Created by: Patrick Ingelmo
www.patrickingelmo.com

Overview

The three-point line was adopted by the NBA during the 1979-80 season. Since then, the 3-point shot has continued to increase in importance and is now an integral part of the game. The new age of analytics in the NBA has also driven some extreme strategic behavior with regard to the 3-point shot. Look no further than the Houston Rockets to understand this change. Driven by stats-nerd GM Daryl Morey, the Rockets have looked to eliminate mid-range shots from their repetoire (http://es.pn/1erBBfi). This analysis will look at the 3-point shot as it has progressed in the 35 years since its introduction.

#load libraries
library(ggplot2)
library(ggthemes)
library(dplyr)
library(reshape2)
library(knitr)

#set directory to where csv files are located
setwd("/Users/pingelmo/Dropbox/data/nba/teams")
#initialize data frame
nba <- data.frame()
#assign files to list
files_list <- list.files()
#read in csv and row bind into one data frame
for(file in files_list){nba<-rbind(nba,read.table(file,header=T,sep=",",skip=2,stringsAsFactors=F))}
#remove rows not containing team data
nba <- nba[!nba$Season=="Season",]
rownames(nba) <- NULL
# convert characters to numeric
nba[,c(1,5:40)] <- sapply(nba[,c(1,5:40)],as.numeric)
#convert character to factor
nba$Season <- as.factor(nba$Season)
#remove asterisk from Team Name
nba$Tm <- substr(nba[,3],1,3)
#season end
nba$season_end <- as.numeric(substr(nba[,2],1,4))+1
options("scipen"=100, "digits"=4)

#load NBA champions data
finals <- read.csv("/Users/pingelmo/Dropbox/data/nba/finals.csv",header=T)
#create new binary championship variable
finals$finals <- 1
#rename to match NBA dataset
names(finals) <- c("Season","Tm","finals")
#join
nba <- left_join(nba,finals)
#remove NAs
nba$finals <- ifelse(is.na(nba$finals),0,1)

#rename fields
names(nba)[8] <- "win_pct"
names(nba)[33] <- "eFG"
names(nba)[34] <- "TOV_rate"
names(nba)[35] <- "ORB_rate"
names(nba)[36] <- "FT_FGA"
names(nba)[37] <- "opp_eFG"
names(nba)[38] <- "opp_TOV"
names(nba)[39] <- "opp_ORB"
names(nba)[40] <- "opp_FT_FGA"


# get rid of NAs, replace with 0
nba$X2P <- ifelse(is.na(nba$X2P),0,nba$X2P)
nba$X2PA <- ifelse(is.na(nba$X2PA),0,nba$X2PA)
nba$X3P <- ifelse(is.na(nba$X3P),0,nba$X3P)
nba$X3PA <- ifelse(is.na(nba$X3PA),0,nba$X3PA)

# new variables
nba$fg_per <- nba$FG/nba$FGA
nba$x2p_per <- nba$X2P/nba$X2PA
nba$ft_per <- nba$FT/nba$FTA
nba$pts_per_fga <- nba$PTS/nba$FGA
# decade variable
nba$decade <- cut(nba$season_end,breaks=c(0,1989,1999,2009,2015),labels=c("1980s","1990s","2000s",'2010s'))
# new 2-pt variables
nba$x2pa_game <- nba$X2PA/nba$G
nba$x2p_game <- nba$X2P/nba$G
nba$x2p_per <- nba$X2P/nba$X2PA
# new 3-pt variables
nba$x3pa_game <- nba$X3PA/nba$G
nba$x3p_game <- nba$X3P/nba$G
nba$x3p_per <- nba$X3P/nba$X3PA
nba$x3pa_share <- nba$X3PA/nba$FGA
#new fga
nba$fga_game <- nba$FGA/nba$G
nba$fg_game <- nba$FG/nba$G
nba$ft_game <- nba$FT/nba$G

Explore Data

The NBA is played in a far different way than it was in 1980. Below is a snapshot showing 1980-2015, listing field goal attempts per game, 2-point attempts per game and 3-point attempts per game. Since 1980, field goal attempts per game have dropped from about 91 to 83 but 3-point shots have increased from 3 to just over 22.

# create data frame
shot_change <- nba %>% filter(season_end %in% c(1980,1990,2000,2010,2015)) %>% group_by(season_end) %>%summarize(FGA_per_game=mean(fga_game),x2PA_per_game=mean(x2pa_game),x3PA_per_game = mean(x3pa_game)) %>% select(season_end,FGA_per_game,x2PA_per_game,x3PA_per_game)

#display contents
kable(shot_change)
season_end FGA_per_game x2PA_per_game x3PA_per_game
1980 90.64 87.87 2.773
1990 87.15 80.55 6.598
2000 82.09 68.38 13.715
2010 81.70 63.56 18.139
2015 83.49 61.11 22.381 
#melt to plot
shot_change <- melt(shot_change,id="season_end")

#plot
shot_change %>% ggplot(aes(x=season_end,y=value,group=variable,col=variable))+geom_point(size=4)+geom_line(size=1.5)+ggtitle(expression(atop("Evolution of Shot Selection",atop("3 point attempts per game have increased from 2.7 in 1980 to 22.3 in 2015"))))+theme_fivethirtyeight()+theme(plot.title=element_text(hjust=.5),legend.position="none")+annotate("text",x=1985,y=11,label="3P Attempts",col="blue")+annotate("text",x=1985,y=79,label="2P Attempts",col="dark green")+annotate("text",x=1990,y=95,label="FG Attempts",col="red")

#3 point shooting by decade 
nba %>% filter(season_end>1980) %>% ggplot(aes(x=season_end,y=x3pa_game,col=decade))+geom_point()+geom_smooth(method="lm")+theme_fivethirtyeight()+theme(legend.position="none")+ggtitle(expression(atop("3-Point Attempts per Game",atop("Colored by decade"))))

#3 point shooting percentage by decade 
nba %>% filter(season_end>1980) %>% ggplot(aes(x=season_end,y=x3p_per,col=decade))+geom_point()+geom_smooth(method="lm")+theme_fivethirtyeight()+theme(legend.position="none")+ggtitle(expression(atop("3-Point Field Goal Percent",atop("2015 League-Wide Average is 34.6%"))))

League-Wide Trends

Since 1980, field goal attempts, 2-point attempts, free throw attempts and pace have all decreased, while 3-point attempts have increased. Unsurprisingly, field goal attempts per game mimics pace. NBA pace decreased steadily from 1980 through the early 2000s, as the half-court offense evolved. Pace has started to increase again, possibly as a factor of the 3-point shot.

# Pace
nba %>% filter(season_end>1980) %>% ggplot(aes(x=season_end,y=Pace,col=decade))+geom_point()+geom_smooth(method="lm")+theme_fivethirtyeight()+theme(legend.position="none")+ggtitle(expression(atop("Pace per Game",atop("Colored by decade"))))

# Field Goal Attempts per Game
nba %>% filter(season_end>1980,fga_game<105) %>% ggplot(aes(x=season_end,y=fga_game,col=decade))+geom_point()+geom_smooth(method="lm")+theme_fivethirtyeight()+theme(legend.position="none")+ggtitle(expression(atop("Field Goal Attempts per Game",atop("Colored by decade"))))

# 2-Point Attempts per Game
nba %>% filter(season_end>1980) %>% ggplot(aes(x=season_end,y=x2pa_game,col=decade))+geom_point()+geom_smooth(method="lm")+theme_fivethirtyeight()+theme(legend.position="none")+ggtitle(expression(atop("2 Point Attempts per Game",atop("Colored by decade"))))

# Free Throw Attempts
nba %>% filter(season_end>1980) %>% ggplot(aes(x=season_end,y=ft_game,col=decade))+geom_point()+geom_smooth(method="lm")+theme_fivethirtyeight()+theme(legend.position="none")+ggtitle(expression(atop("Free Throw Attempts per Game",atop("Colored by decade"))))

# 3-Point Attempts per Game
nba %>% filter(season_end>1980) %>% ggplot(aes(x=season_end,y=x3pa_game,col=decade))+geom_point()+geom_smooth(method="lm")+theme_fivethirtyeight()+theme(legend.position="none")+ggtitle(expression(atop("3 Point Attempts per Game",atop("Colored by decade"))))

The more 3-Point shots the better?

The share of field goal attempts that are 3-point shots has been increasing every year since 1980. This year, almost 27% of a team’s field goal attempts are 3-point shots. And for the first time in NBA history, a team is attempting more than 40% of their shots from 3-point territory. Can you guess who?

# 3 point share by season
nba %>% filter(season_end>1980) %>% ggplot(aes(x=season_end,y=x3pa_share,col=decade))+geom_smooth()+geom_point()+theme_fivethirtyeight()+ggtitle("Percent of Shots that are 3-Pointers")+theme(legend.position="none")

Future of NBA?

If you guessed the Houston Rockets, then you’d be correct. The Rockets have paced well above league average for the percent of shots that are 3-pointers.

# Rockets
HOU <- nba %>% filter(season_end>2000)
HOU$HOU <- ifelse(HOU$Tm=="HOU",1,0)

HOU %>% ggplot(aes(x=season_end,y=x3pa_share,size=HOU,col=HOU))+geom_point()+geom_smooth()+theme_fivethirtyeight()+theme(legend.position="none",plot.title=element_text(size=rel(1.6),hjust=0))+scale_size_continuous(range=c(2,4))+ggtitle(expression(atop("Percent of Shots that are 3-Pointers",atop("Houston Rockets are blue dots"))))+annotate("text",x=2014,y=.38,label="Harden Era")