They are still claiming only 20% of a batteries capacity is used in a typical product.
The reason they use is pulse currents causing products to drop out at 20% of used battery capacity
Which is correct with a really badly designed product (and I never met one)
Correct. Theoretically possible, but finding a real product that does that in the wild would be very hard.
Most battery powered products are:
a) Designed correctly to get at least decent battery life
and
b) Have regulators, bypassing caps, and low voltage dropout circuits etc that handle any required pulse currents just fine. Which is why on all but niche examples, using a power supply to measure the battery cutoff voltage is and industry standard way to do test for battery cutoff voltage. Indeed, the one method that actually lets you easily test what the actual cutoff voltage of the product circuitry is.
Of course, that what I had in mind by saying that I've never met one, even the old and honorable Gameboy just makes it's power LED to flick to tell that the battery are low, but not shutdown the product.
As I says way ago, a batterizer with a configurable (manually or automatically) output voltage to
- tell that the battery is going down
- make sure that the differential voltage does not get absurdly high, like 0.7V when boosting from 0.8V to 1.5V
That may works a bit better, will not make it's assertions of course, but at least would prevent prevent device from shutting without warning.