As I said in the original post. I used the numbers from war-on-ice.com. I looked at the team stats, all situations (pp and pk's included) strength wise. They are cumulative over the 3 season plus what has been played so far this hear period. I did not separate out individual season.
I specifically avoided using just even strength because I wanted power plays and penalty kills to be accounted for. If you are trying to protect a lead having to go on the penalty kill hurts your chances of holding onto it. Likewise going on the power play helps. I want those situations reflected.
Most analysis of possession numbers is limited to 5v5 ES for a reason. You're not going to get an honest gauge of how well a team is playing by lumping in all that TOI together.
All you have to do is look into the loose interpretation of how every penalty is called from game to game to know that it won't give you an assessment of how the game is going.
And it doesn't even take into account the type of PP/PK you're on... is it a 5v3 situation? Is it 4v4? Is the goalie pulled?
Some PPs are just flat out better at getting shots on net (like San Jose - so if you're a divisional rival of SJ, your corsi% on the PK is going to tank as opposed to teams in another division) and some PKs are intentionally passive and successful at it.
And by lumping in all 3 seasons, you're not painting an accurate picture here. Over the past 2 seasons, the Devils have been on par with, if not one of the better teams at pushing play while leading.
The 2011-2012 Devils weren't a very good possession team until they got closer to 2012 and were godawful with a lead -- and that's pretty explanatory when you look at how the roster developed over that season and when you parse the numbers of individual players/lines. That's also dragging down the whole set of numbers over a 3 year sample, which is a little misleading.
Even if you just look at 2013. That's a season that deserves to be looked at alone, league wide (not just NJ). You have half the league not playing against each other and player conditioning all over the place.
If you look at the drop off in Corsi% leaguewide when leading, it's much more significant than the previous or following season -- likely for the above mentioned reason.
You did a lot of work here, and I don't like being the one to rain all over someone when they do this. I'm just going to agree to disagree with how the data was lumped together. The only time I'm really comfortable interpreting multi year stats is when looking at an individual players WOWYs/CorsiRel% to see if there's a trend.