Basketball explained
Points per game lies: six stats that tell you if a player is good
Scoring average is the most quoted number in basketball and the least informative. Six statistics that tell you what a player is actually worth, with the benchmark for each.
Here is the whole argument in one comparison from the 2025-26 season, taken from Basketball-Reference:
| Points per game | True shooting | Usage | Box plus/minus | |
|---|---|---|---|---|
| Jaylen Brown | 28.7 (4th in the league) | .573 | 36.2% | +3.3 |
| Nikola Jokic | 27.7 | .670 | 30.4% | +14.2 |
One of those players finished fourth in the scoring race. The other was arguably the best player alive. Points per game told you almost nothing, because it counts production and ignores what the production cost.
Six numbers fix most of that. None of them is hard.
1. True shooting percentage — the one to learn first
True shooting folds twos, threes and free throws into one efficiency number:
points ÷ (2 × (field goal attempts + 0.44 × free throw attempts)). It is the honest
answer to "how many points does he get out of a shooting possession".
The league average has been remarkably stable: .581 in 2025-26, .576, .580 and .581 in the three seasons before it. So the benchmark is easy to carry around.
| True shooting | What it means |
|---|---|
| Below .540 | Inefficient. On high usage, actively harmful |
| .540–.580 | Below to about average |
| .580–.620 | Solid to good |
| .620–.660 | Genuinely efficient, and rare on heavy volume |
| Above .660 | League-leading — Jokic's .670 territory |
Effective field goal percentage, (field goals + 0.5 × threes) ÷ attempts, does the same
job for jump shooting but ignores free throws entirely. Use true shooting unless you specifically want to exclude the
line.
2. Usage rate — the number that makes efficiency meaningful
Usage is the share of his team's possessions a player ends. Five players share 100%, so the average is exactly 20% and anything else is a statement about role. Brown's 36.2% was one of the highest in the league; Jokic finished possessions at 30.4% while shooting far better.
Efficiency and usage only mean something together. A 65% true shooting at 14% usage is a useful role player. The same efficiency at 32% usage is a superstar, because nobody sustains that as the focus of a defence.
3. Offensive and defensive rating, and net rating
Points produced and allowed per 100 possessions. League average sat around 115–116 in 2025-26, which is useful context on its own: a defence allowing 110 is excellent, one allowing 120 is not. Net rating is the difference, so the league average is exactly zero by construction.
The caveat is that individual defensive rating leans heavily on the team around a player. Basketball-Reference says as much in its own documentation: a perimeter defender who does not generate steals tends to be undervalued by it.
4. Box plus/minus — the best single summary in a box score
BPM estimates a player's points per 100 possessions above league average, adjusted for team quality. The published scale is worth memorising:
| BPM | Level |
|---|---|
| +8 | MVP candidate |
| +6 | All-NBA |
| +4 | All-Star |
| +2 | Good starter |
| 0 | Average starter |
| −2 | Replacement level |
Its defensive half is the weak half, and the documentation says so directly — treat defensive BPM as a guide, not a measurement.
5. PER, and how to use it without being fooled
Player efficiency rating is a per-minute, pace-adjusted production index, rescaled every season so that the league average is exactly 15. John Hollinger built it, and he has been the most honest critic of it: he has written plainly that PER is not a final verdict on a season, particularly for defensive specialists who do not pile up blocks and steals. Later work by Ben Taylor found something sharper still — PER is very stable year to year but consistently misallocates credit, predicting team results worse than BPM and its cousins.
There is also a minutes trap. In 2025-26 Giannis Antetokounmpo posted the highest PER in the league at 32.6 — in 1,039 minutes. Jokic's 32.3 came across 2,265. Rate statistics need a minutes filter before they mean anything.
Same points, different player. Jalen Brunson and Kevin Durant both averaged 26.0 points in 2025-26. Brunson did it on .580 true shooting at 30.4% usage; Durant on .641 at 27.1%. Identical scoring lines, two clearly different seasons — and only one of the two numbers on the back of the card would have told you.
6. The four factors, for judging teams instead of players
Dean Oliver's framework weights what actually decides games: shooting 40%, turnovers 25%, rebounding 20%, free throws 15%. If you only ever check four things about a team, check those, in that order. It is also a decent sanity test for a manager game: any sim where rebounding outranks shooting is not modelling basketball.
What this has to do with ratings in a manager game
A rating is a compression of all of the above. In Backcourt a high three-point rating means a player both takes and makes threes, because the shot mix follows the rating. A high passing and IQ pair shows up as a low turnover rate. The attributes guide has the exact weightings, including how much each one moves the overall number by position.
What a rating cannot carry is context, and that is the part managers underrate. Energy, morale, form and role fit all decide what a rating is worth on the night, which is the game's version of the same lesson these statistics teach: the number is only half the story. The other half is how the number was produced.
If you want to browse the numbers instead of reading about them, every player in Backcourt's pool has a page with his nine attributes and his percentile against his position — start from the ratings index.
Frequently asked
Why is points per game misleading?
Because it counts what a player produced without counting what it cost. In 2025-26 Jaylen Brown finished fourth in the league at 28.7 points a game on .573 true shooting — below the league average of .581 — while Nikola Jokic scored 27.7 on .670. The lower scorer was four times more valuable by box plus/minus.
What is a good true shooting percentage?
The league average has sat around .580 for four straight seasons, so .580 is the line. Above .600 is genuinely efficient, and .640 or better puts a high-volume scorer among the best in the league. Below .550 on heavy usage is the classic empty-calories profile.
What does usage rate tell you?
What share of a team's possessions a player finishes with a shot, a trip to the line or a turnover. The average is 20% by definition, since five players share 100%. A genuine first option lives in the high twenties to low thirties; above 35% you are looking at one of the league's highest-volume creators.
Is PER a good stat?
It is useful and consistent, but it has a known defensive blind spot — John Hollinger, who created it, said so himself — and later testing found that while PER is stable year to year, it predicts team results worse than BPM and similar metrics. Treat it as a scoring-and-production index rather than a verdict.
What is a good box plus/minus?
Basketball-Reference's own scale puts +8 at MVP level, +6 at All-NBA, +4 at All-Star, +2 at a good starter, 0 at an average starter and −2 at replacement level. Its defensive half is the weaker half, which the metric's documentation says openly.
How do these stats map onto a manager game's ratings?
Ratings compress them. A high three-point rating is a player who both takes and makes threes; a high passing and IQ pair is a low turnover rate; rebounding is the game's version of rebound percentage. What a rating cannot show is context — which is why role fit, energy and minutes decide whether a good rating produces a good night.
Try it on your own club tonight
Backcourt is free on iOS and Android. Career mode works offline, and there is nothing to buy that makes your squad stronger.
