Thanks for the description of the test. Makes much more sense now.
So you found 2 to 3% of leak by the valve stem when the pressure difference over the valve stem is 100 psi or 7 bar. When actually shooting the pressure difference will be much larger (at it's maximum somewhat lower than 85 bar and decreasing rapidly) and so the leak will be larger as well. But because of the small space between valve stem and the hole it sits in the critical flow will be reached quite soon.
Nice to see how you proved and quantified there is quite some leak along the valve stem.
The volume of lost air will be small compared to the total amount of air used per shot because of the short time the valve is open.
The valve will probably open a few times (2 to 3) for each shot because of the hammer bouncing. (see this high speed movie: http://airgunaccuracy.wordpress.com/...hammer-bounce/
). I think you can save much more air when you reduce the hammer bounce than when you would when sealing the valve stem.