Alright here are the actual graphs, captured from the scope with a camera.

As before the discharge is a 0.5-ohm load.
My intention was to measure the voltage ~2.5 seconds after first hooked up but I didn't always end up doing this, so I had to estimate some values.

The cell is rated at 800mAH, so for every 8mAH of overcharge, that is 1% of the capacity.

1. 0mAH overcharge / 4.2V -> 3.447V
2. 7mAH overcharge / 4.22V -> ~3.49V
3. 0mAH overcharge / 4.2V -> 3.425V
4. 16mAH overcharge / 4.240V -> 3.491V
5. 0mAH overcharge / 4.2V -> 3.472V
6. 0mAH overcharge / 4.2V -> 3.487V
7. 0mAH overcharge / 4.2V -> ~3.52V
8. 0mAH overcharge / 4.2V -> 3.487V
9. 30mAH overcharge / 4.249V -> 3.550V
10. 0mAH overcharge / 4.2V -> 3.550V
11. 0mAH overcharge / 4.2V -> 3.552V

As you can see there is more than one thing going on in this data, while it is clear:
1) the overcharges do give better voltage under load,
there are probably some other things going on, like:
2) some break-in effect (demonstrated with the last experiment) and
3) something similar to 'zapping' where overcharges can effect not only that run but subsequent runs.

For instance observing the difference between 8 and 9, there is a 63mV difference but this carries over into 10, 11 even though no actual overcharge was done for these steps.

The series of the pics below is left-to-right, then top-to-bottom, ie.
01 02 03 04
05 06 07 08
09 10 11