clock menu more-arrow no yes mobile

Filed under:

From the editor's desk: the evolution of advanced statistics

What once was looked upon with skepticism is now becoming embraced by members of the advanced statistics community.

Tom Szczerbowski-USA TODAY Sports

"Shot quality".

That term tended to be one of the major counter-arguments against writers who predicted that teams like the Minnesota Wild, Toronto Maple Leafs, and Colorado Avalanche were likely to see an overall reduction in production after defying possession based metrics the season before.

Such arguments were frequently discounted by statistics supporters as qualitative analysis in the guise of quantitative data. While intuitively the concept of shot quality would seem to make sense, with no way of consistently tracking and evaluating it, it was almost impossible to assert causality in trying to explain why certain teams could be out-possessed so thoroughly and yet still win.

Nowadays, shot quality is making its way back into the discussion. But this time, it's the statistics writers leading the charge.

Sites like War-On-Ice are looking more thoroughly into which parts of the offensive zone tend to result in more scoring. Termed "danger zones," these areas are classified into low, medium, and high "danger" areas based on scoring chance conversion rates on those particular sections of the ice.

Steven Burtch of Hockey Prospectus has a far more in-depth primer on how scoring chance data factors into our understanding of offensive production that's worth a read. But the adaptation of "danger zone" analysis reflects a merging of quantitative analysis and qualitative observation.

The ground work for statistics like these is laid by largely unsung members of the hockey community who track individual games and record the events they see. It is somewhat ironic that the common retort towards statistics-based writers is that they need to "put down the spreadsheet and watch the games," given they likely watch more games than most hockey fans.

Which ultimately is why I find the antagonism between traditionalists and more progressive writers to be a bit puzzling. The stats community has had more than enough "victories" to allow their work to speak for itself. And the future of NHL statistics are likely rooted in player tracking and situational analysis. Both require keen eyes for the game to fully appreciate.

The merging of qualitative observation and quantitative analysis is well underway. And as more men and women devote more time and attention to better understanding the game we all love, the way we interpret what happens on the ice will change too.


  • So how did the Arizona Coyotes perform in 5v5 high danger scoring chances (HDSC) last season? Unsurprisingly, not well. They had just 45.5% of all HDSC while at even-strength, which was good enough for 26th. The teams worse than them? Toronto, Edmonton, Colorado, and Buffalo.
  • One area that HDSC statistics might significantly improve our understanding of is goaltender evaluation. Goalies are notoriously inconsistent (ask Mike Smith about that). Will weighing scoring data based on location on the ice help us see which goaltenders are more consistent in high pressure situations?
  • Player tracking and situational analysis seem like the two areas the Coyotes attempted to address with the hiring of John Chayka this summer. It will be interesting to see if ownership springs for automated tracking technology like NBA teams have adopted.
  • There was no small amount of negative backlash associated with last weekend's article considering the benefits of bringing back Martin Erat or David Moss. That didn't come as a surprise.
  • One of the ideas I want to maintain here is that no topic is out-of-bounds. It can be difficult to avoid homogenizing opinions when bringing on new writers. My hope is that we continue to produce pieces that both support and challenge conventional thought surrounding the Coyotes, even when members of our staff disagree with each other.
  • For what it's worth, I'm fine with the Coyotes parting ways with Moss and Erat in favor of clearing space for younger players like Max Domi and Anthony Duclair. Even if both forwards are not going to transform the Coyotes into a Stanley Cup contender overnight, the team has to start playing young talent.
  • What are the Coyotes' future plans for Boyd Gordon? He will be 32 come October. The Coyotes aren't looking for him to provide offense, and they clearly did not want to let him go after the 2011-12 season. Will they try to sign him to a multi-year deal after this season?
  • While having Joe Vitale also on the roster makes juggling personnel a bit challenging, I think Gordon's penalty killing ability is a rarity in the NHL, as are his shot suppression abilities. If the Coyotes can convince him to take a bit of a pay cut from the $3 million he will make this year, he'd be a great piece to retain.
    Courtesy Domenic Galamini

Courtesy Domenic Galamini