SportsLine model identifies value in New York Knicks for 2026 NBA Finals
The projection system assigns a 45.2% probability to a New York victory, contingent on Mitchell Robinson’s availability, and highlights a specific five-game series outcome as the only play offering statistical value.

The SportsLine Projection Model has selected the New York Knicks to defeat the San Antonio Spurs in the 2026 NBA Finals, identifying significant value in the underdogs despite the Spurs being listed as favourites. The model, which simulates every NBA game 10,000 times, assigns a 45.2% probability to a New York victory, a figure that exceeds the 35.8% implied odds set by FanDuel.
The series is scheduled to tip off on Wednesday at the Frost Bank Center. San Antonio enters the best-of-seven contest as -198 favourites to lift the Larry O'Brien Trophy, while the Knicks are priced at +166. Victor Wembanyama is the -180 favourite to win Finals MVP, compared to Jalen Brunson at +210.
The model’s confidence in New York is underpinned by their performance throughout the postseason. The Knicks have won 11 consecutive playoff games, marking one of only five instances in league history of an 11-game winning streak in a single postseason. Their +262 point differential over this period is the largest such margin recorded in NBA history.
A critical variable for the model is the availability of centre Mitchell Robinson. The 45.2% probability is contingent on Robinson playing at least 70% healthy, with an expectation of at least 15 minutes per game. He is expected to start Game 1 with a cast on his right hand to protect a broken pinkie finger. If Robinson is unable to play, the model reduces New York’s chances to 36.1%, which aligns closely with the market’s implied odds.
While simulations indicate that the most likely outcome is a Spurs victory in seven games (18.3%), followed by the Knicks in six (16.7%), the model identifies neither series length as offering value. The only specific prop bet flagged as valuable is the Knicks winning the series in five games at +1200, which the model assigns a 10.5% probability against an implied 7.7%.


