When Steph Curry dunked against Philadelphia on March 1, 2025, it was his first in an NBA game in six years. Afterward he joked that it might be his last. That is a fun stat, but it is a poor way to judge his best jump. In warmups he can throw himself a lob, run into it and jump without a defender in the way. Those clips put his best-looking vertical at about 34 inches.
Why the warmup lob is the useful clip
One left-handed self-lob gives us the clearest look in the video. Steph gets a full approach, catches the ball in rhythm and finishes without having to change course around a defender. His head appears roughly seven inches above the bottom of the net. That is the jump I use, rather than counting how many times he has dunked in NBA games.
He is listed at six-foot-two by the NBA. The video allows about an inch for shoes, so the working height is six-foot-three, or 75 inches. If the rim is 120 inches high and the net hangs about 18 inches below it, the bottom of the net is around 102 inches. His head would need to rise about 27 inches to reach that point. Add the roughly seven inches it appears to clear, and you get about 34 inches. That is the arithmetic behind the estimate.
How much can we trust the net?
Less than the rim. A net swings, bunches up and can hang differently from court to court. The camera angle also makes a few inches hard to call. So 34 is not a laboratory result or a claim that every Curry jump peaks at exactly that height. It is my read of the best warmup footage I could find. If a clearer clip shows his head higher against a fixed marker, I would revisit it.
It also matters that he is carrying and finishing the ball. Jumping just to touch a marker and jumping to dunk are related, but they are not the same test. This is why the games versus warmups breakdown asks where elite jumpers actually get their best attempts, instead of assuming the loudest game dunk is always the highest jump.
What about the reported combine jump?
The source video cites a 35.5-inch maximum from Curry's 2009 draft combine. That is about an inch and a half above the warmup estimate. It is a useful comparison, but it is not a contradiction: the combine tests a running jump to a marker, while the clip here asks him to catch and finish a self-lob. They happened years apart, and my film number also depends on where the net hangs. The closer those two numbers look, the more tempting it is to pretend they verify each other. They do not.
For the other players in this series, use the NBA vertical estimates table. It keeps each player's number beside the source video and explains when a combine measurement and a footage estimate are being compared.