Predicting the 2020 NBA Playoffs

Introduction

For our final project, we decided to predict the teams that would have made it do the playoffs in 2020. We used historical data of games statistics since the 1980 playoffs to base our prediction.

Every regular season, there are 30 teams divided into two conferences: Western and Eastern. Each team plays 82 games per season - 41 home and 41 away. At the end of the regular season, the eight teams with the most wins in each conference qualify for the playoffs.

We decided to compare the results of 3 different models to do our predictions, and see which model would be the most accurate.

Link: Project Repository

NBA History

Why we chose 1980 - major change in the gameplay

Starting in the 1979-80 NBA season the 3-point line was implemented with the intention of increasing the chances for shorter players to earn more points and to expand defense away from the basket.

Source Data

To build our models, we wanted to be able to pull as much data as possible to optimize machine learning. After reviewing various NBA data sites, we chose https://www.basketball-reference.com/.

Reasoning:

• Most robust data tables
• Longest history of data
• High accuracy (spot checked against data on NBA.com and ESPN.com)
• Low level of complexity/Javascript on the website, allowing us to scrape without a web driver (chromedriver)

To scrape the data we used HTML parsing with Python’s Beautifulsoup to build a JSON table per our specified date-range to go year-by-year (e.g. page-by-page) to scrape and parse the data needed from two separate tables housed on each year’s summary page (example: https://www.basketball-reference.com/leagues/NBA_2020.html):

• Conference Standings – Collect ‘conference’, ‘wins’, and ‘losses’
• Team Per Game Stats – Collect all other per-game average statistics

This data was then written into a CSV file for Machine Learning analysis.

Link: Scrape, Data

Building the Models

We developed three different types of models, Logistic Regression, Random Forest, and Support Vector Machine (SVM).

Each performance statistic of the regular season data was scaled using a quantile transformer by each year (all columns except RK, Team, Year, Games, and Conference) and normalized. This transformed the data so the performance statistic was comparable for each season.

We used two variations of game wins - count of wins and losses versus win rate.

We split the regular season data by Eastern and Western conferences. From the split data, 1980 to 2018 was used to train the model. We tested the trained models against the 2019 season. Finally, we predicted 2020 playoff teams based on the statistics from the incomplete season.

Link: Preprocessing (WL) , Preprocessing (W%) , Train/Test Models (WL) , Train/Test Models (W%) , Saved Models

Model Scores

Each model was run with all performance statistics, but with varied structure for game wins. Each model was trained twice: first, with the scaled and and normalized data, then with the count of game wins and losses, and once with the % of wins.

We found that using the scaled and normalized count of wins and losses generally improved the model scores.

Western Conference

	Logistic		Random Forest		SVM
	WL	W%	WL	W%	WL	W%
Train	0.927	0.910	1.000	1.000	0.933	0.927
Test	0.933	0.933	1.000	1.000	0.933	0.867

Eastern Conference

	Logistic		Random Forest		SVM
	WL	W%	WL	W%	WL	W%
Train	0.910	0.888	1.000	1.000	0.926	0.913
Test	0.933	0.867	1.000	1.000	1.000	1.000

Feature Importance

Game outcomes (whether wins, losses, or w%) is the most important feature of all the models.

• Defensive rebounds (DRB) is an important feature for almost all models, for both conferences.
• Steals (STL) and turnovers (TOV) are important features for the logistic and SVM models, for both conferences.
• 2-pointers percentage (2P%) and field goals percentage (FG%) are important features for the random forest models, for both conferences.
• 2-pointer attempted (2PA) is an important factor negatively correlated with outcomes for all models in the Eastern conference.
• Free throws attempted (FTA) is an important factor positively correlated with outcomes with the svm model in the Western conference.

Western Conference

	Logistic		Random Forest		SVM
Order	WL	W%	WL	W%	WL	W%
1	L (-)	W% (+)	W	W%	L (-)	W%(+)
2	W (+)	TOV (-)	L	2P%	W (+)	STL (+)
3	STL (+)	STL (+)	2P%	FG%	STL (+)	DRB (+)
4	TOV (-)	DRB (+)	FG%	DRB	TOV (-)	TOV (-)
5	DRB (+)	FG% (+)	DRB	AST	FTA (+)	FTA (+)

Eastern Conference

	Logistic		Random Forest		SVM
Order	WL	W%	WL	W%	WL	W%
1	L (-)	W% (+)	L	W%	L (-)	W% (+)
2	W (+)	DRB (+)	W	2P%	W (+)	2PA (-)
3	DRB(+)	2PA (-)	2P%	FG%	2PA(-)	DRB (+)
4	2PA (-)	TOV (-)	FG%	DRB	DRB (+)	STL (+)
5	TOV (-)	STL (+)	DRB	2PA	ORB (+)	ORB (+)

Click here to see performance statistic abbreviations.

2019 Predictions

The model predictions for the 2019 season showed that the Eastern Conference models were better at predicting the teams more precisely than the Western Conference models. The teams that made it to the playoffs are listed in rank order under the "2019 Actual" column.

Teams with higher than .5 probability are listed below in descending order. If there were fewer than 8 teams that meet this criteria, the team with the next highest probability was listed until 8 teams were selected.

Red indicates false predictions.

Western Conference

		Logistic		Random Forest		SVM
Rank	2019 Actual	WL	W%	WL	W%	WL	W%
1	GSW	DEN (.98)	DEN (.95)	DEN (.99)	GSW (.99)	DEN (.99)	DEN (.99)
2	DEN	GSW (.95)	GSW (.95)	GSW (.98)	UTA (.93)	POR (.99)	POR (.99)
3	POR	UTA (.95)	UTA (.94)	UTA (.95)	DEN (.90)	UTA (.99)	UTA (.99)
4	HOU	POR (.95)	POR (.91)	HOU (.94)	OKC (.86)	GSW (.99)	GSW (.99)
5	UTA	OKC (.81)	OKC (.79)	POR (.89)	POR (.85)	OKC (.93)	OKC (.95)
6	OKC	LAC (.74)	LAC (.72)	OKC (.89)	HOU (.84)	HOU (.82)	LAC (.80)
7	SAS	SAS (.71)	SAS (.72)	SAS (.86)	LAC (.80)	LAC (.78)	HOU (.77)
8	LAC	HOU (.71)	HOU (.66)	LAC (.85)	SAS (.80)	SAS (.69)	SAS (.74)
9		NOP (.53)	NOP (.62)			SAC (.52)	SAC (.51)
10							NOP (.50)

Eastern Conference

		Logistic		Random Forest		SVM
Rank	2019 Actual	WL	W%	WL	W%	WL	W%
1	MIL	TOR (.99)	MIL (.99)	MIL (1.0)	MIL (1.0)	TOR (1.0)	TOR (1.0)
2	TOR	MIL (.99)	TOR (.99)	TOR (.99)	PHI (.96)	MIL (1.0)	MIL (1.0)
3	PHI	PHI (.98)	PHI (.97)	PHI (.98)	TOR (.95)	PHI (.99)	PHI (.99)
4	BOS	BOS (.96)	BOS (.94)	BOS (.96)	BOS (.93)	BOS (.99)	BOS (.99)
5	IND	IND (.75)	IND (.74)	IND (.93)	IND (.93)	IND (.85)	IND (.84)
6	BKN	ORL (.68)	ORL (.65)	ORL (.90)	BKN (.84)	BKN (.85)	ORL (.75)
7	ORL	BKN (.65)	BKN (.60)	BKN (.88)	ORL (.84)	ORL (.80)	BKN (.74)
8	DET	CHA (.49)	CHA (.52)	DET (.80)	DET (.82)	DET (.67)	DET (.50)

Click here to see team name abbreviations.

2020 Predictions

Given that only four teams had clinched a playoff spot before the season was cancelled, there is no ground truth for this set of predictions. However, we thought it would be interesting to see what could have been.

• The team ranks for 2020 at time of the NBA shutdown is listed under "2020 Partial" column.
• Western conference models predicted more than 8 teams, but the top 8 predicted by the models were the same.
• Eastern conference models predicted the same 8 teams would make the playoffs.

The teams with higher than .5 probability are listed below in descending order.

* indicates teams that have made it to the playoffs.

Western Conference

		Logistic		Random Forest		SVM
Rank	2020 Partial	WL	W%	WL	W%	WL	W%
1	LAL*	LAL (.98)	LAL (.97)	LAL (.99)	LAL (.97)	LAL (.99)	LAL (.99)
2	LAC	DEN (.98)	DEN (.95)	DEN (.98)	LAC (.95)	DEN (.99)	DEN (.99)
3	DEN	LAC (.92)	LAC (.89)	LAC (.97)	DEN (.94)	LAC (.98)	LAC (.98)
4	UTA	OKC (.87)	SAS (.87)	OKC (.97)	DAL (.92)	OKC (.98)	OKC (.96)
5	OKC	SAS (.83)	OKC (.83)	UTA (.96)	MEM (.91)	SAS (.87)	SAS (.93)
6	HOU	DAL (.82)	DAL (.81)	DAL (.90)	UTA (.89)	DAL (.85)	DAL (.84)
7	DAL	HOU (.71)	HOU (.69)	MEM (.90)	OKC (.89)	MEM (.80)	HOU (.79)
8	MEM	MEM (.70)	MEM (.67)	HOU (.89)	HOU (.86)	HOU (.79)	MEM (.77)
9	POR	PHX (.59)	PHX (.62)
10	NOP	UTA (.50)

Eastern Conference

		Logistic		Random Forest		SVM
Rank	2020 Partial	WL	W%	WL	W%	WL	W%
1	MIL*	MIL (.99)	MIL (.99)	PHI (1.0)	MIL (.98)	TOR (1.0)	TOR (1.0)
2	BOS*	TOR (.99)	TOR (.99)	TOR (.99)	MIA (.95)	BOS (1.0)	MIL (1.0)
3	TOR*	BOS (.99)	BOS (.99)	MIL (.98)	TOR (.93)	MIL (1.0)	BOS (1.0)
4	MIA	PHI (.95)	PHI (.94)	MIA (.98)	PHI (.92)	PHI (.99)	PHI (.99)
5	IND	MIA (.65)	MIA (.89)	BOS (.97)	BOS (.89)	MIA (.99)	MIA (.98)
6	PHI	ORL (.64)	ORL (.67)	IND (.94)	IND (.88)	BKN (.84)	BKN (.78)
7	BKN	BKN (.64)	BKN (.63)	ORL (.92)	ORL (.86)	IND (.74)	IND (.70)
8	ORL	IND (.64)	IND (.60)	BKN (.86)	BKN (.80)	ORL (.70)	ORL (.68)

Click here to see team name abbreviations.

2020 FINAL PREDICTIONS

We averaged all six models to determine which teams we think would have made it into the playoffs for 2020. All four teams that had already clinched playoff spots (marked with an asterisk) were at the top of the lists (though not necessarily in the same order).

Western Conference

Rank	2020 Partial	Predicted
1	LAL*	LAL (.98)
2	LAC	DEN (.97)
3	DEN	LAC (.95)
4	UTA	OKC (.92)
5	OKC	SAS (.88)
6	HOU	DAL (.86)
7	DAL	MEM (.79)
8	MEM	HOU (.79)

Eastern Conference

Rank	2020 Partial	Predicted
1	MIL*	MIL (.99)
2	BOS*	TOR (.99)
3	TOR*	PHI (.99)
4	MIA	BOS (.95)
5	IND	IND (.84)
6	PHI	BKN (.78)
7	BKN	ORL (.77)
8	ORL	DET (.62)

Conclusion

Using Logistic, Random Forest and Support Vector Machine as tools of Machine Learning (ML) we found that all provided similar predictions, whether the importance of the independent values were the same or not. Though, if you look at our section ‘Feature Importance’ you’ll see the strongest values are related to either ‘Win’, ‘Loss’, or ‘W%’ (win percentage); which we believe should be expected in predicting a teams likelihood of earning a place within the Playoffs.

When we removed the WL/W% from the models, we found that defensive stats played a much larger role in the model. Steals (STL), defensive rebounds (DRB), turnovers (TOV) all increase in importance while providing similar predictive results and accuracy.

We ran numerous tests for previous years showing similar results, with a high rate of accuracy (with some inconsistency in 1999 and 2011 during NBA lockouts). With that being said, with the amount of data we currently collect and model, we feel that we have built three models that can predict playoff contention with a ‘strong’ level of accuracy of telling who from each conference will participate in the NBA Playoffs.

Next Steps

Improving the Model

If we had more time with these models, there are several more things we’d like to try.

• There is a lot of data we could add to the model, like player info or strength of schedule.
• We could also narrow down the features used in each model to determine what really is important and what isn’t.

Continued Testing of Accuracy

We could use these models year-over-year with mid-season statistics to prove accuracy.

Future Models

The natural next step would be to make a model that uses regular season and playoff game data to predict not only who makes it to the playoffs, but who wins the whole thing.