I came across some unexpected behaviour today when using the Longest function when trying to do some pattern matching.

StringCases[#, a___ ~~ b_ ~~ Longest[c___] ~~ b_ ~~ d___ -> {a, b, c, d}] &@"abcbba" 

The intent is to find the two identical characters which are the furthest apart in the string. I would expect this to return {a,b,cb,a}, since that would be the longest possible distance between two identical characters. However, it instead returns {abc,b, ,a}. I don't understand why it's not finding the longer possibility, especially when default behaviour when matching patterns seems to be that earlier patterns try to match the shortest possible sequences. Using c__ instead of c___ makes the behaviour more like the expected behaviour for this case, but I do want the pattern to work when the only sets of identical characters are adjacent to each other.

What am I missing?