A CPU can run only one job at a time, so an operating system must decide who goes next. First-come, first-served is fair to arrivals but suffers the convoy effect: one long job at the head makes every short job behind it wait its whole length.
Shortest-Job-First is provably optimal for the average waiting time — clearing quick jobs first lets many finish sooner. But it can starve a long job indefinitely if short ones keep arriving, and it needs to know burst lengths in advance.
Round-Robin slices time into a quantum and rotates, so no job waits long for its first turn — that is fairness. The price is context-switch overhead: a tiny quantum means constant switching and lower throughput. There is no single best policy; every scheduler trades throughput against fairness.
The simulation stopped unexpectedly — the lesson continues without it. You can move on; nothing you did was wrong.