Appreciate the feedback. I want to make this easy for scenario writers, but I need help in this area. May I ask what you mean by "all these yard tracks?" We don't have signals "inside" yard areas, like Ronceverte. There is a signal to protect the gate leading onto the mainline from sidings/yards; access to a siding to get OFF of the mainline is protected in blocks. If you exit the block completely onto a siding, you free that block for other traffic. In real life, you would then contact the dispatcher for orders to get back onto the mainline.
If I may offer a real-life scene.... A WB freight will stop at Ronceverte WB#1 signal; disconnect 10 cars leaving his train on the mainline; proceeds through the signal into the block containing the gate for Ronc North Yard and backs those cars into a siding there. His train is still holding the block east of the signal (right?) but once inside the yard, the block with the gate is now available. Only with orders and a green dwarf signal can he return to the mainline to return to his train one block east. While in the yard, a non-interfering EB passes on #2 has no effect. However, an interfering EB that comes up #1 (uncommon) and crosses over to #2 would block you in the yard until it past. Theoretically, to make this scene work, all you should need are several "at-speed" destination instructions and a couple/few wayside markers. Does it work? I don't know; haven't done it yet.
That is a real-life operation; translating that to Railworks is what I am still learning; any lessons are welcome! I need to take some time to find and read everything out there on scenario editing. About my design: prototypical of the C&O (and Chessie and still CSX on this line) each siding is not protected by it's own signal; each siding or gate or groups of them is protected within a Block. That block's signals may be as much as 2 miles away. I marked every track at the entrance to that track with a siding or destination marker; scenario writers should add waypoint markers liberally to control the paths of their trains and indicate drop locations. This is what was recommended by other scenario writers in the community. I think the idea is think about routing your trains, not by signaled junctions, but by signaled blocks.
I'm guessing that while this theory is all well and good, RW AI traffic has it's own mind?? I've heard it likes to lock junctions that it is not even using and junk like that. I have accepted the possibility that RW may not allow this pure block type of signal operation at all. RW tends to like the developer to control every single teeny tiny little itty bitty piece of it individually (usually with a whole file full of properties!!); I didn't build the COA that way. Please post again with your results or more detail.... this is what the beta is for (so far I've not gotten a lot of truly usable techical feedback). I'll do what I can to help make it work.