Thank you again for posting this.
It appears that there's actually no noticeable trend prior to 1998-99, then it becomes very noticeable and consistent. Would you agree?
Zero net difference per game in first five seasons (10832-10830), with results all over the map: +2.36, -3.00, +2.64, -1.93, -1.31
Then, a 34586-32346 difference between road and home, with positive results every year, and very positive results most years, with a net effect of 3.75 per game.
I'd be inclined to not credit Brodeur with any shot undercounting prior to 98-99 (interestingly this is when his sv% results started to tank, the team also started to take significantly more penalties for a few years) but then starting in 98-99 give him an extra 0.94 shots and saves per game.
I'd also be inclined to not look at every season as an individual entity (as in, this season deserves 1.55 extra shots per game, this season 0.57, etc) because there's no evidence the effect was significantly stronger in some seasons than others, just strong evidence that it did exist.
Thoughts?