Regarding the theory behind the Motor 1A “L” switch dog position and the resultant behavior - you’re right, this switch dog position delays the 1A pulses (about 7%). As a result, the fifth 1A pulse begins to overlap the Motor 2B pulse (which is used to release G in the 500 point case and E in the bonus scoring case). Additionally, in the bonus scoring, the “L” switch dog causes the 1A pulses to slide relative to the 4 masking pulses. All bets are off with this changed timing, it comes down to how much charge flows through a switch that is opening prematurely and whether it’s enough charge to get the relay down the line to move.
The reason adjusting the 1A switch gap was affecting the behavior, it that you also are changing 1A timing based on the gap width, but to a much lesser extent than the switch dog does.
…and, yes, many thanks to baldtwit for the suggestion to check the dog - we could’ve ended up putzing around for 50 or more posts chasing ghosts.