I'm pretty confident that Grant has access to stats that relay the winning percentage for each team, in real time. ESPN has one on their game tracker. Grant likely has an NBA data stream that he utilizes and predicts chance of winning. Hence when the Kings hit that shot, the chance of winning was likely 99.9%. And, generally, barring something crazy--the projections are accurate. Predicting outcomes is pretty much constrained by time, pace, and the currency (3 pts, 2 pts).
That said, in sports, something crazy is almost a given now--kickers missing PATs, Az cardinals lateraling the ball backwards and end up giving the Niners a TD (hence allowing the Niners to cover), or a few Dummy plays each night.
Side note: I looked casually, but haven't done an actual deep dive. I think a key stat to track is Buddy's TOs. More than 5 and the Kings are pretty much toast.