Spark is generated when a large transistor pulls the coil low to charge it, followed by the transistor opening, the coil field collapses and coil voltage increases from 12V to about 400V. The secondary coil windings step it up by about 100x, so it could peak to 40,000V. A circuit needs to be complete for current to flow, normally the + coil lead back to battery, then the coil primary, secondary (coil is auto transformer), plug gap and ground. There is also a capacitive path in transistor to ground, but less than point capacitor in point ignitions. Opening the coil +, will result in spark, if coil has current, and shock to uninsulated fingers. The ignition circuit is a series circuit at time of ignition. If one understands this, it becomes apparent why capacitor is placed from coil + to ground to reduce EMI.
The iffy part is if the coil is charged. There is a control circuit in the HEI, that closes the transistor to charge the coil based on the time of the last spark event to charge the coil near full just in time for next spark. That is one reason the ballast is not required. It takes a few milli seconds to charge the coil. There is also a hardware limit, it is used at lower RPM. Since RPM varies the time between sparks, the predictive control is busy, and also less than perfect, it adds additional time for varying RPM at low engine speeds, else variations would cause misfires because the time between sparks is very long compared to charge interval. Now days a micro controller makes better decisions based on better timing sensors, and coil current measurements to accurately control ignition.
Some may read this and think I am a know it all, I am not. I learn from experience, just enough to do things I enjoy doing, the best I can.
