Can an NBA team really go 82–0?
No NBA team has ever finished a season 82–0. The best regular season on record is the 2015–16 Golden State Warriors' 73–9, which broke the 72–10 mark the 1995–96 Chicago Bulls had held for twenty years. TRUE 82 exists because that question is more fun as a draft challenge than as a trivia answer.
Why 82-0 is close to impossible
Going undefeated in the NBA is almost impossible because you'd have to have a 99.16% chance of winning every given game for an entire season. In a modern offensive environment, it would require a +38.8 net rating, using Daryl Morey's Pythagorean model (Win% = PF13.91 / (PF13.91 + PA13.91)). And that's just to have a >50% chance of it happening. Dealing with the variance from travel, back-to-backs, injuries, cold shooting, getting jobbed by refs, and having a target on your back, means insane variance that needs to be balanced with massive outperformance cushion. Morey's model might break down at the extremes, but it's built for what's possible in real life, and 82-0 simply isn't.
So if you ever feel True82 is too hard, just know we're going easy on you, because our required net rating is down to +27.1. Not to mention your starters get the ability to play a full 48 minutes for all 82 games (Joel Embiid breathes a sigh of relief).
Who came closest to going undefeated in the NBA?
Nobody has even come close.
- 2015–16 Golden State Warriors, 73–9, +10.7 net rating (+25.4 away from >50% chance undefeated)
- 1995–96 Chicago Bulls, 72–10, +13.4 net rating (+23.1 away from >50% chance undefeated)
- 1971–72 Los Angeles Lakers, 69–13, +10.7 net rating (+23.4 away from >50% chance undefeated)
So just know when you get mad at your results for not being 82-0, the system is being fair to you (probably cartoonishly generous, actually).
How TRUE 82 turns the question into a game
Real rosters are stuck with one era, one payroll, and one front office. TRUE 82 removes those limits and adds its own: draft five players from random franchise-and-era tickets — two guards, two forwards, one center — and a BPM-based season engine projects whether that lineup could run the table. The engine cares about roster fit, so the question stops being "can you name five legends" and becomes "can you build a team."
The engine compresses the value of superstars somewhat, for the same reason you couldn't put together a team of 8 players who score 20 points a game and score 160 points a game. cough Matt Ishbia and Vivek. It also punishes weak defense more than you might expect because on a team of legendary playmakers the elite offense doesn't compensate for being an easy target on the other end, compared to a normal team.
What kind of roster actually has a chance?
One that does everything better than everyone. It's not enough just to have a stack of pantheon players. Your team has to:
- be able to defend its butt off to stabilize the game-to-game variance.
- have spacing. Which means at least three players who are (or would have been, if it existed) respected from the 3-point line. You might get away with two shooters, but it almost never works.
- Have a super-role player. In a team of legends, there will be one who needs to contribute without touches. Just ask Chris Bosh. Actually, ask Mario Chalmers, because he was the #5 on the Heatles.
- Fills all the box score categories. When you're mad that 24-25 Kevin Durant looks like the Mario Chalmers of the team, know it's because he had below average stats in everything except efficient scoring. The engine is built on BPM, so it's gotta see all the counting stats filled, sorry.
So - not just trying to recreate a Best of Team USA. A team that has every weakness covered. To the extent BPM can detect that.