August 22 2014 01:30PM
Some of you may be keeping up with the ongoing diatribes occurring in Edmonton following the news of the Oilers' hiring of Tyler Dellow as an analytics advisor to coach Dallas Eakins.
For those of you who unfamiliar with Dellow, he is a longtime member of the sports bloggerati, an Oilers fan, advanced analytics guru and a pretty sharp-minded (and sharp-tongued) fellow at that. He ran his blog, mc79hockey, for a long time and used advanced stats to answer questions that would honestly make most everyone else’s eyes glaze over. Questions such as how the Anaheim Ducks handled faceoff losses in the seconds immediately following puck drop, or zone entries with and without possession, or why Taylor Hall, a phenomenal puck possession and scoring chance player, suddenly dropped off early last season were meticulously researched. He’s also the one who broke the story on Colin Campbell’s e-mails to Stephen Walkom about “little fake artist” Marc Savard.
Why I’m writing about this topic here on FlamesNation isn’t to post an apologia to Dellow or advanced analytics in general. My opinions are on record elsewhere and I’m hardly the voice to put to rest any debate over analytics. Rather, I’d like to bring up the issue here because of what it means in relation to the world of hockey, journalism and as a bellwether of things to come for the Flames organization.
Throughout this drama there has been a consistent tone taken by many members in the media with whom Dellow had had a confrontation, usually through Twitter. They have tended to frame the conversation as being an existential struggle between the traditional media members who have access to the team and see all the practices and games versus an antagonistic, arrogant blogging community who don’t see the game for what it is and would operate the sport based entirely on arcane metrics and vague mathematical terms prone to self-replicating errors. Steve Simmons has voiced his opinions on the matter previously and to Dellow personally. For those unfamiliar with Simmons or Dellow, I'd strongly recommend listening to the audio clip of their conversation. It helps to illustrate the animosity from many in the media and gives one a take on Dellow's personality as well.
Some in the media have attempted to boil down the entire conversation to a confrontation between the played-the-game, saw-him-good crowd versus the arrogant, pretentious, cheeto-eating, mother’s-basement-dwelling number crunchers.
This is not the case, at all.
Most proponents of advanced analytics, especially those involved in the blogging community, have a strong democratic inclination due in part to the nature of the medium. Contributors and commenters will often participate in the process by watching the games, using the numbers to clear the noise from the picture, and then arriving at a reasonably sound conclusion on what is lacking or desirable about a given player, team or situation. These opinions are typically shared on an open forum. For the most part those participating in the discussion have a measure of respect for the work being done and the potential value in the data gathered. Discussions often get heated and sarcasm and ill-will is always available, but provided the comments are well-moderated, good blogs usually promote good discussion.
This bilateral arrangement is antithetical to the traditional media model where a
professionally trained collection of sports journalists are given preferential
access to the team, report on what they see and offer opinions based on their personal history and
observations. They construct a narrative that tells the story of sport using tools they have learned from experience and professional training.
When things work well you get Peter Gzowski’s Game of Our Lives. The best of these journalists can raise the level of discourse and inspire respect and admiration for the game, the athletes, and the fan community. In my lifetime, John Short was an prime example of one such sports journalist, and he told a story that I’ve taken to heart as a general philosophy in the sporting world that can be paraphrased thus: “you can cheer for whomever you want, as loudly as you want. But if you boo, we’re done.” Short had time for everyone's opinion provided it treated the topic and people involved and he extended the same courtesy.
It is generally when the less-admirable aspects of the two
differing perspectives of traditional media and blogging collide that trouble occurs. The blogging community,
steeped in a more egalitarian environment, can often become adversarial with
the mainstream media members who construct unassailable narratives when those stories run counter to empirical data. Greg
Wyshynski mentioned in conversation with Dustin Nielsen on TSN 1260 on August 12th, (to paraphrase) that the blogging
community had its roots in being an ombudsman to the mainstream media, calling
them out when they eschewed logic and facts for convenient narratives or lazy
In the end, they can end up in shouting matches where positions become polarized and entrenched and the opportunity for compromise and understanding dwindles.
So where does this all-or-nothing idea come from?
Nobody, at least not a single person I can find, is suggesting that advanced analytics should be used as a sole source for all hockey decisions. Instead, they are being suggested as a complement to aid in separating the signal from the noise, when it comes to reviewing games or making player asset decisions.
A while back I was listening to Lowetide’s show on the radio when he was interviewing Bruce McCurdy from Cult of Hockey. Bruce had mentioned that the best description of advanced stats he had heard was from someone in baseball (if I recall correctly) who described them as being excellent diagnostic tools, but poor predictors. It struck me how appropriate that description was, and while I would suggest that advanced analytics can provide a measure of predictive value to a team’s decision-making process, they are at their best when they are being used to dissect trends and provide supplementary information to improve player deployment/situational strategy. Advanced analytics also have a very strong role to play in player assessment and acquisition. Instead of talking about a team getting an “ugly win” in vague terms, they can now look at the numbers to see where they got lucky and try to mend bad habits before they become ingrained.
I would also be remiss if I didn’t mention that one statistic in particular has had a surprising relevance in relation to the recent Stanley Cup winners of the past few seasons: 5v5 FenwickClose, an offshoot of Corsi. This suggests that there are some predictive values in the statistical data being gathered.
In all the years I have been reading sports blogs, especially those that make regular use of advanced analytics, I have yet to read any one of the major contributors argue that the entirety of roster decisions should be made according to advanced statistics. Coaching strategies, deployment strategies, and general priorities employed by teams in their approach to the game have all come under scrutiny and been the target of those who suggest that these would benefit from a healthy dose of Corsi, Fenwick and QualComp perspectives. Certainly there have been voices arguing for the acquisition of a player here or there based on some analytical models (think of Kent Wilson's long-distance man-crush on Frans Neilsen a few years back), but this hasn’t yet been taken to the extreme of constructing an entire roster. For illustration, this is a project I have recently begun and whose results I will post at a later date.
So why do some in the mainstream media perpetually frame the issue as though this is a zero-sum game where either the old school method wins out or the soulless mathematical approach bleeds the game of any passion and fun and reduces every match to an equation?
Well, it fits a nice narrative for starters. However, if we dig a little deeper, I think we could uncover something of a psychological motive that isn’t uncommon in professional circles. Specifically, that the resulting affirmation of the blogging community implied by the Dellow hire may be interpreted as undermining or at the very least diminishing the information monopoly traditionally held by mainstream media. Seems kind of far-fetched but...
Remember Napster? If you don’t, Google it and read up. Napster fundamentally changed the business model upon which over half a century of the music industry had been built. And what really galled the music industry was that they had absolutely no say in how it happened.
Some artists and companies recognized that this genie was not going back into the bottle, so they accepted their new standing in the music world and what that would mean for their revenues and adapted. Others had a hissy fit, stamped their feet and effectively shut Napster down. It didn’t stop file sharing and the hegemony enjoyed by online music sites like iTunes today is a direct result of how the companies willing to keep an open mind and look for opportunities in the new model gained an advantage.
Sports journalists are facing a similar problem today in that the information being gathered by the NHL isn’t going to stop. In fact, it is going to increase. A lot. And this information is going to be made public, which means that a whole bunch of bloggers sitting in their proverbial mother’s basements around the globe and eating the figurative cheetos of life are going to write algorithms and code to sort, sift, diagram and extrapolate that data six ways from Sunday. Some already have. Most other areas of the journalism world have already adapted to the extent that unverifiable, and alarmingly limited, social media outlets like Twitter are accepted as bonafide sources of information.
Professionally speaking, sports journalists have a choice, the same choice every profession has when faced with a technological movement that alters existing paradigms. They can challenge themselves to understand the thinking processes behind the analytics movement, attempting to discuss them on the same level as the bloggers with respect for the data but disagreement over its finer interpretation, adopting an inclusive attitude towards the community and their informed opinions (Scott Cullen of TSN and Elliotte Friedman of CBC are two excellent examples of this); or journalists can double-down on blocking out the advanced stats community and refudiate the arguments with clichés and straw-man arguments about having superior perception borne of their unique experience that makes them able to innately determine the finer details of the game and its talent.
If you think the latter is really an option I would invite you to explore the history of human reactionary movements and counter-revolutions for successful examples. It makes for a quick read.
The disturbing part of the latter response is that the responses from media have thus far, at least in the case of Tyler Dellow, been disturbingly personal. Mark Spector and Derek Van Diest have been the notable examples.
I’d like to take a moment to point out that not all of the Edmonton media are critical of the Dellow hiring, and Robin Brownlee wrote an insightful and outstanding piece over at OilersNation about the situation from the more traditional side of the fence, that is, setting personal motives aside and getting to the actual topic at hand.
What the Dellow hiring means for hockey isn’t so much that it will be a watershed moment in the sport, because several teams have been working with advanced stats people for a few years now. What makes it different is that Dellow is/was a major presence in the blogging community, a harsh, some might say caustic, critic of the traditional decision-making processes in the hockey world, and that this was, or at least quickly became, a very public hiring.
Adding a blogger to any aspect of the hockey operations department is going to be noticed around the league. That Dellow was a noted critic of many hockey observers and media members was only going to ruffle feathers. All of this happening in an intense hockey market like Edmonton, with a rabid media industry and intense blogging community both focused on a franchise whose last glimpse of competitive relevance was nearly a decade ago meant that any story written about the addition of Tyler Dellow to the organization was going to be less about the position he was meant to fill than the person in question and the countless emotional investments in the team disguised as editorial pieces.
In the short-term this hiring may mean more for the blogging/amateur statistics community than hockey. Why?
Start by carefully considering the following: the Oilers, under the leadership of Craig MacTavish, Kevin Lowe and Bob Nicholson, agreed to hire blogger Tyler Dellow as an analytics resource on the advice of head coach Dallas Eakins.
Craig MacTavish was an NHL player, an NHL and AHL coach,
and is now an NHL GM.
Kevin Lowe was an NHL player, an NHL coach, an NHL GM, a
member of the Hockey Canada selection committee for several Olympic
tournaments, and is currently an NHL President of Hockey Operations.
Nicholson was the head of Hockey Canada for many years and is currently working
alongside Kevin Lowe as Vice-Chairman of the Oilers Entertainment Group.
Dallas Eakins was an NHL player, an AHL player and AHL coach, and is now an NHL head coach.
Those are all men with a vested interest in the traditional NHL practices, the way they used to do things when they were players, coaches, and so forth. The NHL is historically averse to change (glowing pucks notwithstanding), and these four men have been steeped in North American hockey culture.
All these men looked at what Tyler Dellow has done with advanced statistics and felt his work merited adding him to the organization and that his input held value, to the extent that the Oilers now appear to have proprietary rights to Dellow’s previously-posted work.
Taken in tandem with the hiring of Kyle Dubas in Toronto by Brendan Shanahan, another former NHL player and former NHL Vice President of Hockey and Business Development, as well as the purchase by the Leafs of sites like Extra Skater, these moves signal that the window into the historically-closed world of NHL management has opened, if only slightly. To quote a modern media sage: this is kind of a big deal.
It also indicates that the management of hockey is perhaps moving in a direction that those who claim to be closest to it are vocally opposed. This is a disconnect between the institution and the professed expert observers and that is very important. Usually that kind of noticeable demarcation presages a significant shift towards a new working model. Be aware of this as the next few years of media coverage unfold.
Journalism, the News, and the Industry of Information
Now, this brings me to another point in regards to journalism to which I had alluded. Specifically, legacy, expectations, and message control.
Alain de Botton, in his book News: a User’s Guide, writes of the nature of media in presenting selected world events within a marketable context. In other words, how the news corporations keep themselves financially viable by picking out what stories they believe you need and/or want (if only subconsciously) to hear about and then presenting them to you in order to ensure that you feel it is important that you read (ie: buy) more news.
Within the landscape of the 24-hour news cycle, the serial demands upon virtually all media today, create a situation wherein perspectives and narratives will be constructed with an inherent short-term viewpoint in mind in order to maximize the attention span of the audience.
Solutions to problems or projects that are likely to take decades or even centuries to alter or correct don’t sell and are therefore less likely to be trumpeted as positive, viable options. Catastrophes happen in an instant and are compelling, putting things back together takes much longer and nobody is really all that interested. What can I say? Entropy sells.
This is how news media typically deal with expectations, by
keeping them short but inflated. Grand solutions on the immediate horizon make
for far better reading than incremental improvements spread over an entire
lifetime and beyond.Think Obama and Change, or the War on Terror (Drugs, Crime, <insert vague noun here>). They are presented as defined events whose progress can be quickly and efficiently tracked through discrete incremental stages when in fact they are amorphous, shifting, deeply complex issues that require generations of smaller changes and countless backwards steps before any end result can be reasonably anticipated.
Nobody wants to try and sell that to an audience.
On another note, message control is ubiquitous in the modern media landscape and it extends from the commercial and business world to politics and the sporting world as well. It is often believed that message control is in the hands of those doing the delivering.
It is not.
It is in the hands of the one receiving it: you. It is extremely important that the audience be able to discern when they are being told something sincerely and factually, and when to suspect that the message or messenger is not being so. This isn’t just critical thinking, but parsing the message, messenger, and the medium in order to properly appraise the information being conveyed. It is a lot of work, but necessary given the extent to which information is often parsed in order to achieve political and social goals.
Take the following hypothetical as an example of how this relates to the topic of hockey journalism: when the Oilers encounter struggles during next season there may well be those within the media who will attempt to weave a narrative that equates the team’s situation as being the result of misguided influence from advanced analytical work. Kyle Dubas and Tyler Dellow could end up being fired and the entire advanced analytics movement appear to be a catastrophic failure. But it won’t be, because another team will hire another statistician or another analytical specialist and try again. The legacy of this change is most likely to be incremental and meandrous, but it will be, regardless of how the story is framed at the time.
Okay, enough of my rant about the modern philosophy of news
Back to hockey
Consilience – loosely translated it means taking a collection of information that appears to be unrelated or is separated by various fields of study and deriving from that pool an inference or conclusion.
Consilient intelligence is, in my opinion, one of the great strengths of mankind. Gone haywire it can result in absurd conspiracy theories, but when manifested correctly it is tremendously rewarding. Pythagoras, Newton and Einstein all displayed aspects of consilient intelligence in developing their mathematical theories. Today, the specialization and compartmentalization of expertise is dominant and that can deliver highly intricate results in a narrow field, but the cross-pollination of human knowledge is where the really great leaps forward often occur.
Consilient thinking in hockey is also a necessary development that perhaps the advanced analytics movement could help usher in. Putting together the old and the new, the scouts who see the players and can absorb information about a player’s ability from countless small details of the way he plays, alongside the statistical models that inform the observer about how a player may perform in the more mundane aspects of the sport or provide methods of quantifying that player’s decision-making processes.
For my part, I think that the best scouts are those who
intrinsically see some of the things that the advanced analytics tease out, if
only on an unconscious level. For instance, Nicklas Lidstrom is arguably one of
the best defensemen to have played the game of hockey, yet he did not boast
many of the traditional assets considered to be key for NHL defensemen. He did
not hit, fight, block a great many shots, possess a powerful slapshot, or tower
above everyone in a display of raw physical force. Yet Hakan Andersson saw
enough of him to recommend him as a 5th round pick for the Detroit
Red Wings. I believe Andersson saw Lidstrom play and understood that when he
was on the ice, his whole team got better. Better at defending, better at scoring
and better at passing. He saw in a holistic sense what the numbers would later
reveal. Andersson may have understood on a deeper level that Lidstrom was a
player whose greatest strengths were to improve the play and effectiveness of
the five other teammates on the ice with him in such a nuanced way as to be
almost invisible to many other observers.
Andersson's draft record suggests that he is able to observe and process information that would score well on an advanced analytics study as evidenced by his selection of highly skilled puck-possession-type players like Lidstrom, Pavel Datsyuk and Henrik Zetterberg, to name a few examples.
Scouting, be it professional or amateur, is a tough business
and the margins on return are typically very slim. The salary cap structure of
the league has levelled the playing field significantly, although internal budgets
can provide the illusion of greater parity than actually exists. Consider that the league average success rate for finding a player through the draft who can post a career of at least 200 NHL games is approximately 18%, depending on the year, with no more than five percentage points separating the high and low end of the spectrum, If a small
portion of the overall hockey operations budget were dedicated to one or
two individuals tasked with providing some quantifiable information
based on clear, empirical data that increased the rate of success in this area
by even 2%, it would seem irrational not to pursue it on the basis of an ideological
or personal basis.
Granted, this is the sports business we are discussing here, so nothing is too irrational or illogical not to be considered a good idea by some, nevertheless, it would seem the road ahead is clear to a growing number of people.
The Calgary Flames will eventually follow suit in some form or other by hiring an analytics specialist, whether that be in the area of scouting, development or coaching, it doesn’t really matter. What will matter is what the response will be from media community and the fans.
In Edmonton’s case, there is a fairly clear disconnect between the two. The media have stated their case and it has largely been found to be wanting, both in perspective and professionalism. The blogging community, the closest things fans have by way of representation within a public forum, are generally quite supportive of the decision and have stated their case in a far more objective manner than the media. The fan not connected to the blogging community and advanced analytics doesn’t really seem to care much either way, preferring to wait for results before passing judgement.
The schism between traditional journalism and the blogging community is nothing new. The case of the Oilers hiring Tyler Dellow merely exposed the rift and brought forward the arrogant and uninformed opinions on one side of the floor. What is needed is for both sides to make the other better. Reporters have access that bloggers do not, while the online community can crowd-source ideas and information-analysis in a way that a single reporter cannot. No one group controls the truth of the matter but collectively the sport itself could benefit tremendously from a more…consilient approach.
All this gets back to the underlying fact that sports journalism has historically reached some pretty lofty highs (Gzowski). Over my lifetime I have seen a fair bit of evidence to suggest that the standards have been steadily slipping.
Allan Mitchell, aka Lowetide, a man with a foot firmly planted in both worlds of media and blogging, wrote this piece about the general trend of modern mainstream media back in 2012. To quote him directly: “I’ve been reading newspapers, listening to radio and watching television since the 1960′s and the trusted voice of (mainstream media) seems to have grabbed a cab downtown and headed for someone’s house. Call me crazy but the one thing mass media had as an advantage–staff size, fact checkers, balance and truth–seems to have slipped away one night not so long ago”. (parentheses added)
Did I mention that Tyler Dellow is also a practicing lawyer based in Toronto? No? Would that change your perception of his opinion and work in this area? Interesting, because it is also notably absent in many of the articles critical of Dellow and dismissive of his work. But I suppose some facts are more important than others.
*Note - Steve Dangle has his take up on the Nations. You can read, well, actually watch it here.