0)),
    axis.title.x = element_text(margin = margin(t = 20, r = 0, b = 0, l = 0)),
    plot.title = element_text(margin = margin(t = 0, r = 0, b = 20, l = 0))
      )+
    labs(
      title="Michael Jordan v. Bulls (1983-2000) & Wizards (2001-2002)",
      y="BPM",
      x="Season (Starting Year)"
    )

#Kobe Bryant
kb<-fullnba %>%
  filter(Player != "Kobe Bryant", Tm == "LAL", between(Year, 1996, 2015)) %>%
  ggplot(aes(x=Year, y=BPM, fill="black"))+
    geom_point(alpha=.2, size=2)+
    geom_point(data=justkb, aes(x=year, y=mean, color=lakeryellow), size=3)+
    scale_colour_manual(values = lakeryellow)+
    scale_x_continuous(breaks=seq(1996, 2015, 1))+
    geom_hline(yintercept = 0, alpha=.1, size=2, linetype="solid")+
    geom_vline(xintercept = 2013, alpha=.05, size=8, linetype="solid")+
    geom_point(aes(x=1999, y=4.3),color=lakerpurple, fill=lakeryellow, size=3, shape=21)+
    geom_point(aes(x=2000, y=5.75),color=lakerpurple, fill=lakeryellow, size=3, shape=21)+
    geom_point(aes(x=2001, y=3.85),color=lakerpurple, fill=lakeryellow, size=3, shape=21)+
    geom_point(aes(x=2008, y=6),color=lakerpurple, fill=lakeryellow, size=3, shape=21)+
    geom_point(aes(x=2009, y=4.75),color=lakerpurple, fill=lakeryellow, size=3, shape=21)+
    annotate("text", x = 2013, y = 15, label = "INJURED", family="sans", size=3)+
    theme_minimal()+
    theme(
      legend.position="none",
      panel.grid.minor = element_blank(),
      panel.grid.major = element_blank(),
    axis.title.y = element_text(margin = margin(t = 0, r = 20, b = 0, l = 0)),
    axis.title.x = element_text(margin = margin(t = 20, r = 0, b = 0, l = 0)),
    plot.title = element_text(margin = margin(t = 0, r = 0, b = 20, l = 0))
      )+
    labs(
      title="Kobe v. Lakers (1996-2015)",
      y="BPM",
      x="Season (Starting Year)"
    )


#LeBron James
lj<-cavbpm %>%
  ggplot(aes(x=Year, y=BPM, fill="black"))+
    geom_point(alpha=.2, size=2)+
    geom_point(data=cav2bpm, aes(x=Year, y=BPM), alpha=.2, size=2, fill="black")+
    geom_point(data=miabpm, aes(x=Year, y=BPM), alpha=.2, size=2, fill="black")+
    geom_point(data=lakbpm, aes(x=Year, y=BPM), alpha=.2, size=2, fill="black")+
    geom_point(data=justlj, aes(x=year, y=mean, color=cavswine), size=3)+
    scale_colour_manual(values = cavswine)+
    scale_x_continuous(breaks=seq(2003, 2018, 1))+
    geom_hline(yintercept = 0, alpha=.1, size=2, linetype="solid")+
    geom_point(aes(x=2011, y=11),color=cavsgold, fill=cavswine, size=4, shape=21)+
    geom_point(aes(x=2012, y=10.9),color=cavsgold, fill=cavswine, size=4, shape=21)+
    geom_point(aes(x=2015, y=11.1),color=cavsgold, fill=cavswine, size=4, shape=21)+
    theme_minimal()+
    theme(
      legend.position="none",
      panel.grid.minor = element_blank(),
      panel.grid.major = element_blank(),
    axis.title.y = element_text(margin = margin(t = 0, r = 20, b = 0, l = 0)),
    axis.title.x = element_text(margin = margin(t = 20, r = 0, b = 0, l = 0)),
    plot.title = element_text(margin = margin(t = 0, r = 0, b = 20, l = 0))
      )+
    labs(
      title="LeBron v. Caveliers (2003-2009 & 2014-2017), Heat (2010-2013) & Lakers (2017)",
      y="BPM",
      x="Season (Starting Year)"
    )

BPM: GOAT v GOAT

The GOAT is Revealed

With a career average BPM of 8.1 and solid performance in both regular season and playoff games, LeBron James is called King James for a reason.

« Back to Projects