Here are a few spacetime diagrams. If you haven't come across them, they're simply plots of the position of an object over time, the custom being that time goes up the page. So a vertical line represents an object that isn't moving, and a line slanted to the left indicates one moving to the left. You may have come across these (usually with time horizontally) as "displacement-time graphs" in high school physics. We take them a little more seriously in relativity, since spacetime is a thing - these are maps of spacetime, and the lines are the 4d objects inhabiting it. You see one 3d slice at a time.
So here's your scenario, with my extra "D" observer. A and B are marked in red and are stationary, C and D are marked in blue and moving to the right.
Note that, in this frame, C and D are closer together than A and B - the horizontal distance between the lines ("the space between them") is shorter.
We could mark on the diagram the start and end of the experiment - when C passes A and then B. Let's do that with fine red lines:
Now comes the interesting part. C and D don't use the same definition of simultaneity. The lines in spacetime they call "now" are not horizontal lines on this diagram - they're sloped. In fact they are the fine blue lines on this graph:
Note that these lines still go through the A/C and B/C meetings. But if you look at the lower fine line, you can see it passes through B much later than the fine red line did. So according to C, at the start of the experiment B's clock was ahead. And the part of B's line that is "during the experiment" according to C (the part between the fine blue lines) is much shorter than the part that is "during the experiment" according to A and B - that is, it starts ahead and ticks slower.
This is also why length contraction doesn't lead to a contradiction. Length is measured at one time, and the frames have different definitions of "at one time". A and B measure length horizontally; C and D measure it along the fine blue lines. So there are four different measures: the A-B and C-D distances along a horizontal line, and the A-B and C-D distances along a fine blue line (warning: "distance" here is a slight lie - you have to measure "interval", which is ##\sqrt{\Delta x^2-c^2\Delta t^2}## rather than the Euclidean distance you see on the graph, ##\sqrt{\Delta x^2+c^2\Delta t^2}## - this has the consequence that some lines that are longer on the graph are shorter in reality).
You asked about symmetry. That's why I've drawn D, even though I haven't used that letter yet. I can draw the equivalent graph using the frame where C and D are at rest. It looks like this:
This is a mirror of the first graph - now the blue lines are vertical and the red lines are sloped and closer together. The "same experiment" is A moving from C to D, and the same graphs with horizontal and slanted fine lines (just left-right reversed) can be drawn:
That's the symmetry you were missing. But you need to add D to get it, or in this frame there's nothing for A to move to and you can't do the same experiment as you did with the first frame. If you ignore D for a moment, the only thing you can measure is the time between A and B reaching C - and here you can see that the time between those events as measured by C is shorter than the time measured by A and B, by comparing the vertical distance between the crossings.
I hope that helps. You may need a more thorough intro to Minkowski diagrams if you haven't seen them before, but they really are the easiest way to visualise special relativity.