Arkansas Fight’s Glossary of Advanced Basketball Stats

Meet all of our advanced stats

By Adam Ford Oct 22, 2019, 8:58am CDT

Confused by any of the stats you’ve seen at Arkansas Fight? You’ve come to the right place!

This glossary has two parts: team stats and individual stats.

Team Stats

Basketball is much simpler than football, so if the EV+ system we use for football stats confuses you, then don’t worry, this is much easier to grasp.

PPP, EPR, and TS

The ultimate goal of basketball is to get the ball in the hoop more than your opponent. But comparing raw points isn’t good enough because teams play at different paces. We need a universal number to track how good a team is at scoring. Points per possession (PPP) is that number. Because both teams have the same number of possessions per game (or at least within one), then the winner of PPP will tell you who won. If a team scores 82 points on 67 possessions, then they scored 1.22 PPP. If another team in a different game has 85 points on 74 possessions, then they may have had more points, but they had a lower PPP: 1.15 PPP. Pretty simple, right?

Of course, PPP isn’t descriptive enough if the goal of our analysis is to learn about the teams involved. So we can break PPP down into parts: creating shots and making shots. Everything that happens on the basketball court is in support of these goals.

Creating shots includes everything that is valuable but not worth points. Rebounds and turnovers, mostly. Those things aren’t worth points, but they give a team a chance to score points (or end an opponent’s possession without points). The statistic that measures a team’s ability to generate shots for itself is effective possession ratio (EPR). EPR is simply the number of shot opportunities per possession.

The formula is simple:

EPR = (Possessions + Off. Rebounds - Turnovers) / Possessions

Offensive rebounds give you another chance to shoot, so they make EPR go up. Turnovers end a possession without a shot, so they make EPR go down. In that sense, rebounds and turnovers do the same thing. If you want to see more how EPR works, read my recap of Arkansas’ 2018 NCAA Tournament loss to Butler, where I point out that despite Arkansas being out-rebounded 45-25, rebounding wasn’t the problem in the loss:

Because rebounds and turnovers are essentially the same thing, we combine them into a single stat: Effective Possession Ratio, which is the ability to convert possessions into shots. Grabbing offensive boards and avoiding turnovers help a team’s Offensive EPR go up; forcing turnovers and securing defensive rebounds help a team’s Defensive EPR go up. The goal of EPR is maximize the number of shot opportunities.

But here’s the thing: Arkansas actually had more shot opportunities than Butler. The Hogs had 66 shot opportunities. Butler had 65.

***

Butler got more defensive rebounds because Arkansas missed more shots ... Arkansas didn’t get many defensive rebounds because Butler didn’t miss many shots. Butler got a lot of defensive rebounds because Arkansas missed a lot of shots. What looks like a rebounding problem was, in fact, a shooting problem.

Now that we understand creating shots, we can move on to making shots. Total field goals made doesn’t tell us much: some shots are worth 2, others are worth 3, and a team that gets 50 free throws probably doesn’t need to make many field goals to win. Since you don’t score on a possession that ends in a turnover, and you could get multiple shots on a possession where you get an offensive rebound, then we can refine points per possession to points per shot opportunity, also known as true shooting (TS). Here’s that formula:

TS = Points / (Field Goal Attempts + 0.44 * Free Throw Attempts)

The denominator is shot opportunities. Why the 0.44 coefficient on free throws? Because, according to stats, about 44% of free throws actually end a possession. The other 56% are the first of two shots or an and-one, where the possession is recorded as a field goal.

If a team turns the ball over 24 times but wins anyway, then it’s probably because they had a very high TS. That’s how the EPR/TS breakdown allows us to dig deeper into how a game went. We get PPP with a simple formula:

PPP = EPR * TS

All the basketball team stats we do start with the big three stats: PPP, EPR, and TS.

Other team stats

These “supporting stats” help us get a better understanding of EPR and TS.

Effective Field Goal % is field goal percentage adjusted for the fact that three-pointers are worth more. If a player goes 4 of 8 from the floor and all shots are two pointers, he had a 50% field goal percentage: 8 points on 8 shots, divided by the two-point value of each shot. But if two of those shots are three-pointers, then he actually got 10 points on 8 shots. Ten divided by 8 divided by 2 is 62.5%, so that’s his effective field goal percentage. EFG is basically saying this: if all shots were worth 2 points, what percentage would you have to shoot to get the same number of points? In our example, you’d have to shoot 62.5% (5 of 8) to get those 10 points. So going 4 of 8 with two made 3’s is the same is going 5 of 8 on all 2’s. It’s a much more accurate measure of shooting percentage. We’ll still use 2FG% and 3FG% on a granular level, but this is a good summary stat.
Three point rate is the percentage of field goal attempts that are three-pointers. Some defenses try to force 3’s, others try to limit them. Studies show that trying to limit three point rate is the best way to prevent teams from getting hot against you. FTA per FGA is the ratio of free throw attempts to field goal attempts. Teams that get to the line a lot are good at this stat.
Turnover % and Steal % are the percentage of possessions that end in a turnover or steal. Offensive Rebound % is the percentage of each missed shot that is rebounded by the offense. For the defense, it’s Defensive Rebound %. For all missed shots (offense or defense), it’s Total Rebound %. Assist %, meanwhile, the percentage of made field goals that were assisted.
Assist-to-turnover ratio is... exactly what it sounds like. It’s a good measure of the quality of your ballhandling.
Floor % is the percentage of possessions where the offense scores at least 1 point. Teams with low floor percentage often go into long droughts and may be dependent on three-pointers. Floor percentage is not adjusted for the number of points actually scored (that’s PPP), but is instead more like our “success rate” stat in football.

Predictive stats

We can preview games and predict final scores in football using the EV+ system, and now we can do the same thing for basketball. All these numbers are opponent-adjusted:

Marginal efficiency is PPP adjusted for opponent. If a team records 1.25 PPP against an opponent that was allowing just 1.06 PPP, then their marginal efficiency was +17.9%, meaning they did 17.9% better than the average allowed by that team.
Marginal EFG%, marginal 2FG%, marginal 3FG%, marginal offensive rebound %, and marginal turnover % are all opponent-adjusted numbers that support marginal EPR and marginal TS. Once we’ve opponent-adjusted everything, we can directly compare teams that have played vastly different schedules.

Individual stats

There are two kinds of ways to evaluate individuals: overall and single-stat.

Overall

The methodologies for these are very complicated, so look out. They are too long to explain here:

Offensive rating is the number of points produced by a player per 100 possessions. Produced isn’t the same as scored: it includes scored, but it also includes the value created by grabbing offensive rebounds and dishing assists... and subtracts the value lost by turning it over. The formula is really, really long. A similar stat, defensive rating, tells you the same thing but for defense; however, it is very unreliable because there is no way to measure a lot of the good (and bad) things that players do on defense. Steals, defensive rebounds, blocks, and fouls are the only stats we have to go on. DRtg, then, is rarely used.
Player Efficiency Rating (PER) is an advanced stats evaluation mostly used in the NBA. The average is always set to 15, so any PER above 15 is above-average. Get around 35 and you’re at GOAT-level. We won’t use this stat much.
Win shares are the number of victories created by a single player’s stats. It takes all stats, including both offensive and defensive rating, into account. It basically just adds up all the good things a player does, subtracts the bad things they do, and determines how many games could be won based solely on all those numbers (using some all-D1 averages as benchmarks). Add up all the win shares by every player on a team, and you’ll get (roughly) the total number of wins by that team. If you adjust win shares for playing time, you get win shares per 40 minutes. A guy with a high WS/40 but few minutes should probably get more playing time.

Single-stat

All of these statistics are adjusted for playing time:

Assist % is the percentage of made field goals that player assisted while he was on the floor. Offensive Rebound %, Defensive Rebound %, and Total Rebound % are the percentage of available offensive, defensive, and total rebounds that player grabbed while on the floor. Block % is the percentage of two-point attempts that player blocked while on the floor. Steal % and Turnover % are the percentage of possessions that player ended with a steal or a turnover while on the floor. You get the picture, hopefully.
Usage Rate is the overall percentage of possessions ended by that player while they are on the floor. “Ended” here means turned it over or took a shot. Obviously, on a perfectly-balanced team, every player would have a 20% usage rate, since there are five players on the floor at once. So any usage rate above 20% is an above-average player in terms of usage. Usage is not always good: missing a bunch of shots or turning it over a bunch will create a high usage rate without much in terms of results!

Have any questions about this glossary? Drop a comment!