i think you're incorrect.
better defensive systems, in theory, mean a decrease in quality scoring opportunities and a decrease in shots on net.
It also means a decrease in non quality scoring opportunities...
What numbers peoples have discovered is that better systems tend to decrease both relatively equally, and because of this save percentage is predominately maintained.
There have been multiple different ways this has been looked at and proven:
1) looking at all goalies who have switched team's part way in the season, and then looking to see if the delta is larger than what you would expect just from sampling size (ie: chance)
2) looking at coaches who have switch team's part way in the season, and then looking to see if the delta is larger than what you would expect just from sampling size (ie: chance)
3) looking at a coach's career and how goalies tended to perform with and without the goaltender while making adjustments for aging curves
4) looking at the repeatability of % of shots are shot quality out of total shots
5) looking at the repeatability of shot against location as a talent for a defender
6) looking at the repeatability of on-ice sv% for a goaltender
I can show you links for all of these. They are almost all done by different people, but it all came up to the same answer. Teams and players over the course of a large sample (3k+ SA usually) have little to no effect on save percentage. They do on a micro sense, as **** happens, but over a macro sense you have to remember that **** flies both ways.
The last one is a really intriguing one for both Jets fans as it takes into consideration two controversial players: Pavelec and Byfuglien.
Many people blame Byfuglien for the low sv% he has experienced on the ice
(when he has, although they do not praise him when he doesn't have this), but absolve Pavelec of the low sv% he has experienced.
Here's the thing though, when you look at every Thrasher/Jet defender who has played more than one season for the Jets, then take their relative sv% (on-ice minus off-ice) for once season, and then plot it against the next season.... you get a R2 of 0.006... Basically completely random dots.
So, let's think about this for a minute.
ALL defensemen WANT to stop high quality shots. So if a higher save percentage is something you want to obtain, why is it that no one is consistently better or worse at it???