What the lookup computes.
Nearest-name matching converts your input and every named color into Lab space, then computes Delta E, a measure of perceived difference, and ranks the results. #E92C4B gives crimson, whose value #DC143C sits about 4.6 Delta E away, with indianred next. #5A8FCB gives cornflowerblue at about 4.5, with dodgerblue and steelblue close behind. #FF6347 gives tomato at a distance of exactly zero, because that's tomato's defined value.
The finder shows the closest name plus four alternatives, each with its Delta E, so you can see when two names are nearly tied. The distance number matters as much as the name. Under 5, the name is a fair description; past 10 or so, treat it as a general direction rather than a label you'd put in a brand doc.
Why not just find the closest RGB values? Because channel arithmetic doesn't match human vision. Our eyes are far more sensitive to some hue shifts than others, so a naive nearest-neighbor search in RGB space will sometimes crown a name that looks obviously wrong next to your swatch while a better candidate sits two entries away. Lab space and Delta E were designed to fix that mismatch.