Do you not understand how enormous half a second is in this race?
It's also more than that. Just to be clear since I can't remember if you're quoting a wind-adjusted time:
- Usain's WR is 9.58 (9.62 wind-adjusted)
- Jesse Owens PR was 10.2 (10.3 wind-adjusted)
The reason this is listed as 10.3 on his personal Wiki because it was tailwind-assisted 10.2. Usain had an 0.9 m/s tailwind, too, which adjusts his time to 9.62. But we're not done yet. Even capable hand-timers have been shown to subtract an average of ~0.15s to the racer's time due reaction time (the normal range runs from 0.1s-0.3s). Now, some of this would be canceled because you are hand-timing both the start and the finish, but the anticipation effect means that hand-timing still favors the racer.
This was a well-known and well-studied phenomenon in club swimming prior to the advent of the Colorado timing system (which automated the start time). The reason is that, before the race begins, the timer must wait, and reacts to a light/sound for the race to start. Meanwhile, they can visually track the racer into the finish, which means their reaction to the finishing time is greatly diminished, and more accurate. So the race gains time at the start, but loses less time at the finish.
Realistically, you are comparing a 9.62 to a 10.4: about an ~0.8s difference indicating nearly a 1 m/s difference in average speed.
I do not believe Epstein's 9.77s to be a realistic projection.