Hey, battery technician here. You see, what happened here is that the hungry little caterpillar has entered the digital age and I don't know what the hell is going on here.

It probably cannot measure correctly because this is an unexpected usage. Correct place is the bottom of it obviously. The screenshot just coincided with the upper side of the spike.

My guess (I don't know if this is actually right) is that the battery is showing a lower charge while under load. So when it wakes up to do it's little periodic checks/activity (could be trying to fetch updates, could be scans, could be cleanup, idk) it draws additional power and the old battery can only support that load for so long.

Explanation based on theory as I understand it: Using an analogy, batteries aren't a cup full of electricity that just gets sipped or dumped, they're closer to a water balloon with a little valve. The system may only use a little water at a time, but the "battery full" percent is the measure of how much pressure is in the balloon to push the water out. So when the valve is mostly closed (phone is idling) the pressure is pretty normal, slowly dropping. But as batteries age when you open the valve further you start to realize the battery can't actually push out the pressure you need (the balloon has stretched out). The pressure drops, to the levels it would be at if the balloon was brand new, but almost drained. The valve closes more as the device demands less power, and it seems like the expected pressure returns to normal, because the balloon can handle that lower pressure demand.

In short "New battery almost dead under load = old battery mostly charged but under load" but "new battery charged with no load = old battery charged with no load" which is why you see the slingshotting on the battery graph.