Every October, the cottage industry of NBA prognostication churns out confident forecasts: this team will contend, that one will tank, and some promising young core will finally break through. By June, most of these predictions have curdled into embarrassments. The gap between preseason certainty and postseason reality is not a bug of basketball analysis—it is a feature of a league designed to reward the unpredictable.
The 2025-26 season offered a particularly brutal reckoning. Teams that were supposed to rebuild instead made playoff runs. Franchises anointed as contenders stumbled into the lottery. And the consensus picks for individual awards missed badly enough to suggest that perhaps the entire exercise of preseason prediction is less science than astrology with better graphics.
The contenders who weren't
The most instructive failures came at the top. Several teams entered the season with championship expectations built on offseason acquisitions and healthy rosters, only to discover that basketball chemistry cannot be purchased at the trade deadline. Injuries played their usual role, but the deeper issue was a systematic overvaluation of on-paper talent and an undervaluation of continuity and coaching stability.
The inverse was equally revealing. Teams written off as transitional projects—too young, too thin, too dependent on one player—instead developed faster than anyone anticipated. The lesson is familiar but apparently unlearnable: the NBA's salary cap and draft structure are specifically engineered to compress the talent gap between franchises. Predicting a 20-win difference between teams is almost always overconfidence masquerading as analysis.
Why we keep getting it wrong
The structural problem with NBA predictions is that they must assume health, assume chemistry, and assume that last season's performance is a reliable guide to this season's outcomes. None of these assumptions hold reliably. A single torn ACL can transform a contender into a lottery team. A coaching change can unlock a roster that looked hopeless. A 22-year-old can make a developmental leap that no model anticipated.
The analytics revolution has not solved this problem—it has merely made the predictions more precise in their wrongness. Advanced metrics can tell you what a team should do given certain inputs. They cannot tell you whether those inputs will materialize. The result is a kind of false precision: we know exactly how good a team would be if everything went according to plan, and everything almost never goes according to plan.
Our take
The annual ritual of grading preseason predictions serves a useful purpose, but not the one intended. It is not really about accountability for analysts—nobody loses their job for picking the wrong playoff teams. It is about reminding us that professional basketball remains gloriously resistant to forecasting. The NBA's competitive balance mechanisms work. The human element—injuries, egos, unexpected growth—defies modeling. Perhaps the most honest preseason prediction would simply acknowledge that we are all guessing, some of us just with fancier spreadsheets.
