To prevent trains from getting stuck nose-to-nose at bi-directional signals, they have to be placed very carefully. There are basically only two situations where they should be used:* Side line with single track and '''only one train''' going back and forth crossing the main line (without . Without the crossing, there wouldn't have to be any signals on such side line. But to prevent above described self-block, there must be signals on '''each segment'''. Since the track is used in both directions by one train only, these must be bi-directional signals. If there is more than one train using such sideline, bridge must be used instead of the crossing otherwise the trains would get stuck nose-to-nose at the bi-directional signal eventually.

While it sounds complicated, it simply means that the signal will "look" down the line at the next signal (or signals if there is a switch and multiple tracks) and if he it sees all red, it will be red, if he it sees at least one green, it will turn green.<br>If It should be noted that incorrectly placed chain signal where block signal should be will not cause an accident, but it will make the trains wait longer, thus the rail network will be less efficient as a whole.

* ''At crossings'' to prevent train from blocking the crossing, if the train cannot continue after beyond it. Correctly placed signals are '''Chain before''' the crossing and '''Block after''' the crossing. If the track after the crossing is empty -> the block signal is green -> the chain signal before the crossing is green. If the block after the crossing is occupied -> the block signal is red -> the chain signal before the crossing is red as well so no next train will won't enter the crossing thus it will not block the traffic on the other track.