Over the coming days and weeks here at Inside the Senate, I'll be passing along various statistical breakdowns of some of your favourite Senators based on their 2013-14 seasons. While finishing five points out of a playoff position meant the season did not go according to plan, the team itself and, on an individual level, a number of its members did a lot of good over the course of 82 games. It simply just didn't pan out.
As hockey fans and, more broadly, fans of any sport will tell you, the game is cruel. The Senators were conceivably worthy of a better fate last season and with the benefit of a few bounces here or there very easily could have made it into the dance. Alas, hindsight is just that.
Before we dive into a player-by-player look at how the Senators fared last year, I thought it would be of some value to do an overview of the stats we will be profiling. Let's begin...
To begin, let's be clear: hockey analytics are very much in their infancy.
While the current iteration of stats which are in vogue have been useful in unearthing patterns insofar as what traits are possessed by winning hockey teams or quality players, the work done within hockey's analytic community is nowhere near what has been accomplished by baseball's sabermetric community.
This isn't a knock against the hockey fans and analysts putting in the work across the sport, but rather making light of the fact that baseball had quite a head start. The Society of American Baseball Research (SABR) — for which the term sabermetrics was derived — was established 33 years ago. There is some catching up to do.
Furthermore, while Twitter uses #fancystats as a term of endearment for the game's advanced stats, the stats themselves — unlike Iggy Azalea — aren't terribly fancy at all. There may be a slight learning curve if you're new to the world of hockey analytics, but don't let it hold you down. We're not giving lessons in physics.
The core concept behind the most mainstream hockey analytics surrounds puck possession. In hockey you win by scoring more goals than your opponent. The more you have the puck the less the other team does. The more you have the puck, the more opportunity you have to score goals. The less the other team has the puck, the less opportunity they have to score goals. There is no need to bend your mind in order to grasp why this is important. You probably feel as though you've wasted your time reading this far.
At this point you're likely wondering why stats are necessary at all. Why not just run a stopwatch when a team has the puck? You're not wrong at all. In fact, plenty of teams and coaches will track this information via their own means. Unfortunately the NHL does not track possession time which doesn't help Joe Fan at home watching the game on his couch or in the stands. We need stand-ins. Luckily, they exist.
The above paragraph serves as a convenient segue into...
CORSI AND FENWICK
If you have a passing knowledge of hockey's stats movement you've probably heard the term Corsi thrown about. If you're more familiar than the casual observer, you likely know all about Corsi and have heard Fenwick being thrown around as well. By the end of this section, you'll know the purpose of both and be able to wow your friends, family and loved ones at parties, dinners and gala events.
The first time you dust off this info for said people, however, may mark the last time you're invited to a party, dinner or gala event. You've been warned.
Corsi and Fenwick share the same purpose, but have a different wrinkle in their composition. Both stats are measures of shot attempts towards the goal and both serve as adequate placeholders for actual possession time. The difference comes in what goes into their total.
Corsi counts shots on goal, missed shots and blocked shots while Fenwick counts unblocked shot attempts —shots on goal and missed shots. While the stats can be presented in the form of their totals — in 2013-14 the Senators had 3938 shot attempts for and 3575 shot attempts against while playing 5-on-5 hockey — the statistic may be more easily digested in the form of a percentage — the Senators had 52.4 per cent of the shot attempts on the ice at 5-on-5 in 2013-14.
You'll note that in the above paragraph I used the term "shot attempts" in lieu of Corsi. Many get hung up on the term Corsi as a statistic when, in reality, it is simply named for former Sabres and current Blues goaltending coach Jim Corsi who began counting the stat to determine the amount of work his goaltenders faced in a game. Fenwick is named for Matt Fenwick, who writes at the Battle of Alberta blog, and proposed that the alteration be made to Corsi in an effort to get a better idea of who is creating more scoring chances.
In sum, there is no craziness behind the names of these respective stats, they're named for the guys who got the ball rolling on the concepts. No acronyms, no Illuminati subliminal stuff, no math, just names.
Still with me? Awesome.
Another key difference between these measures is what they tell us about the run of play. Corsi correlates to how much possession time a team will have while Fenwick will tell us more about which teams are getting more scoring chances. That nuance is a key reason why both are useful when evaluating teams and players.
There will be teams that post much stronger possession numbers than scoring chances and vice versa based on measurable skills like shot blocking and tougher to quantify abilities like finding space in the offensive zone for clear attempts on goal. The point is good teams do both well.
Given that many more attempts happen in a game that fall under the heading of Corsi, it is a more reliable way to track teams over a shorter span due to — buzz word alert — the larger sample size. If you're unfamiliar with sample size, it's essentially the amount of information you have to draw your conclusion from.
For example, you have a much better idea of how likely you are to land on heads when flipping a coin (50 per cent odds) if you flip it 50 times rather than 5. Because Corsi draws from a larger pool of events on a game-by-game basis, it is a better predictor of a team's performance than Fenwick over shorter stretches. What's interesting, however, is how likely a strong Fenwick team is to be a playoff contender.
Chris Boyle of Eyes on the Prize put together a solid piece exploring that very concept in April 2013. Apologies for the link to a Habs blog for those who get offended over such minutiae, but good work is good work and should be commended as such. Well done, Chris, your parents named you well.
At this point it should be noted that the majority of valuable data from these headings is extracted when teams are playing 5-on-5 hockey with a close (tied or within a goal) scoreline. Why? The majority of a hockey game is played at even strength and, as we've seen countless times across countless sports, when a team is up by a large margin, for whatever reason, the foot tends to come off the pedal and things tend to balance out — this phenomenon is referred to as "score effects."
The proverbial foot may not even come off the proverbial pedal, but this type of swing happens time and again. By putting the most weight in close 5-on-5 play you get the most accurate picture of a team's true ability.
In essence, if your team passes the Corsi and Fenwick test, it's in a good place to win hockey games. Similarly, if your favourite player performs well by those measures, he's helping the team win the possession battle and, in turn, helping them win games.
At this point I'm sure (or hoping) you see the value in these stats when evaluating teams or players. However, I'm sure many of you are ready to argue that players are used differently. What about the defensive specialists or the match up players? These stats pay no attention to their day-to-day struggles.
Well, about that...
Thanks to the great work of people who track such things, we can actually get a sense of who is starting their shifts in what zone. This is commonly represented by what's called a player's "OZ %" which shows how often a player goes out to take their shift with an offensive zone faceoff at even strength. Hypothetically, if Player X was to hop over the boards for four faceoffs and take two of them in the offensive zone, his OZ% would be 50 per cent.
Again, simple stuff.
For a more tangible example with respect to the Senators, last season Mika Zibanejad took 57.1 per cent of his faceoffs in the offensive zone (7.1 per cent more than the average Senator when Zibanejad wasn't on the ice) while Zack Smith took 41.3 per cent of his faceoffs in the offensive zone, which is over 13 per cent less than the average Senator without Smith on the ice.
While this is useful for seeing the differences in the way a player will be handled, there has been research done which indicates the impact of these starts may be relatively minimal on a player's actual on-ice play.
While the true impact of zone starts has yet to be flushed out, players who post strong possession numbers despite heavy defensive zone starts are of note. It indicates an ability to earn the puck back (via faceoff or takeaway), get the puck moving the other way and put shots towards the opposing goal.
For example, Mark Stone (small sample size alert) recorded a 56.9 Fenwick percentage last season in his 19 games despite starting 46.7 per cent of his shifts in the offensive zone — a 7.5 per cent drop off from the average Senator who didn't line up with Stone. For a player who has drawn praise from coaching and management for his ability to win pucks from opponents using his stick and body positioning in addition to his obvious offensive skillset, the numbers certainly check out.
Regardless of how much impact Zone Starts truly have on how things play out between whistles, at the very least it's interesting to see how a player is used within a team's system and how their abilities are characterized. These numbers give us that tangible comparison.
If you've read this far, congratulations! You certainly haven't won anything but, provided you were new to the topic, you've learned a little bit more about what is sure to be a growing area within the game as we charge fearlessly into the future.
There are more facets of analytics to ideally be covered in future posts — a player's quality of competition, quality of teammates, zone entries/exits, etc. — but for our purposes going forward these will serve as our big three. You'll only see more reference to concepts like Corsi, Fenwick and Zone Starts in mass media as this season gets underway.
The demand for this data has grown considerably in recent months and, as a result, you can expect that demand to be filled. It's my goal that you'll know exactly what TSN's panel is talking about if these concepts come up during regional Senators broadcasts.
If I can leave you with anything that will hopefully rid your hesitance to embrace these ideas it would be that nobody is saying analytics ought to be the be all and end all of player evaluation. The origins of that strawman are unclear — if that mystery is ever solved I can only hope its place of birth is set alight with vigour.
Even the most unabashed analytics homer would (or should) acknowledge the need to have additional information beyond what can be derived from hours of labour in Microsoft Excel. At the end of the day, an NHL locker room is a workplace and, like any workplace, there is more to creating a team than assembling a group of people who hit their numbers.
Looking forward over the long term, these numbers will likely serve two purposes for you:
1) They'll confirm a lot of suspicions you had about the game. If we went back and figured it out, I'm willing to bet Wayne Gretzky had an excellent Fenwick over the course of the 1985-86 season. That wouldn't be a surprise. Good players are good players. Good teams are good teams.
2) They'll make you raise your eyebrows and change the way you look at the game. If these numbers were discovered to simply reaffirm what we had been told about the sport they would be pointless. Prepare to read pieces that challenge your opinions and see lines of numbers that confound how you perceived a sequence within a game. Good players are good players and good teams are good teams, but how you perceive what qualifies as either could change permanently, ideally for the better.
Those may seem like conflicting concepts on the surface but I'd argue that you can have suspicions confirmed while having perceptions altered. You can be correct on a topic and have come to that conclusion from a different way. Broadening your perspective is half the fun of things.
At the end of the day, this is all about hockey, a game, and what's actually happening out there. At its best this is a series of ingredients which can be found in a Stanley Cup winner. At its worst it's something to complain about at the pub with a game on in the background. Hockey talk is just that — let's chat about it.
Questions? Concerns? Need to vent? Shoot me a note on the Twitter at @chrisjlund and I'll do my best to dispel whatever concerns you at that exact moment.
Thanks for sticking it out and say tuned for the next installment of this feature later on in the week.
Want to read more? Here's some suggested reading that has greatly impacted my knowledge and curiosity on the topic of hockey analytics:
The Faker’s Guide to Advanced Stats in the NHL by Sean McIndoe
Eyes on the Prize's "Fancy Stat Summer School"
Broad Street Hockey's Advanced Stats Gallery
Behind the Net's Advanced Stats FAQ
Our friends at Silver Seven also compiled a series of useful analytics links you can find here.
Want to see the stats yourself and get busy being more clever than your friends?
Behind the Net
The Hockey Analysis Stats Database
Corsi measures shot attempts and is a good stand-in for possession time.
Fenwick measures unblocked shot attempts and is a good indicator of who is getting more scoring chances.
Zone Starts tell you where players take their faceoffs.
More interesting features to come.