The condenser not only changes voltage at the plug, it also changes the shape of the voltage curve, as the crank (and with it, the flywheel) rotates...kinda like point gap affecting dwell angle.
Electrical energy follows the path of least resistance. Somewhat counter-intuitively, voltage at the plug should be higher with a healthy condenser in the circuit. However, the current would also be released more rapidly, i.e. "all at once". On an oscilloscope, you'd see a sharper curve, ideally approaching a square waveform. With an inadequate condenser, voltage would build more gradually (visually, a smoother curve), with the wire & cap providing some capacitance...until it reached a level high enough to jump the air gap. Think of the flywheel magnet, it's long enough to span several degrees of crank rotation. Electrical energy builds until it's rotated past the primary coil armature. In physics, I once learned how to estimate arc voltage by the length of the arc (a.k.a. spark gap); there is a linear relationship between voltage and maximum gap size. That's why I said "somewhat counter-intuitively"...the combination of a dying condenser, with a secondary gap at the end of the HT lead, especially if there's oxidation at the ends of both can result in this kind of arcing, at a voltage that's lower than what the plug normally sees. The car makers taught a lot of folks a hard lesson, back in the 1970s, when plug gaps were opened up from the typical .035" to .060"...failed plug wires became almost epidemic - and were easily diagnosed, at night.