The 1000km railcar between the two clocks is unnecessary. All we need is are two ends, moving at the same speed. But, a 1000km car would work too, if you can get the funding. To be clear, in my setup both the \(S\) and the \(S'\) clocks are the same distance apart in \(S\). This means that in \(S'\), the distance between the \(S'\) clocks is \(L\gamma\). As Neddy said, that's not a flaw, it's the expected result. In \(S'\), the clocks obviously don't line up at the same time - the events are not simultaneous. That's the whole point of the experiment. The \(S'\) clocks measure the two events to be non-simultaneous. The \(S\) clocks measure the events to be simultaneous. The rail clock at B is measured by the \(S'\) clocks as being way past to the right of the corresponding ground clock at the time that the A clocks line up. The rail clock at A is measured by the \(S'\) clocks as being still way off to the leftt of the corresponding ground clock at the time that the B clocks line up. They are lined up at the same time in \(S\). They are lined up at different times in \(S'\).