The first two components he removed (the filter Y-capacitor and the opto-isolator) are supposed to be open-circuit to direct current. And since they were not at

The first two components he removed (the filter Y-capacitor and the opto-isolator) are supposed to be open-circuit to direct current. And since they were not at fault (ie, they were not shorted), removing them from the circuit didn't change anything. The fault probably stopped after he removed the transformer because in the process of desoldering and yanking it from the board, he caused enough strain/vibration that whichever wire was shorting the primary and secondary windings moved and was no longer causing a short. That sort of intermittent fault is common in hardware, and quite annoying to debug.