The main plug is between the valves, right in the centre. If that stopped working it would be obvious.
If the small spark plug at the side stopped working it would make no discernible difference. That is the theory.
I do agree that if you had the choice of which plug failed, then you would choose the secondary plug, because a centrally located spark is better than an offset spark. However, lots of engines have substantially off centre spark plugs. The single offset plug still fires the cylinder quite effectively, if not quite as efficiently (as a centrally located plug).
My understanding is that having twin plugs creates a wider flame front, and tends to mean a faster and more complete combustion, though I suspect the effect is quite minor (or twin plugs would be a lot more common). With twin plugs, if the primary plug fails I can't see why the secondary plug wouldn't still fire the cylinder effectively enough, though the possibly less complete combustion may show up in the exhaust emissions (emissions possibly being the real reason why twinned plugs are fitted). Power would be affected, but not to a major degree (I suspect...).
CF1, CF2 and early CF3 engines had the HT wires pair cylinders 1&4 and 2&3. On these engines, a coil would either power both main plugs or both emission plugs. It would be obvious or not at all discernible. If it was obvious, swap the coil from paired cylinder and drive home normally.
"Emission plugs"? As I said, I suspect that attempting to meet emissions requirements is the main reason for the use of the twinned plugs, because of the ostensible wider flame front / faster and more complete combustion. But plenty of other engines make good power with a single plug, and still manage to meet emissions requirements...
In all instances of 16v Twinspark, wasted spark strategy is used. I suspect this made more difference to cleaning emissions, or at least giving catalytic converters an easier time than extra spark plugs ever did.
IMO, 'wasted spark' ignition (the 'wasted' spark occuring BTDC on the exhaust stroke) is at least partially a cost saving exercise. With wasted spark it is possible to trigger the ignition very accurately once each crank rotation from the crank position sensor, meaning that there is no need to fit an additional sensor to detect camshaft position (keeping in mind that the crank and cam shafts don't rotate at the same speed, so with single spark the ECU must 'know' which stroke the crank is on in order to fire the plug on the correct piston upstroke, which can only be done with a camshaft position sensor if the spark only fires on every second crank rotation).
And / or, not relying on a camshaft position sensor means that the ignition timing can be more stable and accurate. It eliminates timing 'scatter' caused by the timing chain or belt allowing the camshaft to momentarily slightly over and / or under rotate relative to the crankshaft (which does happen, due to inertias in the camshaft and shaft drive, and belt drives suffer less bcecause there is less mass and inertia in a belt than in a chain). This causes small more or less erratic fluctuations in the valve timing, and if the ignition timing is reliant on a camshaft position sensor, in the ignition timing as well ('wasted spark' eliminates this ignition timing 'scatter'). However, these affects are minor, lots of very good engines trigger the spark timing from camshaft position sensor(s) with adequate accuracy.
I've heard it postulated that the 'wasted' spark helps to ignite unburnt hydrocarbons in the exhaust gasses as they leave the cylinder, thus improving emissions. I have serious doubts as to the veracity of this theory. Near the end of the exhaust stroke any gasses will be very depleted of either oxygen or fuel, or both. If the exhaust gasses still contained significant unburnt fuel, then the gasses would be extremely oxygen depleted, the oxygen having been consumed a lot earlier in the combustion process. If the gasses were oxygen rich, then the gasses would contain very little unburnt fuel, the fuel having been consumed a lot earlier. Either way the gasses would be very difficult to ignite, the 'wasted spark' is highly unlikely to achieve it (IMO...).
Perhaps having 2 effective pressure wave fronts asymmetric in each cylinder was useful for obtaining efficiency and torque without the need for really hot camshafts. That is debatable as is benefits and deletion of twin spark technology when the JTS replaced it.
IMO, twinned plugs are a dead end. There are theoretical benefits, which may be partially realised in practice, but even AR has abandoned it. I suspect that cost vs benefit isn't good enough, and at least as good results can be achieved in other ways with single plugs. This isn't to say that it is not an issue if one plug fails in a twin plug design, the engine has been designed around having the two plugs operating...