FWIW, the magneto doesn't give a fig about where the keyway is in the flywheel, as it will fire whenever the magnets have done their things. Generate enough juice to create a field in the coil, and then drop that juice(voltage) to ground and collapse the field. That is all they do.
So, if you pulled both spark plugs to allow for faster cranking, and to supply the high voltage terminals with something to work on(lay the plugs on metal while cranking), you should see sparkage at both plugs are just about the same time. As noted above, the same electrical voltage is applied, and the current flows roundabout to finally get to ground.
I bet a nickel that you have grounded the coil somehow, or mounted it somehow, that is causing the spark to go right to the block, and bypass that rigorous route through coil-plug-block-plug-coil.(more or less).
To put it in different words, this particular coil setup grounds differently than a single plug coil, even though all it does is create one jolt of juice that flows across two plugs, and the block before getting back to the coil.
I am trying to some mental gymnastics to figure out if the coil itself needs to be grounded, or just one side of the high-volt secondary.(along with the 'generate' winding that creates the low-volt primary).
Either way, with two plugs in series, the ground is totally different from that of a single cylinder application.
I started yapping about the magneto not caring about keyway. That was the start point of my thought about the flywheels, one working, the other not. The layout would have to be the same as far as the permanent magnets around the periphery of the flywheel to be able to generate base voltage and signal voltage. The NSNSNS pattern of magnetic fields around the flywheels on the outside would have to be the same to operate the magnet the same. I'd take a hacksaw blade and move it around both the working and non-working to determine if I get the same pattern of magnetism. Perhaps draw a picture/schematic relative to the keyway on both, show where magnetism starts/ends and see if they are similar. To my thinking, for the magneto to work, it needs a particular setup of fields, or it won't work. While poking around with a hacksaw blade, you can mentally compare the strength of the fields on each flywheel.(if they are totally different in strength, maybe one flywheel got dropped and lost magnetism) ... I understand that B&S used to offer a re-magnetizing service for the occasion when the electronic ignition was taking the place of the 'points' ignition. Along those lines, if a field is weak, perhaps it can be strengthened by B&S or by cooking up an electromagnet that induces stronger permanent magnetic field in your weak flywheel. I *think* an electromagnet with an iron/steel core to 'concentrate' the field could be placed on/near the flywheel magnetic spot, and have the winding energized such that it is in accord with the flywheel magnets, and make them stronger. Dunno, never tried it. I do know you can magnetize screwdrivers by winding around them and applying DC(quick/short bursts or it gets HOT). I think the same can be done...
tom
