Boston Scored Its Way Into a False Spring—Then the Defense Burned It Down

Start with 108.4. That is the number of points per game the Celtics allowed in their final 10 regular-season games, up five full points from the 103.4 they gave up in the prior 20. Nobody wanted to talk about that number in late spring, because the offense was doing something genuinely spectacular—119.6 points per game in those same 10 games, a nine-point leap from the preceding stretch that made Boston look like a team hitting its stride at exactly the right moment. The problem is that both numbers are true simultaneously, and when you hold them next to each other, the "stride" looks a lot more like a team sprinting toward a cliff with its eyes closed.

The close-game record is where the illusion fully collapses. Boston finished 10-10 in games decided by five points or fewer—a 50% win rate in the situations that playoff basketball is almost entirely composed of. Their overall winning percentage was 68.3%. That 18-point gap between how they performed in general and how they performed when margins tightened is not a rounding error; it is a structural diagnosis. I wrote about this gap when the Brown trade dropped, and I wrote about it again when Tatum called the whole situation "weird" at the ESPYs. The number keeps forcing the same conclusion: this was not a 56-26 team in any meaningful playoff sense. It was a team that feasted on margin, then starved when margin disappeared.

Now inventory what the front office actually did about it. Mitchell Robinson arrives from New York on a three-year deal, bringing rim protection that was genuinely absent—Robinson is a legitimate deterrent at the basket, and his signing at least acknowledges that the interior was a liability. Neemias Queta, who shot 65.8% from the field last season and was quietly one of the more efficient bigs in the league, got locked in with a four-year extension. Those are real defensive additions, and they address a real problem. But Robinson came with a hand injury from the Knicks' playoff run, and his durability heading into 2026-27 is already a question the organization is managing rather than ignoring.

Here is the part that does not add up: the trade that defined this offseason was not about defense. Trading Jaylen Brown—27.9 points, 6.6 rebounds, 4.8 assists per game on 47.4% shooting, the engine of that late offensive surge—for Paul George and picks is a bet on offensive reconfiguration and future flexibility. Whatever Paul George brings, and he brings real two-way credibility, the transaction was framed around scoring construction, not around fixing the thing that actually ended Boston's season. The defense allowed 108.4 points per game when it mattered most. The response was to overhaul the offense.

My read is that the front office correctly identified the symptom—clutch-game fragility, close-game collapse—but misread the cause. The 50% close-game rate was not primarily a Brown problem; Brown was Boston's best player in high-leverage moments. It was a defensive identity problem, a team that could not tighten its scheme when opponents adjusted and margins shrank. Robinson and Queta help at the rim. But rim protection alone does not solve a late-game defensive breakdown that showed up systematically across the final stretch of a season. The depth chart at forward is also a concern that predates this offseason and survives it: the drop from Tatum to Sam Hauser—who averaged 9.2 points per game in 24.5 minutes—represents a production gap that leaves Boston dangerously thin if Tatum's current day-to-day status lingers into training camp.

The Celtics will score next season. Pritchard at 17.1 points per game, White at 16.1, George integrating into a system built around Tatum—the offense is not the worry. The worry is the same one it was in April, and May, and when I started tracking this defensive trend back in the regular season's final weeks. Boston's offense gave them a beautiful regular-season record and a first-round exit. The offseason moved the pieces around the scoring column. The defensive column, the one that read 108.4 when it should have read something closer to 103, has not been answered. That number is still waiting.