Expanding slightly on oscarfonseca's excellent explanation, For loops, by default, use an indexing output tunnel (if you look carefully, you'll see the tunnel is a square surrounding a pair of square brackets, [ ], suggesting an Array). Your exponential is a series that you want to sum. Notice that the wire coming out of the tunnel is slightly thicker, the sign of a 1D array. When you pass it through the Sum function, you add up all those elements and get a number, now carried in a thin "scalar" wire.
What about the other calculation, 1/(1-x)? This never varies, and is (probably) the (single) number that you want to compare to your summed series. By default, you are also bringing this out through an indexing tunnel, so you are creating an array, when you really want a single number. The only change you need to make to your code is to change the indexing tunnel from the 1/x function to a "last value" (the filled-square) tunnel. Right-click the tunnel, choose Tunnel Mode, Last Value, and the errors will disappear. Now run your code. With N = 5 and X = 0.5, I got a deviation of -3%, which dropped to -0.1% for N=10.
Bob Schor
