Not very good math here. S Engineering stated, E640 draws about 9 watts when on but idle, and about 4 watts when off. So 9W times 80% times 7 hours = 50 watt hours, and 4 X 7hours x 20% adds 5.6 watt hours. Also, the inverter itself probably adds another 20-30 watt hours over 7 hours when turned and not connected to anything.
So 80 WH or so of the potential 110 watt hours was used in seven hours without taking a single shot, leaving only about 30WH for actual shooting. If 400 shots took the indicator down to 1/4 that means 3/4 of the power fired 400 shots and the math says the battery would be depleted at 533 shots. Had you done he shoot over a 1 hour period you would likely have gotten around 500 shots at 1/16 power.
All of this assumes the numbers you and Engineering were accurate and that you didn't leave the inverter turned off the night before. Everything stated here was "approximately".
Also, everything stated here applies equally to all other battery packs (including Profoto, Innovatronix, etc, except that the other systems use SLA batteries that deliver only about 1/2 their capacity to the load, so you could expect to see about fewer pops per charge, relative to VML, at a given WS setting on the light(s).
Finally, your original posts refers to "Einsteins" implying your were using more than one Einstein. If you were using two Einsteins, the quiescent draw would be double that stated above and would indicate the quiescent draw would use the entire capacity of the VML in around 7 hours with no shots taken.
Please see
http://www.robgalbraith.com/bins/multi_ ... 1314-11226 for comparison to other battery packs.
Finally, in any number of VML tests taken by myself, my staff and by Rob Galbraith, at 640WS power and one shot per 10 seconds using PW Intervalometer, One Einstein 640 consistently produced 460 shots per charge and one AB1600 at 640 WS produced 550 shots. The difference is that AB draws less idle current than E640.