I took a grid of each board to show here. It will be easier to understand.
In version one, you've overlapped the gap between pads with a trace. This is a unique approach to say the least; one I don't recommend.
Going from Gerber files to a board is a photographic process (it may be digital now, but the same concepts apply). The copper for the pads is made by using an aperture to expose the pad area. There is some tolerance in this process both in terms of pad size and of pad position relative to the drilled hole. The traces are made by a difference process. The trouble with this method is that the traces don't connect the pads as is the usual case. There's no
guarantee that the traces connect the pads if the tolerances are off. It probably will work ok, but why bastardize the technology?
For version 4, it took me a while to even figure out what I was looking at.
Turning off different layers made it more clear. Square pads and a big thick trace connecting them.
This will work fine for board production. But to be honest, from a use point of view, I don't think this is optimal. The thick trace isn't needed for anything you're likely to build on a proto-board, and the large pads and traces end up being kind of close together, increasing the chance of solder shorts between pads.
If you're trying to duplicate the proto-board you showed originally exactly, this won't quite do it. The area in between the pads will be covered with solder mask - it won't be a 1x3 area of blank copper. More on this shortly.