That's because you're using ".*". That covers any characters and matches as much as possible, so the 4th group gets rolled into one of the previous ones. In your example, you also have dashes in the middle of what you're expecting to be the second group, so that makes it more complicated because it now has even more possibilities for matching a group. Is the 3rd group always a date?

If yes, you can try something like this: ^[^-]+ - (?:[^-]+-?[^-]+) - [0-9/]+ - (.+)$

(^ at the beginning and $ at the end tell pattern matching to look at the whole input value to match the pattern. The first group is "anything but a dash", the second one is "at most one dash in the middle of 2 words", the third one is "digits and slashes" to account for variations on the date format, but you could refine further if you know the format is always the same. Then the last is "whatever is left all the way to the end")

I'm a bit late to this party but I'd like to join in on the fun ~

You can also resolve this problem by "starting from the back". This pattern works for me, and is a bit more straight-forward (in my opinion):

This will work even if your input does not contain the date, or if the date is formatted differently (such as with hyphens). The only constraint here is that it uses the last two hyphens are delimiters, starting from the end. That is, it assumes that the text you want is indicated by the finding the second-to-last hyphen and starting there (and going until the end). It also means that if the second-to-last word is "Spread-sheet", or if the document name is "New-Document" or something, that this won't work. But you have to define *some* kind of rule for the patterns to work anyways =). Oh, it also assumes that this is the whole input. That is, you're not finding this in a bunch of text, but rather you are given this input and you are running pattern matching on this whole Test Value as though it were the start-and-end of the input. You could always do a pattern match to capture this particular text and then perform these steps.

Basically what this pattern says is "Get me anything except a dash, then a dash, then anything except a dash, and then make absolute sure that we are at the end of the input right after that". The last [^-]* is important because being specific about what you want to match, it will try to get as big of a match as possible. Regexes are both hungry and greedy. Specifically, the last [^-]* ensures that you are actually starting that section of the match between the last hyphen and the end of the input. If you instead used .* for that part of the input, then it would "prefer" to start that part of the match at the character right after the first hyphen which would result in a valid match and a much bigger one at the same time.

You might notice this also captures an extra space at the start. You can get rid of that by instead using this:

[^\s-]*\s-[^-]*$

That is, "everything except a space or a hyphen, then a space, then a hyphen, then everything except a hyphen, then the end of the input".

If the spaces are "optional", then you can do 

\s*([^-]*-[^-]*)$

That is, there are zero or more spaces before the first part of the match, and then we only capture the part that doesn't include the space.

If you prefer to "visualize" this as capturing part of the whole match, then you could instead look at it this way:

^.*-\s([^-]*-[^-]*)$

That is, "Start of the input, then a bunch of stuff, BUT make sure you also match a hyphen, a space, everything until a hyphen, then a hyphen, then everything until a hyphen, and then the end of the input. Oh, and make sure you capture everything starting from the first hyphen+space which we mentioned, all the way to the end of the input". 

Different way to visualize the same solution, but it matches your entire input and only captures a specific part.

Regexes are fun ~