Statistics is all about using the past to explain the present and predict the future. You can’t compare players very effectively without measuring how they performed at various aspects of the game. But, even still, you can’t compare players very effectively with basic statistics like points per game (PPG). Is Kobe better than LeBron? Well, he averages less… Arguments like these fall apart because of the following question: what makes one statistic a better measure of a player’s value than another? Is PPG more valuable than PP48Min? Which is worth more- stealing the ball, or blocking it?
It all comes down to how much a statistic is correlated with winning percentage; in other words, how much does Player X’s superiority at Stat Y allow his team to win? This is the reason analysts don’t cite statistics like “Minutes” to say Dwight Howard is better than Patrick O’Bryant. A player’s “Minutes” simply aren’t very correlated with his team’s winning percentage. In fact, almost none of the basic NBA recorded stats (points, assists, steals, blocks, turnovers, rebounds, etc.) are correlated enough with winning percentage to mean anything. Rather, it takes clever combinations of these various statistics to come up with meaningful measures- that’s the job of basketball analysts (APBRmetricians, if you will).
In his book Basketball on Paper, Dean Oliver notes how four “advanced” statistics can predict about 99% of a team’s wins (and thus, winning percentage). What are these stats?
1. Effective Field Goal Percentage (eFG%)
As Dean Oliver succinctly puts it, “shooting the ball is the most important” factor to a team’s winning percentage. You don’t win without scoring, and you don’t score without shooting well. Effective Field Goal Percentage differs from normal FG% in that it gives some credit to hitting threes (after all, threes are more valuable than twos). It’s easily calculated as [eFG% = (FGM + 0.5*3FGM)/FGA]. The Phoenix Suns shot a league best 55.1 eFG% last year, with the Pacers last with 47.4 eFG%.
2. Turnover Rate (TOr)
Seems obvious: you can’t shoot the ball if you don’t have it in your hands. Turnover rate measures how often a team gives away the ball on a per possession basis. When you take a bad shot, it may have a 10 to 20% chance of going in, but when you turn the ball over, there is 0 probability it ends up in the hoop. Turnover Rate is calculated as [Turnovers/Possessions] where Possessions is equal to [FGA - OR + TO + 0.4 * FTA].
3. Offensive Rebounding Percentage (OR%)
Now let’s say you didn’t turn the ball over, managed to get off a shot, and it still didn’t go in. At this point, the possession can only be kept alive by one thing: an offensive rebound. Offensive Rebounding Percentage basically measures how many offensive rebounds you pulled down out of all the offensive rebounding opportunities you had. Another way to think of it: all the offensive rebounds you don’t pull down are received by your opponent as defensive rebounds. Therefore, offensive rebounding percentage is easily calculated as [OR / (OR + Opponents Def Reb)]
4. Free Throw/Field Goal Percentage (FT/FG%)
According to Dean Oliver, “the biggest aspect of “free throws” is actually attempting them, not making them.” Over the course of NBA history, teams that get to the line are generally the ones that enjoy success. What does FT/FG% mean? It documents how many times you draw fouls versus how many total shots you attempt. It’s simply [FTA/FGA].
It’s important to note that these four statistics apply on the offensive and defensive ends. You might shoot 60% from the floor, but if you allow your opponent a ton of offensive boards, things might not end well.