It's cool to see the issue from a different perspective, but the math is just as suspect as all the other arguments, for and against. In this case -- going by a player's percentage of goals scored by his team -- the outcome is affected by factors beyond that player's control, and obscured by factors that make the results erroneous...
* I'm assuming you're not accounting for games the player missed. A good number of guys on your final list lost a lot of games to injury. Not excluding the goals the team scored while he was out drives his percentage down, and the more games they missed, the worse it gets.
* Along the same lines, did the team's goal production drop while they were out, or in the season directly before or after their contributions? Those numbers would be very telling.
* What if a guy played for a team that scored a ton regardless? A number of those guys played for teams that were still offensive powerhouses without them. That can drive their percentage down dramatically even though it's completely independent of their performance. The same is true for guys that played for teams that struggled to score outside of their performance. That can drive their percentage up dramatically.
There are other points to be made here, too, but most become pretty obvious once you start asking these questions.
It's cool that you did this work, and it's certainly an interesting way to look at it. But the injury numbers would have to be factored in for this to be even remotely accurate, that would be an insane amount of data to parse through, and it wouldn't be worth doing because of the many other issues that make this a dodgy perspective to use when trying to answer this question.