Ray Allen shooting NBA league has become one of the most successful sports products ever, partly thanks to perfecting the advanced statistical metrics to absurd detail. Heavy use of computer statistics entered American professional sports in the early 2000s, first in baseball and later spread like a plague to other sports.

Today, even the occasional basketball fans notice how much the game has changed in the last decade and a half. To a die-hard student of the game like myself, it often feels like someone ripped the soul out of the chest of basketball with pliers. The game strategy has become automatic. Players execute their preprogrammed roles, going in and coming off the game with minute precision. There’s very little room for surprise. Even before the season begins, we can tell who will compete in the conference finals with high probability. The analysts hired to crunch numbers figured out 3-pointers and free-throws are the most efficient way to score, so the teams do exactly that. All the fucking time. It can be painful to watch everyone pulling threes like seventh-grade maniacs in the gym class, but apparently, it works. Teams today score more points than ever.1 The games are crazy fast and more fun to watch. At least according to the general public. And it reflects in the most essential metric for the league owners — the revenue almost quadrupled between 2001 and 2019. They gently stripped the game of unnecessary improvisation, making it more efficient and predictable. Predictable, not as uninteresting, more like consistently high-scoring. Fans don’t care for tactical finesse, as long as buckets are raining.

The league is home to crème de la crème basketball players from all around the world. Growing up, all of them destroy competition in little league, high school, and college. But as they move up the ladder, there’s less and less room to dominate. Eventually, by the time they enter the NBA, most of them will have to adjust to a lesser role they can execute consistently. There’s only a handful of LeBrons and Kevin Durants that continue to dominate, every night, at the elite level.

No matter how elite, the teams look for predictable players. Individuals who can produce stable numbers season-in-season-out, so the coaches and team management know what to expect. Athletes who have been in the league for many seasons secure contracts because they repeatedly prove reliable in their role. Flashes of brilliance don’t work in the long run. On the other hand, the presence of 2× NBA champion Kawhi Leonard secures elite offense and defense and guaranteed contention for the title. He routinely delivers, and his contract reflects that. Hiring 36-year old J.J. Redick comes with a promise of 40%+ 3pt FGM on high volume shooting. He’s done it in almost each one of his 15 NBA seasons. Just to name a few. The scouts know everything about every player, so it’s unlikely they’ll exceed their projected limit. But the moment they demonstrate flakiness and inconsistency, they’re one foot out the door. Those who can’t stay consistent at the NBA level must take their talents elsewhere, to lower quality competitions, where they can perform predictably.

Making software has some similarities with professional sports. The technology and the software change all the time, and it’s challenging to stay indefinitely relevant. However, predictability is timeless. Great basketball players, designers, engineers, plumbers, and carpenters have one thing in common – people trust them to deliver. But it’s not only for their nifty dribbles, clean code, or precise woodwork that they’re great – it’s because they show up every time.

  1. In 2003-04 NBA season, on average teams scored 93.4 points per game, compared to 111.8 points per game in 2019-20 (Source).