STATS WITHOUT THE SPIN · CAREER AVERAGES
Jalen Brunson and the Late-Bloomer Problem
A career average can be mathematically correct and still be the wrong number for the question you are asking.
The average remembers every version of a player
Jalen Brunson entered the NBA as a second-round pick and spent his first four seasons in Dallas. He developed from a reserve into a starter before taking on a much larger scoring and creation role in New York. His career average blends all of those versions together.
If the question is “What has Brunson produced across his entire NBA career?” that blend is appropriate. If the question is “How good is Brunson now?” early seasons with fewer minutes, fewer shots, and less responsibility can pull the answer away from the player he became.
Development is not a statistical error
Career numbers are not misleading because they include development years. They become misleading when fans use them without identifying the timeframe. Late bloomers, role changes, injuries, trades, and coaching decisions all affect what appears in a career line.
The opposite problem also exists. Selecting only a player’s best month or season can exaggerate his normal level. That is why I prefer a recent multi-season sample when evaluating the current player and a complete career sample when evaluating historical achievement.
Role explains part of the change
More responsibility can reveal skills that were always developing, but it can also make efficiency and decision-making harder to maintain. A fair evaluation asks what changed: minutes, usage, teammates, spacing, coaching, health, or the player’s own skill level.
Brunson is valuable as an example because his growth is visible in both opportunity and performance. The lesson is larger than one player: do not let a convenient single number erase the timeline that produced it.
Use the correct window
Facts Over Feelings shows career information because the All-Time / Career tool evaluates the entire body of work. Peak Season exists separately because a career average cannot answer a peak question. For an active late bloomer, both views are necessary.