A gamma match is a 'one-sided' device, meaning it will be found on one side of the element's center or the other. It can't be placed 'behind' the element. If it is, it's no longer a gamma match. You might be confusing a gamma with a delta or beta maybe.
In any case, placing that match in front of, or behind the element isn't going to change the gain or 'sharpness' of the resulting radiation pattern, or it's F/B ratio.
A gamma match 'skews' the radiation pattern of a beam? Yes, VERY slightly. And so do a 'ton' of other things near that directional antenna, that's normal. The 'perfectly' formed, 'smooth', pattern representations you see are the results of statistical 'smoothing' done to make the resulting pattern uniform. All the antenna modeling programs do that. And YOU do that when graphic the results of taking measurements and plotting them on paper. You sort of 'average' out the plotted points, make it a 'smooth' curve.
And then when you come down to actually putting a directional antenna into use, does your rotor really point things that accurately? There's no 'play' in the whole structure that the wind doesn't 'play' with? Those "one degree" marks on that rotor indicator's dial are for filling spaces, not because the rotor is that accurate.
(They do make handy reference marks though.)
[Reminds me a lot of the 'jiggle' correction thingy in the newer digital cameras, but you'd better have some extra memory for that thing to work.]
Why don't you find those matching devices on top of, or below the driven element? Mostly because it increases the 'sail' area of the antenna, which isn't the best idea in the world.
Think about it, then tell me how I'm wrong.