About 2 years ago my 144 would run fine, and sometimes all of a sudden just die. Had to leave it for 30 seconds, then it would start up fine and drive for another day or so. The FI models of the 80's had that problem with the fuel pump relay, but what could possibly do this on a B20B? Turned out that the distributor cap center electrode (that pushes onto the rotor) was worn down (or broken long before I discovered it) and in wet weather the spark would track along the inside of the cap rather than the gap that had formed between the electrode and the rotor. Letting it "heat soak" dried out the cap, and the car ran fine again for the next day or so. Just replaced the cap, ran fine after that.

An inductor (coil) tries to keep a constant current flowing by manipulating the voltage across it, that's why when the contact breaker breaks the current you get the voltage rising to something like 30V if I remember correctly. You also have a multiplication due to the primary and secondary turns ratio. So on the secondary where the only real remaining path for current flow is, you get about 1000x higher voltage at 1000x less current (law of conservation of energy). But all the energy is stored in the inductor, not in the capacitor (or condencer).

A capacitor (condencer) tries to keep the voltage constant by manipulating the current through it. When the contact breaker opens, for a brief moment the current will flow through the capacitor instead of through the air gap of the contact breaker. So the contact breaker lasts longer. It's only a brief moment, due to losses. The capacitor also helps to dampen the current/voltage waveform, because the faster the current/voltage rise the higher frequency noise is generated and the more interference you get with you car radio, etc. But that's getting a bit too technical for this forum.

Capacitors have numerous failure modes. The chemicals inside can dry out, the end caps (connections) can burn off, contaminants can cause high resistance, just to name a few. Some failure modes can cause a parallel resistance, so effectively the breaker does not completely interrupt current flow through the coil. This will produce weak spark, which you might not notice when looking at the plug at atmospheric pressure.

Hope this helps, I'm an electronic engineer who graduated 9 years ago and did electronics as a hobby for over 18 years. Can't help much with mechanical problems, but sure can with electronic problems.