Thank you Mr. Hicks. I've noticed that after a long long charge past the time when it first reads 100% resulted in very little reduction in the first mile of riding, whereas use after the charger light first changes is now resulting in up to 14% reduction immediately on the first hill less than 1/8 of a mile, which bounces back up to maybe 8% only loss after a short rest.
In other words, the battery shows quicker loss when full than later on in the same ride.
I'm doing the extra long "after 100% charge" right now again to see if I get that result of not losing so much immediately.
Previously I had measured with a voltmeter at 100% and found it wasn't fully charged. Leaving it on for an hour or so helped, but only long charging after 100% got the battery to full voltage.
I suspect some weaker cells are in there.