March 31 2012 11:51AM
If you're going to quote blocked shots in analysis, make damn sure that you're looking at an overall percentage of shot attempts blocked rather than the raw number of blocked shots.
And even then, it still may not matter.
You may remember a short time ago when I looked to dig through a few of the NHL's real-time stats and found that blocked shots actually correlate negatively with winning. When you think about it a little more though, it does make sense. Teams that block a lot of shots are more likely to both:
A - Spend time in their own zone
B - By proxy, give up a lot of goals
This is one of the principle factors behind Corsi, a puck possession metric that counts the number of shot attempts directed at the opposing net minus the shot attempts directed at the defensive goal. Teams that have a high Corsi number are very likely to spend a lot of time in the offensive zone, and vice versa.
So why do many analysts continue to bring up blocked shots as an indication of defensive play? Well, I don't really have an answer for that. Our Pass it to Bulisian friend Daniel Wagner, and a fellow write at the Backhand Shelf, introduced me this year to "Blocked Shot Percentage".
Blocking a shot is not inherently negative. In general, if an opponent is shooting the puck and you have the opportunity to block the shot, you do so. At the same time, blocking a shot is not inherently positive. Preferably, you are the one shooting the puck rather than blocking it.
Wagner's piece was inspired by something written by Derek Zona, who divided a player's even strength blocks divided by even strength shot attempts against in an effort to find out who the best shot blocker on the Edmonton Oilers was in December.
While Ladislav Smid had blocked 51 shots compared to Andy Sutton's 27 at that point, it's worth noting that Smid was on the ice for 434 shot attempts against his own net, and Sutton just 204, allowing Smid to rack up his shot block totals.
But it's easy to find that data for players, but I wanted to see if it had any real world application, as in, does ESBS% correlate well with winning? The answer is "no, but it doesn't correlate with losing like blocked shot totals do." I enlisted the help of Josh Weissbock who stripped team data from timeonice.com for the last three seasons, giving me team numbers in goals, saved shots, missed shots and blocked shots for and against each team.
The r-squared value (0 implies no correlation, 1 implies perfect correlation) of blocked shots to point percentage was -.144. For ESBS% to point percentage, it was only -.032. Not necessarily a good thing, still, to block a high percentage of shots, but it's still better than looking at the raw data alone.
We can logically assume why blocking shots isn't necessarily a positive, but can we also speculate as to why blocking a high percentage of shots doesn't clearly predict good teams? The easy solution would be to look at NHL scorer bias, since different arenas will have different criteria on what a shot block is, but I'm looking a bit beyond that.
If you check behindthenet.ca's shot block tables, you'll see that many of the players who lead the league in having their shot attempts blocked are defencemen. Defencemen not only score at much lower rates than forwards (I calculated 3.9% to 9.6% at this season, respectively) but they also shoot from further away, usually taking shots from the points.
Getting a shot through the first defender and blocked by a scrum in front can create all kinds of chaos and rebound chances, and those are the types of shot blocks that we mostly see.
Most of the shots turned away by skaters aren't generally opportunities that would have resulted in goals in the first place. Perhaps if we looked at away numbers only, we'd start to see a more favourable correlation between blocks and wins, but I still have yet to think of any reason why blocking lots of shots as opposed to simply not giving them up is a positive.