Actually, that is false. It is not "too often" it is not "often" at all. As for your assertion that he is so much "riskier" than Doughty,Keith, Weber, Vlasic, etc... please read the following before posting that crap again:
http://canucksarmy.com/2014/2/17/the-reports-of-p-k-subban-s-demise-have-been-greatly-exaggerated
Here is a key conclusion that the stats support:
THE CONCLUSION
First of all, there are no bad players on Team Canada. Every single one of these players has elite strengths, and I could only really see an argument that Jay Bouwmeester doesn't belong on the team. That being said, the evidence points to P.K. Subban being Canada's best offensive weapon from the blueline, and probably the second best defender behind Alex Pietrangelo too.
The argument that he's "risky" or untrustworthy is fueled mainly by biases and misconceptions about what "good defense" is. When Subban is on the ice, he's more likely to get a chance to score the next goal than he is to give it up, which is basically all that matters in this discussion.
At the end of the day, the drop off in defensive ability from Subban to Keith is marginal in small samples, as is the drop-off in offensive ability between Subban and Bouwmeester. We're talking like maybe one shot per game. This adds up over a long season, but keeping Subban out of the lineup isn't going to make-or-break Canada's chances to win the gold medal.
With that in mind, it's still reasonable to say that scratching Subban is poor decision making, as you're incurring the opportunity cost of not having a super-elite blueliner in your lineup, instead replacing him with a good-to-very-good one. And in doing that, Mike Babcock and Lindy Ruff are incurring more risk of losing a tight game than P.K. Subban has ever carried.