NBA analysis - Basketball Essay Example
- Data Description
The data for the National Basketball Association (NBA) teams that played in the playoff from 1994-1995 to 2002-2003 was downloaded from (NBA Reference 2007) - NBA analysis introduction. The data contained team names, percentage of games won during normal season, total points scored during normal season, total points scored by opponent during normal season, and number of games won during the playoffs.
essay sample on "NBA analysis"
? We will write a cheap essay sample on "NBA analysis" specifically for you for only $12.90/pageMore Basketball Essay Topics.
1.1 Data variables
The following variables are used in this study:
PCT | : | Winning Percentage of Team during Normal Season |
PTS | : | Average of points scored by team during Normal Season |
DPTS | : | Average of points scored by opponent teams during Normal Season |
POWINS | : | Number of wins in the Playoffs |
1.2 Descriptive Statistics of Data
During the period of the study from 1994 to 2003, the data for 259 playoff game is described in Table 1. Table 1 show that the mean winning percentage during normal season for teams that played in the playoffs is 0.5 with a standard deviation of 0.16. The number of points scored is equal to number of points scored by opponents which makes perfect sense. The following table will analyze the data for teams that played the final, quarter final, and others teams that only in the playoffs.
Table 1: Descriptive Statistics of Wining Percentage (PCT), Points Scored (PTS),
Points Scored by Opponents (DPTS), Playoff Wins (POWINS).
Variable | Number | Mean | Mode | Std. Deviation |
PCT | 259 | 0.5 | 0.6097 | 0.16196 |
PTS | 259 | 96.384 | 95.85 | 4.87160 |
DPTS | 259 | 96.384 | 97.93 | 5.05113 |
POWINS | 259 | 1.36 | 0 | 3.027 |
NBA teams are classified in Table 2 according to their achievement. Four groups are created as follows: NBA winners, NBA runner ups, NBA quarterfinalists, and rest of teams played in NBA playoff. Comparison between NBA winners and their runner ups in Table 2 shows that Mean PCT of NBA winner is higher than runner up teams indicating the relationship between winning the NBA and season winning percentage. NBA winners also have higher Mean PTS and lower Mean DPTS than runner ups.
Table 2: Analysis of Data according to NBA Position in Playoffs
Team | N | Mean PCT | Mean PTS | Mean DPTS |
NBA Winner | 9 | .7478 | 99.9900 | 92.7895 |
NBA Runner up | 9 | .6833 | 99.2667 | 93.7712 |
NBA quarter final | 18 | .6944 | 99.5251 | 94.0087 |
Rest of Playoffs | 223 | .4672 | 95.8690 | 96.8267 |
Total | 259 |
- Regression Analysis
The relationship between offense represented by PTS, defense represented by DPTS, normal season winning percentage represented by PCT, and Winning playoff games represented by NOWINS, is tested using the Ordinary Least Squares (OLS) regression analysis.
2.1 Regression Analysis I: Winning NBA vs. Offense
Regression analysis was applied with the independent variable average points scored during normal season (PTS) and the dependent variable number of playoff wins (POWINS). Table 3 summarizes the results of the regression analysis.
Table 3: Regression results for POWINS as a function of PTS
Independent Variable
Normal Season Points Scored (PTS) |
Dependent variable:
Number of Playoff games Won (POWINS) |
Constant | -15.2773 |
Coefficient PTS | 0.172564 |
T Stat | 4.635 |
R^{2} | 0.077151 |
Adjust R^{2} | 0.07356 |
Standard Error | 2.913131 |
Observations | 259 |
Table 3 shows adjusted R-squared of .074 which means the independent variable the average points scored during normal season predicts 7.4% of winning games in the playoffs and winning the NBA Championship. T-stat for this variable is 4.635 so it can be seen as statistically significant however it doesn’t seem to explain much. Winning NBA Championship can be estimated using the equation:
Playoff Wins(POWINS) = -15.2773 + 0.1725(PTS) + E
2.2 Regression Analysis II: Winning NBA vs. Defense
Regression analysis was applied with the independent variable average points scored from opponents during normal season (DPTS) and the dependent variable number of playoff wins (POWINS). Table 4 summarizes the results of the regression analysis.
Table 4: Regression results for POWINS as a function of DPTS
Independent Variable
Normal Season Opponent Points Scored (DPTS) |
Dependent variable:
Number of Playoff games Won (POWINS) |
Constant | 15.38339 |
Coefficient DPTS | -0.14554 |
T Stat | -4.01425 |
R^{2} | 0.059002 |
Adjust R^{2} | 0.05534 |
Standard Error | 2.941638 |
Observations | 259 |
Table 4 shows adjusted R-squared of .059 which means the independent variable the average points scored from opponents during normal season predicts 5.9% of winning games in the playoffs and winning the NBA Championship. T-stat for this variable is -4.014 so it is statistically significant however it has a limited ability to predict winning the NBA championship. Winning NBA Championship can be estimated using the equation:
Playoff Wins (POWINS) = 15.38339 + 0.14554(DPTS) + E
2.3 Regression Analysis III: Winning NBA vs. Offense and Defense
Regression analysis was applied with the independent variables average points scored (PTS) and average points scored by opponents during normal season (DPTS) and the dependent variable number of playoff wins (POWINS). Table 6 summarizes the results of the regression analysis.
Table 5: Regression results for POWINS as a function of PTS & DPTS
Independent Variable
Normal Season Points Scored (PTS) &
Normal Season Opponent Points Scored (DPTS) |
Dependent variable:
Number of Playoff games Won (POWINS) |
Constant | -1.17317 |
Coefficient PTS | 0.355392 |
T Stat PTS | 9.187244 |
Coefficient DPTS | -0.32916 |
T Stat DPTS | -8.82269 |
R^{2} | 0.292327 |
Adjust R^{2} | 0.286799 |
Standard Error | 2.55598 |
Observations | 259 |
Table 5 shows adjusted R-squared of .292 which means the independent variable the average points scored and average point scored by opponents during normal season predicts 29.2% of winning games in the playoffs and winning the NBA Championship. T-stat for this variable is 9.1872 and -8.823 which means both variables are statistically significant and PTS is a little more significant than DPTS. This is significantly better than the first two regressions. However, it is still rather low. Winning NBA Championship can be estimated using the equation:
Playoff Wins (POWINS) = -1.17317 +0.35539(PTS) – 0.32916(DPTS) + E
Second, the relationship between winning games is analyzed in relation to offense and defense.
2.4 Regression Analysis IV: Winning Games vs. Offense
Regression analysis was applied with the independent variable average points scored during normal season (PTS) and the dependent variable number of season wins (PCT). Table 6 summarizes the results of the regression analysis.
Table 6: Regression results for PCT as a function of PTS
Independent Variable
Normal Season Points Scored (PTS) |
Dependent variable:
Percentage of Winning (PCT) |
Constant | -0.89017 |
Coefficient PTS | 0.014423 |
T Stat | 7.721161 |
R^{2} | 0.188292 |
Adjust R^{2} | 0.185134 |
Standard Error | 0.146171 |
Observations | 259 |
Table 6 shows adjusted R-squared of .188 which means that offense represented by the independent variable the average points scored during normal season predicts 18.8% of winning games. T-stat for this variable is 7.721 so it is statistically significant. Winning games can be predicted using the equation:
Game Win Percentage(PCT) = -0.89017 + 0.14423(PTS) + E
The following is the scatter diagram of the PTS relating to PCT. The line representing the equation developed in the previous section is also shown in figure 2.
Figure 1: Scatter Plot of PCT vs PTS with Line Fit
2.5 Regression Analysis V: Winning Games vs. Defense
Regression analysis was applied with the independent variable average points scored during normal season (DPTS) and the dependent variable number of season wins (PCT). Table 7 summarizes the results of the regression analysis.
Table 7: Regression results for PCT as a function of DPTS
Independent Variable
Normal Season Points Scored by Opponent (DPTS) |
Dependent variable:
Percentage of Winning (PCT) |
Constant | 2.055452 |
Coefficient DPTS | -0.01614 |
T Stat | -9.34003 |
R^{2} | 0.253419 |
Adjust R^{2} | 0.250514 |
Standard Error | 0.140185 |
Observations | 259 |
Table 7 shows adjusted R-squared of .253 which means that defense represented by the independent variable the average points scored by opponent during normal season predicts 25.3% of winning games. T-stat for this variable is -9.34 so it is statistically significant. Winning games can be predicted using the equation:
Game Win Percentage (PCT) = 2.055452 + 0.01614(DPTS) + E
The following is the scatter diagram of the DPTS relating to PCT.
Figure 1: Scatter Plot of DPTS vs. PTS
2.6 Regression Analysis VI: Winning Games vs. Offense & Defense
Regression analysis was applied with the independent variable average points scored during normal season (PTS) and average points scored by opponents (DPTS) and the dependent variable number of season wins (PCT). Table 8 summarizes the results of the regression analysis.
Table 8: Regression results for PCT as a function of PTS & DPTS
Independent Variable
Normal Season Points Scored (PTS) & Normal Season Points Scored by Opponent (DPTS) |
Dependent variable:
Percentage of Winning (PCT) |
Constant | 0.527437 |
Coefficient PTS | 0.032799 |
T Stat PTS | 58.304 |
Coefficient DPTS | -0.03308 |
T Stat DPTS | -60.9773 |
R^{2} | 0.947714 |
Adjust R^{2} | 0.947305 |
Standard Error | 0.037171 |
Observations | 259 |
Table 8 shows adjusted R-squared of .948 which means that offense represented by average points scored and defense represented by average points scored by opponent predicts 94.8% of winning games. T-stats for these variables are 58.304 and -60.9773 so they are both statistically significant with defense being little more significant than offense. Winning games can be predicted using the equation:
Game Win Percentage (PCT) = 0.527437 + 0.032799(PTS) – 0.03308(DPTS) + E
- Conclusion:
In predicting the winning of NBA playoff games in relation to offense and defense, three regression analyses were completed. The results of the first regression show that offense predicts 7.4% of winning playoff games and the NBA Championship. The results of the second regression show that defense predicts 5.9% of winning the NBA Championship. Offense edge over defense in predicting winning NBA championship is negligible and can be neglected. Results from third regression shows that offense and defense together predict 29.2% of winning the NBA championship. Therefore championships are not won by defense but they are won by both defense and offense.
In predicting the winning of basketball games in relation to offense and defense, another three regression analyses were completed. The results of the fourth regression show that offense predicts 18.8% of winning the NBA Championship. The results of the fifth regression show that defense predicts 25.3% of winning the NBA Championship. Defense is little better at predicting winning of games than offense, but the difference is still negligible and can be neglected. Results from sixth regression shows that offense and defense together predict 94.8% of winning basketball games. Games are not won by defense but they are won by combination of both defense and offense.
It is concluded that “Offense and defense wins (94.8%) games and offers a fair chance (29.2%) in winning Championships.”