Aging curves: when basketball players peak and decline
Basketball players improve into their mid-twenties and fade gradually after thirty, but the average hides huge variation. This guide explains how aging curves are built and how to use them wisely.
Every basketball fan carries a rough picture of a career. A teenager arrives with raw talent, improves quickly through his early twenties, reaches a prime somewhere in the middle of his twenties, holds a high level for a few years, and then declines, slowly at first and then more quickly, until he retires. Teams build contracts around that picture. Scouts project prospects along it. Fans argue about when a star has "lost a step." The picture has a name in statistics: the aging curve, a graph of how performance changes with age.
Like most useful pictures, it is true on average and misleading for any single person. Some players peak at 22 and fade by 28. Others keep improving into their early thirties. A few remain productive deep into their late thirties, while some are finished at 30 by injury. Behind the smooth curve on a chart lies a messy collection of individual paths.
This article explains how aging curves are built, what they show about when NBA players typically peak and decline, why the average conceals important differences between types of players, what biases to watch out for, and how to use the idea without overstating it. It connects to our discussions of regression to the mean and of how draft prospects are evaluated with data, since projecting a young player and evaluating an older one are two sides of the same question.
How an aging curve is built
At first, it sounds easy: take every player, record his production at each age, and average. If you do that, however, you get a misleading curve, for a reason that is fundamental. The players who are still in the league at 35 are not a random sample of players; they are the ones who were good enough to stay. Players who declined faster or less successfully have already left. The older ages therefore contain only the survivors, and the average at age 35 looks better than a typical 35-year-old's would have been. This is survivorship bias, and it flattens the apparent decline.
The standard way around it is the delta method. Instead of comparing all players at age 25 with all players at age 30, you follow the same players from one year to the next. For each player who played at both 29 and 30, you compute the change in his production between the two seasons. Then you average those changes across everyone at that age transition. Stringing the average changes together produces a curve that reflects how an individual player typically moves from year to year, instead of a comparison of different groups at different ages.
The method has its own difficulties. Players who play at 30 but not at 31 are missing from the 30-to-31 comparison, which still leaves a bias toward those who held up. Players who barely played in one season give noisy changes. And the choice of performance measure matters enormously: a curve for points per game looks different from one for efficiency, which looks different from one for an all-in-one rating. Analysts using slightly different assumptions reach slightly different curves, and none is the final word.
What the typical curve looks like
With those cautions in mind, a consistent general pattern appears across studies.
- Rapid improvement through the early twenties. Players who enter the league at 19 or 20 are usually far from their ceiling. They add strength, polish their skills, learn the speed of the professional game and gain the experience to read defenses. The biggest year-to-year gains in value come in this stretch.
- A plateau in the mid-to-late twenties. Most analyses place the peak in overall production somewhere around the age of 25 to 28, with a fairly flat top rather than a sharp point. A player in this window is usually as good as he will be, or very near it.
- Gradual decline after the age of 30. Decline is typically slow at first, perhaps a small fraction of a player's value each year, and then accelerates in the mid-to-late thirties. The decline in a given year is usually modest compared with the year-to-year noise in a player's numbers, which is why a single down season rarely proves a player is finished.
You will see variations in the details. The peak age is later if you measure by skill-based contributions such as shooting efficiency and passing, and earlier if you measure by athletic contributions such as rebounding, blocks and steals. It is later if you count only players who stayed in the league a long time, and earlier if you account for those who flamed out. The principle holds: a long climb, a short flat top and a slow slope down.
An important caveat is the era. The age at which players enter the league, the way they are trained, the care they receive and the way minutes are managed have all changed over the decades. Players now arrive young, and teams actively manage workloads. Our article on load management describes one of the modern tools. These changes may stretch careers, though they make comparisons across eras hard.
Different skills age at different speeds
The average curve is a blend of components that age differently. Understanding the components is more useful than memorizing the blend.
| Type of contribution | General aging pattern |
|---|---|
| Explosive athleticism: leaping, quickness off the dribble, recovery speed | Peaks early, declines earliest |
| Rebounding and shot blocking | Tied partly to athleticism and size; decline moderately early |
| Shooting touch and free-throw accuracy | Tends to hold up well, sometimes into the late thirties |
| Passing and decision-making | Often continue to improve with experience and hold late |
| Defensive positioning and anticipation | Can compensate for lost quickness for a time |
| Durability | Gets worse with age on average, with huge individual variation |
This is why two players of the same age can follow different paths. A slashing guard who lives on speed and explosiveness off the dribble loses value as his athletic edge shrinks, because his game rests on a skill that fades early. A shooter who thrives on spacing and timing may hold his value much longer, because the main tool is mostly intact. A big man whose game rests on touch, footwork and positioning may age better than a rim-running center who depends on vertical explosion.
History offers some well-known illustrations. Kareem Abdul-Jabbar remained a productive center into his early forties, helped by a skyhook that did not depend on leaping and by exceptional conditioning; see our profile in Kareem Abdul-Jabbar: career, style and legacy. Karl Malone won the Most Valuable Player award at 35, a testament to how well a strong, skilled and relentlessly conditioned power forward could age, a story told in Karl Malone: career, style and legacy. Tim Duncan, whose game was built on fundamentals and positioning rather than flash, remained an elite big man into his late thirties, as discussed in Tim Duncan: career, style and legacy. In the WNBA, Sue Bird played until she was 41, a guard whose game was built on vision, shooting and command of the offense. These cases are not typical; they are the survivors, and they remind us that the shape of a player's game helps determine how he ages.
Why individual variation is so large
If the curve is an average, how wide is the spread around it? Very wide. The year-to-year change in any player's production is a mixture of three things: the age effect, which is small and steady; random fluctuation, which is large; and genuine events, such as injuries, role changes and coaching changes, which can be large too.
The typical annual drift due to age in a player's late twenties and early thirties is small compared with the noise in his numbers. A player who drops by a few percent in efficiency from one year to the next may be aging, or he may be having a normal variation. This is the same problem discussed in small samples. Only when the decline persists over several seasons, and when the underlying inputs such as shot selection, athleticism measures and minutes also change, can you say with confidence that age is the cause.
Injuries add another layer. A serious injury can reduce a player's athleticism at any age, and its effect on the curve is hard to separate from aging. Injuries are also more common as players accumulate minutes, so older players face a higher chance of missing time. A player with years of heavy mileage may decline faster than one who entered the league later or played fewer minutes, though the evidence for any particular "mileage" effect is mixed.
There is also the matter of role. As players age, many shift from creator to supporting roles: they take fewer shots, handle the ball less, guard less demanding opponents. Their per-minute production may stay solid even as their role shrinks, and what looks like a decline in totals can be a deliberate choice. A veteran who accepts a smaller role and a high-percentage shot diet might be more efficient than ever.
Beyond the NBA
Aging curves are usually drawn from NBA data, because that is where the numbers are deepest, but the same logic applies elsewhere, with different shapes.
In college basketball, the curve is compressed into a few years. A freshman and a senior are separated by a stretch of rapid physical and skill development, so older college players are expected to be more productive than younger ones, and the right comparison is always by age. That is why scouts pay attention to a prospect's age relative to his class: a young freshman who matches an older teammate's production is showing more promise than the raw numbers suggest. Production at a young age is one of the more reliable pieces of evidence that a player still has room to grow.
In the WNBA, the pattern is similar in outline, but many players arrive in the league at 21 or 22 after four years of college, and stars often remain highly productive into their thirties. Several of the league's best have stayed at an elite level well past the age at which NBA players typically begin to decline, which fits the idea that skill-based games age gently. The shorter seasons and smaller samples mean the curves are less precise, so the caution about noise matters even more. You can follow the league through the WNBA section of the site.
Internationally, players develop in club systems that vary by country, often turning professional in their teens, and national-team careers can stretch across many years. Aging patterns there are harder to measure because the data are scattered across competitions of different strength.
Reading a veteran: a short worked example
Suppose a wing is 33, and his scoring average has slipped for two seasons running. Should you conclude that he is declining? Walk through the questions.
First, what happened to his minutes and role? If he moved from the first option to the third, scoring would fall without any loss of ability. Second, how did his efficiency move? If his shooting percentages held steady while his volume dropped, the decline is in usage, not in skill. Third, what happened to the athletic components? If his rebounding and steals per minute have fallen, and his free-throw attempts have dropped because he drives less often, there is evidence of lost explosiveness. Fourth, did he miss time? Injuries can explain a bad stretch without telling you about the curve ahead. Fifth, how large is the change compared with the noise? A drop of two points per game could be a sample.
Having answered those, you might conclude that he has shifted from a high-volume creator to an efficient spot-up scorer who still defends his position, a change in role more than a collapse in value. Or you might find that every athletic indicator has dropped and his shot attempts at the rim have fallen away, which points to real decline. The same age and the same scoring numbers can lead to different conclusions depending on the answers, and that is exactly why the average curve should start the analysis, not end it.
Biases and traps in using aging curves
Several mistakes recur when people apply aging curves.
- Treating the average as destiny. The curve says what usually happens, not what will happen to a particular player. Using it to dismiss a 33-year-old without looking at his skills and conditioning is lazy; using it to assume a 24-year-old will keep improving is overconfident.
- Ignoring survivorship. Statements such as "players improve until 30" can come from an analysis that includes only players who lasted. The honest version considers those who left.
- Confusing age with experience. Some improvements in the early years are due to experience, not physical growth, and some players develop late because they started playing seriously late. Comparing players by age alone misses this.
- Applying a single curve to every skill. As the table shows, different abilities age differently. A projection should be built from the components, not from a blended total.
- Forgetting regression. A player who has an exceptional season at 27 is likely to fall back, not because of age but because of luck. Mixing up regression and aging produces exaggerated stories about early decline.
- Overreading a year. One down season at 31 is common even for a player who has not declined. A trend of several seasons, supported by changes in the inputs, is better evidence.
How the idea is used
Teams use aging curves for contract decisions: how much a player's expected production will change over the length of a deal, how much risk comes with a long commitment to a veteran, and how much patience a young player deserves. They also use them in the draft, where a player's age is an essential part of the evaluation. A 19-year-old who produces at a given level in college is generally viewed more favorably than a 22-year-old with the same numbers, because the younger player has more room to grow. Our article on how draft prospects are evaluated with data develops this point.
Fans can use the idea in simpler ways. When you read a projection of a team's next season, check how the model treats age. A roster full of players in their early twenties has more upside; a roster full of players past 32 has more downside, particularly at positions that rely on athleticism. When evaluating a debate about a veteran, ask whether the signs of decline are in the athletic components or the skill components, and whether his role is changing. Stats like win shares, BPM and VORP can summarize overall value, but the component statistics reveal what is changing.
You can compare players of different ages on the players page and see how current production stacks up on the leaders page.
What to take away
The aging curve is a useful summary of a long process: rapid growth in the early twenties, a plateau in the middle of the decade, and a slow, then faster, decline after thirty. Behind it are skills that age at different speeds, athletic gifts that fade early and craft that lasts, along with injuries, role changes and luck that make every career different. Treat the curve as a starting assumption, adjust it for the kind of player you are looking at, and be wary of both easy stories about decline and easy faith in permanent youth. The best forecast comes from combining the average pattern with the details of the player in front of you.