Right, he has no ability yet he's one of the best on the team at controlled zone exits:
http://pattisonave.com/luke-schenn-skills/
I'm not saying the guy is PK Subban but he's a pretty decent passer/puck handler. Certainly better than MacDonald and Grossmann in that regard, not that that says a whole lot.
This article ignores Schenn's QoC, which gives him a huge handicap when looking at possession stats. You can't directly compare his numbers to Grossmann/MacDonald when they're playing against top lines, not that it's even impressive to be edging them out to begin with.
Also, it assumes that Schenn bares no responsibility for his low on-ice shooting %. It's not a coincidence. Schenn allows some really dangerous chances. As I type this, Schenn just let Grabovski skate around him 1 on 1 and get a free quality scoring chance in the slot.
All of this stuff is well researched man. The effects of QoC are minimal at best on a players on ice results:
http://nhlnumbers.com/2012/7/23/the-importance-of-quality-of-competition
Same goes with quality of chances and the goals against stuff. Corsi is more reliable in small samples. Goals are pretty infrequent events over such small samples. This causes a ton of variation and makes it unreliable in comparison to shot attempt metrics. Essentially he's catching blame because his goalie doesn't make saves when he's on the ice.
It's an interesting start, but Eric's conclusion is VERY premature (the data he really needs isn't available/possible). For one, he assumes that small variance in QoC numbers compared to individual Corsi (much wider) indicates that the frequency of "matchups" is overstated. That's a big leap to make without concrete data to support it. With stats like QoC, you see smaller variants in large part because you're pooling from a much larger set of players, so the data will regress closer to the mean.
Actually, a good point to take away from that article is that, while stats like Rel QoC are useful in determining overall usage, there's not a one-to-one statistical relationship we can draw from to compare players of different usages (as of yet). Advanced stats are tremendously useful, but they have to be looked at judiciously and taken with a grain of salt, because they don't tell the whole story.
And you're right that goals are infrequent events, but they're pretty important events. It's convenient to look at Corsi, Fenwick, shot differential, etc., since they have a wider data set to pool form. However, that doesn't make them perfect stats. If the goalie is having a tougher time making saves when Luke Schenn is on the ice, that doesn't necessarily indicate a statistical anomaly that will correct itself over time.
