This is a little tricky to compare straight up. You can control a game, have an 8 point lead with 1 minute to go, give up a layup and a 3 and win by 3. So it looks like a close win but really it wasn't. Reverse could be true too.
But as a proxy it's still a reasonable measure, especially with a decent sample size.
Maybe an interesting stat would be W/L in games where it was 4 pts or less at the 2 minute mark. I'm probably overthinking this.
They do say that in baseball, 1 run games are basically attributable to the manager's skill.