Revisiting the Four Factors
Is a 20-year old analysis still relevant?
The Four Factors are four (really eight: each factor is applied to both offense and defense) team stats that have been shown to be highly correlated with winning basketball games. They are, in order of importance:
Effective Field Goal % (eFG%) - the percent of shots you make, juiced up to account for the added value of a three
Turnover % (TOV%) - how often you turn the ball over
Offensive Rebound % (ORB%) - how often you get the offensive rebound when it’s available (e.g., your team misses a shot)
Free Throw Attempt Rate (FTA Rate) - the ratio of free throws taken to field goals taken
Dean Oliver first described the importance of these four factors to basketball success around 2002 while working as a quantitative risk analyst for groundwater remediation projects. He refined and published the results in his book Basketball on Paper, which took off and launched Oliver’s career in basketball that has taken him to four franchises and his current role at ESPN.
But 2002 was a long time ago: LeBron wasn’t in the NBA; Michael Jordan was. Victor Wembanyama and the Thompson twins hadn’t even been born yet. And the game was played very differently: In 2001-02 the Dallas Mavericks led the league with 105 points per game; that would be tied for last in 2024-25, when the top team scored 122 points per game. Two major changes in play-style1 have driven this massive increase in scoring:
Threes: The Boston Celtics led the league in 3 pointers attempted in both 2001-02 and 2024-25 … but with twice as many attempts in ‘24-25. The Philadelphia 76ers made just 2.6 threes per game in ‘01-02; there wasn’t a single game this year in which a team made fewer than 3, and that was only because the Pacers missed 27 attempts in a game.
Pace: The game has sped up, with the average team getting eight more possessions per game this year. Again the fastest pace team in ‘01-02 would be the slowest in ‘24-25. Teams scored 270 more fast break points this year than they did in ‘01-02!
This have driven significant shifts in the Four Factors:
With all that change, you might think the Four Factors aren’t as indicative of success as they were when Dean Oliver first organized them over 20 years ago. It is true that teams and analysts rely on much more advanced statistics these days …
… but the Four Factors are still as powerful as ever.
The Four (Eight) Factors accounted for 96% of the variance in team wins2 this season. That’s actually up from 86% in ‘04-05. Turns out that making shots, protecting the basketball, grabbing rebounds, and getting free throws is still a winning strategy. Shocking!
Nevertheless, I was curious if I could find another set of factors that had a strong correlation with winning basketball games in 2025, accounting for some of the stylistic changes in the way the game is played. I looked into two groupings:
Shot Geography
The analytically minded offenses of today’s NBA aim to optimize shot selection to maximize points per field goal attempt. As a reference point, teams average ~1.08 points per FGA, but the best shots that teams try to take and make as much as possible are:
At the free throw line (1.55 points per trip3)
At the rim (1.33 points per FGA)
From three, ideally the closer range corner threes (1.16 points per FGA)
If we put these factors4 into a linear regression, we can explain 59% of the variance in wins. Not bad, but not nearly as powerful as the original Four Factors. What is interesting is that these factors would have explained just 48% of wins in 2005.
Movement Factors
The second major change I explored was movement. Not the LeBron-spurred player movement between teams (that’s for another time) but players and the ball literally moving around the court. I measured this using pace (number of possessions per 48 minutes), fast break points as a share of total points, average player speed (tracking data is fun!), and percentage of field goals assisted, to measure ball movement.
Again, the explanatory power is less than the original Four Factors, at 46%, but it is up from explaining just 19% in ‘04-05, a doubling of explanatory power in two decades.
So, I’m sorry to say I did not find a new set of crazy predictive stats that will get me a job with the Nuggets tomorrow. But the change over time in the predictive power of these shooting and movement stats is interesting, and I think shows how much the game has changed and how teams are seeking out advantages beyond making more shots and getting more rebounds.
Same Sport, Different Game
NBA Playoffs
As I and many others have noted, the game slows down in the playoffs. Teams are fully locked in, and every possession matters that much more. So I wondered, how does the explanatory power of the Four Factors change?
In the two-season sample I explored, it’s mixed. In ‘04-05, the Four Factors jumped from explaining 86% of wins in the regular season to 97% in the playoffs, while in ‘24-25 we’ve seen the exact opposite, falling from 96% to 83% through Game 3 of the Conference Finals. People are saying these playoffs have been a throwback to 90s era play styles, and I suppose that’s even showing up here.
WNBA
There are differences in the women’s game vs the men’s as well, and it turns out the Four Factors are even more powerful indicators in the WNBA, explaining a full 98.2% of variance in wins during the 2024 season.
In terms of Four Factor priority, shooting is more important in the WNBA. eFG% and FTA Rate make up 61% of the Four Factor power in the WNBA vs 52% in the NBA5, with TOV% and ORB%6 correspondingly lower. The priority order is the same:
Non play-style things such as rule changes, players’ abilities, and an increased focus on analytics have also contributed to more scoring.
That is, the Adjusted R^2 for a linear regression model using the four factors to predict wins is 0.96.
Uh yeah, free throws are crazy efficient, unless you’re Deandre Jordan.
Namely percentage of field goals taken from three, percentage of three point attempts made, percentage of three point field goals taken from the corner, percentage of field goals taken from the restricted area, and free throws attempted per field goal attempted, all for a team’s offense and defense (the opponent’s offense).
To cherry pick an example, the 2025 Chicago Sky have an ORB rate 2x better than the NY Liberty. The Sky are 0-4, the Liberty are 4-0.
We still love you Deandre Jordan ❤️