They claimed (and still do) 7500 pages, but that was probably machine tested along the lines of:
1) click
2) wait five second
3) repeat 'til the battery's flat
4) report number of clicks.
And who can tell what sort of effect the actual testing approach itself might have had on the battery drain?
For instance, we know that it does use some small amount of power between page changes for monitoring the buttons. I'm guessing the pause between pages which an actual
person requires to actually
read the page may make up a lot of the difference. If they waited my above assumed 5 seconds between flips in their testing, but it takes me 30 seconds to read the page then right there I've increased the 'waiting' power consumption by a factor of 6
per page.
Then there're the times I set the Reader aside 'just for a minute' while I'm doing something else 'real quick' and come back to it half an hour later.
I don't know how we'd even figure that into the equation without knowing the exact power consumption during the 'wait' state (and good luck getting
that info).
