Selke Trophy voting: a valuable exercise in futility

Selke Trophy voting: a valuable exercise in futility

The evaluation of defensive play throughout NHL history has always been a work in progress, especially when you get outside of the bubble of NHL coaching and management and delve into the media and general public’s opinion on it. While we can go back as far as the start of the NHL in 1917/18 to see offensive statistics like goals, assists and points, in terms of defensive play we’re left to nothing but the word of people at the time who saw them play, and, if we’re lucky, some grainy footage. Here’s a selection of quotes about Frank Nighbor, who is essentially known as the godfather of defensive forwards, playing in the 1920s:

“So far ahead of all players in defensive ability, and in outguessing the opposition, there is only one Old Master, and aspiring players have a star to aim at.” – The Ottawa Citizen

“Instead of poking the puck off someone’s stick, he had a knack of trapping the puck with a hook-check and bringing it back to his own stick as if the puck were on a string.” – King Clancy

“It was nothing to see his own defensemen resting on their sticks and his goaler stifling a yawn as (Nighbor) massacred eight out of 10 plays that came through his center slot.” – sportswriter Red MacKenzie

“Frank Nighbor is the greatest defensive hockey player I have ever seen.” – Hall of Fame builder Tommy Gorman

And that’s basically it. All I know about Frank Nighbor’s defensive play is from the opinions of people that I have heard are important saying this guy was really good defensively. No video or stats (outside of the impressive deep-dive into his slot defense in that one game) to cross-reference that, it just is what it is.

Fast forward 100 years and it has become a joke in hockey circles to reference plus-minus as an indicator of good defense. Giveaway and takeaway stats are unreliable. Corsi (which measures the number of shot attempts that occur when a player is on the ice) is only effective to use if you’re referencing a specific formula of it that applies a specific weight to how often a player is starting in the defensive zone vs. offensive zone, the expected goal calculations that factor in shot distance and type, how your teammates compare in this metric, and how the opposition compares as well. 

Despite the wealth of resources we have at our disposal now from an array of on-ice analytics and individual player tracking to unprecedented access to video, you will still never find a consensus opinion on something like the Frank J. Selke Trophy, awarded to the league’s best defensive forward.

The Inexact Science of Evaluating Defensive Play and Voting on the Selke

Prior to 1996, a differing number of voters each year from the Professional Hockey Writers Association (PHWA) submitted Selke ballots with a first, second and third place ranking of their top defensive forwards in the league from that season. From 1996 to the present, it became a ballot of first through fifth place. All these votes were added up with a 5-3-1 point designation in the three-player ballot years, and a 10-7-5-3-1 point designation in the five-player ballot years to tabulate a winner based on the forward who received the most total points from the various number of ballots submitted. 

As an example, Ryan O’Reilly won the 2019 Selke Trophy with 1,001 voting points, earned from his inclusion on 151 of the 171 ballots submitted (48 first place votes (480 points), 43 second place votes (301 points), 31 third place votes (155 points), 18 fourth place votes (54 points), and 11 fifth place votes (11 points)).

You can see the disparity in opinion for gauging defensive play in the voting for the Selke Trophy itself. From the time the Selke was first awarded in the 1977/78 season up to the 2018/19 season, there have been 1,931 forwards that have played at least 100 games in the NHL in this span, a reasonable number to assume they found some kind of footing in the league. In this same span, 535 unique forwards have found their name on at least one Selke ballot, so roughly 28 per-cent of the players in the league. That’s a pretty decent chance of getting at least some kind of recognition as one of the very best defensive forwards in the NHL. 

If you’re Patrice Bergeron, receiving at least one vote means that in his career he has been included on 1,272 ballots that stated he was a top five defensive forward in the NHL in a given season. If you’re Trent Hunter, receiving at least 1 vote means that in his nine-year 500 game career, one voter thought he was the fifth best defensive forward in the NHL in the 2006/07 season. Maybe a relative worked for the New York Times that year and got a vote, or maybe that voter meticulously poured through data and footage to come up with an informed opinion that Trent Hunter was indeed the fifth best defensive forward in the NHL that year. Did you watch enough of 2006/07 Trent Hunter to dispute that he wasn’t? Because I didn’t. We didn’t even have publicly accessible Corsi stats in 2006/07, so it was basically the Wild West for public opinion on defence. Maybe he really was that good.

This difference in opinion reached its pinnacle in 2007/08 where on 133 different five-player Selke ballots, 81 different forwards found their name on at least one of them. 294 forwards played at least 60 games that year (rough cutoff where players start to get penalized for being included on ballots due to a lack of games played), so that’s again 28 per-cent of the league being represented in some fashion as one of the top five defensive forwards in the NHL, not too far off from one player per forward line. If we compare that to the 134 Hart Trophy ballots from that season, awarded to the most valuable player to his team, only 20 different players received a vote, and the field of players is expanded to 465 when you bring in defencemen that played at least 60 games, and goalies that played at least 40.

Theoretically, this doesn’t make sense. With the Selke, you’re just looking at one aspect of the game at the forward position: who is the best at helping his team keep the puck out of the net. If someone is trying to gauge overall value for the Hart Trophy on the other hand, they have to look at every aspect of the game and try to come up with a weight for everything. Most valuable should incorporate offence, defence, transition play, powerplay effectiveness, penalty kill effectiveness, etc. 

And then you have the task of comparing different positions; how do you compare a forward to a defenceman? Or a forward to a goalie? And while some voters don’t bother considering this, you also have to think about their surrounding talent. Does that player really elevate his team more than someone else, or would they be just fine without him?

And then you see that Alex Ovechkin received 128 of the 134 first place votes. Oh, 65 goals. That’s right. Forgot about that.

While there is more involved to offence than just goal and point totals, points are still a pretty great and effective measure of offensive talent. It’s a telling stat for which players are creating the most opportunities while still being effective at delivering on those opportunities to put the puck in the back of the net, the entire goal of the game.

With defence, the goal is just as simple. You’re trying to keep the puck out of your net and get the puck back from the opposition so you can put it in the other team’s net. However, there’s no number or metric even remotely close to having the impact that point totals have on measuring offence. The measure of defence should really be seen as:

Everything that a player is doing without the puck when the opposition has it (with consideration given to what that player is doing without the puck when his teammates are in control of it to anticipate if something may go wrong).

So, if we want to take on the task of properly evaluating this and come up with an informed Selke ballot, we have to watch all 1,271 NHL games in a season. 

60 minutes per game… That’s about 53 days of hockey to watch every season with an eagle eye for every nuance of every forward’s defensive game to give a fully informed opinion.

Analytics have done a ton to bridge this gap but they’re still far from perfect. The focus has continually gone from less common events like goals scored against while on the ice to more common events like shot attempts while on the ice, to the quality of shot attempts while on the ice to try to make some sense of this incredibly lengthy time players spend defending. 

But even with all the with-or-without-you stats out there which show the on-ice performance of different players away from their teammates to try to isolate their impact, it still doesn’t isolate what an individual player is doing on the ice at any given time. 

Hockey is a very random and chaotic game; trying to make sense of it with an on-ice metric is nearly impossible. Player tracking is a huge step up, but there’s still a lot that can’t be measured in terms of seeing how certain plays develop, or whether a specific player could have done anything about a certain negative event to reasonably prevent it from happening.

Using Selke Votes as a Guide

So, who even cares about the award if you can’t properly measure defensive play? I think this thought makes sense if you only want to see it as a black and white case of finding out who the hands down best defensive player is, but if we look at it instead as a list of forwards that people around hockey consistently bring up as the best defensive players, chances are it will act as a good tool to use for identifying players that are good defensively, so we can watch them and draw our own conclusions about them. 

This is the same concept as scouting. There’s a consensus big board for each draft year but scouts are still influencing picks based on their own eye tests and opinions. 

The general managers making the picks may watch the player they’re drafting a handful of times, but those handful of games are still something to go by, even if it’s not completely comprehensive. If you’re drafting first overall and there are two players you’re deciding between, maybe an NHL team puts together a video of every single shift they played from the previous season, so they can get the most informed take on those two prospects as possible. But are they going to cross-reference that with every other draft-eligible player’s full catalog of shifts just to be sure? No.

Of course, this doesn’t mean just because someone has a vote for the Selke that his or her opinions are informed. Some may just look at the plus-minus column or points, or purposefully submit their ballots to try to pump a certain player up and bring another player down, but the thought is if you have this many people submitting ballots who are around the game of hockey, you’ll get a lot of good takes in there, and the cream should rise to the top, even if it’s not nearly a foolproof method (because nothing is).

NHL Awards voting often gets labeled as voters being biased towards the team they cover. However, if you see someone from a team-related affiliate vote for someone on the team they follow that you personally never considered, instead of rolling your eyes and accusing them of being biased, give it some thought and try to isolate that player next time you see him play. 

Watch him with increased attention away from the puck. A biased vote doesn’t necessarily mean it’s completely unfounded, the voter still has some reason for including them relative to another player they watch a lot on their team. And if that voter watches a lot of that team, their opinion counts for something. 

If there’s a decision between a playeryou KNOW is good compared to a guy you’ve HEARD is good, obviously you’re going to go with the guy you know is good. If you’re interested in understanding why they voted this way, watch their player of choice play and see what conclusions you can draw on your own. The more you watch, the better.

Eye Test for Defense

When you are watching those players, or watching a retired player in one of the classic hockey games on TV when they air, here are some things you can look for to help gauge defensive effectiveness.

Physical Aspects of Defense

Skating

This is an obvious one. Whether it’s top speed and acceleration to factor in on the backcheck, or agility and edgework to prevent puck-carriers from going around or to block off different passing lanes, skating is a common trait you’ll find in a lot of great defensive forwards. Selke mainstays like Fedorov and Gainey for example were known as excellent skaters.

In today’s game, a fast-riser in the Selke conversation is Anthony Cirelli, and when you see how he moves, it’s easy to see why. Here’s Cirelli chasing down one of the best skaters in the league, Mathew Barzal, on the backcheck:

Strength on the Puck/Hand-Eye Coordination

When you hear a commentator saying “active stick” chances are this is what they’re talking about. Taking positioning out of the picture, the ability to strip pucks, bat passes down or clog passing or shooting lanes is an artform in itself. 

This isn’t strictly physical; it requires understanding of leverage and deception as well. There’s only one real option to showcase this one in Pavel Datsyuk, but players like Mark Stone and Ryan O’Reilly (especially considering his ability to avoid taking penalties with his stickwork) are good examples from today’s game.

Physical Strength/Range

Outside of the strength players use on their stick, using their body or larger frame is a big part of what can make a player effective defensively. This doesn’t just mean throwing the body around but using a bigger frame to bully players off the puck, or win contested puck battles. Big, imposing centers throughout history like Bobby Holik, Michal Handzus and Martin Hanzal could do this, and you can see that same thing at play with bigger forwards like Anze Kopitar, Sean Couturier and Aleksander Barkov today. Tomas Hertl is effective at doing this as well, with an example of him using his body to help win a puck battle in the clip below:

Effort Level/Tenacity/Hustle

Effort isn’t something that can strictly be seen as moving your feet, as I’ll show later with awareness, but the value of a player zooming around the ice being a pest to puck-carriers can’t be discounted when evaluating defensive play. 

This is the mindset that typically gets associated with a fourth line grinder, players with a high motor that aren’t typically looking to conserve energy on the ice, but instead go all-out every shift. An example of this is Mike Richards’ shift from the playoffs in 2010, where he disrupts a play at his own blueline, chases down a player on the backcheck, then hunts down a loose puck to win a puck battle and score a goal:

Mental Aspects of Defense

Awareness/Positioning

When evaluating defensive play, everything that player is doing on the ice at all times matters. Where they position themselves is everything for influencing how a play develops. This is where you can really start to differentiate centres from wingers. 

A lot of people claim there’s a Selke bias against wingers, but in reality a centre just has a lot more responsibility in the defensive zone, along with usually being the support playing higher than the wingers outside of the defensive zone. In their own end they’re helping out both down low and in the slot, and are far more engaged around the puck than a winger is. There are also a lot more decisions they have to make in support.

As an example of awareness and positioning, we can isolate Patrice Bergeron in this penalty kill sequence vs. New Jersey:

The play in itself is simple, but if you break it down you can understand what Bergeron is doing here and how he’s helping to cut down scoring chances and eventually secure the puck to safely clear it out.

Slide 1

You can see that when this sequences starts as Jack Hughes moves into position with the puck on the half wall, Bergeron is put in a situation here where he has two threats to defend:

  1. #63 in white Jesper Bratt in the slot, who is in a high danger part of the ice to score, but as a left-handed shot isn’t in the best shooting position
  2. #97 Nikita Gusev at the top of the circle, who is further out from the net than Bratt but as a right-handed shot is in a very dangerous position to score off of a pass for a one-timer.

Slide 2

Within a second Bergeron swivels his head around to keep track of the passing option to Gusev.

Slide 3

As Gusev slides up the ice to try to give Hughes a passing lane, Bergeron gets his stick in the lane to take that option away, while still staying close enough to Bratt in the slot to take away a scoring option if Hughes decides to go inside.

Slide 4

With that option taken away the Devils powerplay resets and works the puck back to the point to draw #63 in black, Brad Marchand, away, while Bergeron keeps his eye on the puck movement and moves into the shooting lane, still in Bratt’s vicinity. 

Slide 5

The Devils then work the puck back to Hughes, who now has more space to work with. Despite Hughes having more of a lane to the net, the Bruins still have a defender in the way of that lane, and Bergeron’s duties remain unchanged. 

He stays close to Bratt in the slot, while whipping his head around to keep track of Gusev getting set to receive the pass for a one-timer.

Slide 6

As Hughes skates closer into the net, Bergeron just becomes a player that gets lost in the jumble in the slot from Hughes’ point of view, out of the picture tied up with Bratt. However, you can see Bergeron knows what’s coming and separates himself from Bratt, ready to take away the passing lane to Gusev he knows Hughes is going to try to use.

Slide 7

As expected, Bergeron closes the passing lane, intercepts the puck, and makes the clear.

This all occurs in a matter of seconds, but you get a sense of how quickly an effective defensive centre has to think in terms of keeping tabs on different threats, identifying which threat poses the highest risk to score and which ones they can trust their teammates to handle, and then ultimately making the calculated risk to make a defensive play, whether that’s closing a passing lane, stick-checking the puck-carrier, or eliminating them with a hit. 

This attention to detail doesn’t have to display itself in “hustle”. As you can see, Bergeron barely moves at all, but rather his mind is still working quickly to process what is going on on the ice.

Support

Coupled together with positioning and awareness is the ability to recognize when your teammates may be in trouble and step outside of your usual lane to make a good play. As a centre, it’s a crucial aspect of their role.

Going back to Bergeron, you can see as the puck gets dumped into the Bruins defensive zone, he recognizes Zdeno Chara’s lack of footspeed to handle Jordan Eberle on the forecheck and, after taking a peek behind him on both sides to track where the other Islanders forwards are coming into the Bruins zone in case something goes wrong, he goes back behind the net to support Chara to make sure the Bruins secure the puck:

Deception/Influencing the Play Without the Puck

Defensive play is often thought of as reactive. The goal of hockey is to take the puck and put it in the net, so the team in possession of it dictates what happens, with the team on defence reacting. 

However, good defensive players will still be able to influence what the opposition does through deception, making the puck carrier think they have a lane just to take it away, or doing things that are counterintuitive to what the puck-carrier expects. 

They can also use certain areas of the ice to their advantage to force the opposition’s hand, such as stepping up on plays in the neutral zone at the blueline because of the threat of the opposition going offside, or playing up tight against the defence when they’re holding the blueline on the attack.

The idea of trying to influence the offence to do something you want them to do is a defensive tactic that can of course be applied to different sports. Richard Sherman, one of the best cornerbacks in recent memory in the NFL, was famous for making opposing quarterbacks think he was out of position to induce a throw to a wide receiver he was covering, only to close that gap.

The same thought can apply to defensive play in hockey. 

A lot of effective defensive plays happen when the defender gives a player space only to take that space away, or effectively makes the puck carrier think they aren’t involved, only to suddenly step up. 

An example of this would be this play by Artemi Panarin. As a forechecking strategy he doesn’t use his skating ability and tenacity, he uses his smarts and deception to crouch behind the net so he can jump out unexpectedly at the Sharks’ defender circling behind the net with the puck.

Commitment

This can be tied into both positioning and effort level, but it basically means that the player isn’t looking to cut corners with their defensive assignments. While certain players are less involved around the puck and always have one foot out of the defensive zone for when possession changes, a good defensive player is one that sticks to the coach’s gameplan and doesn’t waver from it to cheat for offense. 

While dependent on defensive structure and far less common today, certain players are sometimes tasked with shadowing top forwards, and it’s a very visual example to display this level of commitment. Steve Kasper on Wayne Gretzky is an example of this; you can tell he only has one duty out there, and that’s removing Gretzky from the game:

Leave a Reply