Heat Maps: On Shot Location and the Canucks' Sputtering Power-Play

Thomas Drance
January 14 2014 11:57AM

 
Henrik Sedin in his office on the right side half wall.
Image via wikimedia commons

For fans of the team, Vancouver's power-play has been extraordinarily frustrating to watch this season (and for a couple years now, truth be told). While the Canucks are generating shots-a-plenty with the man-advantage - in fact they're generating 5-on-4 shots at an elite rate this season - the finish just hasn't been there. The club ranks in the bottom-10 in the league in percentage of converted power-play opportunities after 46 games, a ghastly number that makes zero sense when you consider the skill possessed by the club's top-end personnel.

Some of the club's lack of power-play success is a result of pure, unadulterated bad luck; and indeed Vancouver's 5-on-4 team shooting percentage is second worst in the league. While the puck has begun to bounce in Vancouver's favour somewhat more often over the past month, and will probably continue to do so, I tend to think there's more going on here than just "bad bounces".

Read past the jump.

The Adam Oates experience in Washington - and in particular the oddly sustainable success of Washington's vaunted 1-3-1 power-play - has left me open to the idea that perhaps shot quality, or at least shot location, matters more with the man-advantage than it does at even-strength (where it counter-intuitively matters only a little, and in an ephemeral way). The Capitals don't generate a high volume of looks with the man-advantage, but they're one of the best power-play units in the league thanks to a 19% 5-on-4 shooting clip over the past 92 games. The primary driver of that high percentage seems to be Alexander Ovechkin's shot location.

Perhaps the Canucks simply aren't getting their best shooters - i.e. Ryan Kesler, Daniel Sedin, Jason Garrison and Alex Edler - into the best positions to be effective power-play goal scorers. And maybe that's partly why they're shooting such an abysmal percentage this season. Hey it's worth a look, I figure.

A recent post from the stellar, shot quality enthusiast Montreal Canadiens blog "Habs Eyes on the Prize" introduced me to the new "Heat Maps" tool that you can find over at sportingcharts.com. Based on my experience watching games, counting scoring chances, and using NHL.com's play-by-play pages for research purposes, the NHL's shot location data is an absolute mess. So take these charts with a massive grain of salt. 

The Formation

Let's quickly work out what we're talking about before we proceed any further.

The Canucks have mostly played an overload variant at 5-on-4 over the past few season, with Henrik Sedin quarterbacking things from the half-wall on the right-side of the ice. The team actually hasn't changed their formation too much this year despite firing power-play coach Newell Brown this offseason.

Here's how it generally works: Henrik Sedin plays on the half-wall, Ryan Kesler plays net front, Daniel Sedin ostensibly plays in the high-slot although him and Henrik interchange at the half-wall a good deal, Jason Garrison occupies the right point, and Dan Hamhuis or Alex Edler have spent the majority of the season on the left-point (where neither is ideally suited because they're not right-handed shooters). 

Here are a couple of screenshots of Vancouver's power-play formation this season to try and illustrate what I'm talking about visually. Here's one from the second game of the season (Henrik has just passed the puck to Edler, who bobbled it and spun allowing Oilers forwards to close high on him. Edler would shoot the puck wide and the Canucks would retrieve it and score some thirty seconds later):

The things to note here are that Vancouver's iteration of the overload has three men high (in this case Garrison, Henrik and Edler). Daniel and Kesler are both in the slot in this example, with Daniel at net front. Most often that's Kesler, but the two do rotate a fair bit. Notice that Henrik gets zero respect as a shooter...

Here's another screen shot from December.

 

With Edler out of the lineup the Canucks have moved Dan Hamhuis into the left point spot. This screenshot comes from a bit of an odd power-play (the Sedins were on the ice when the penalty was drawn, and PP2 got the offensive zone start at the outset of the man-advantage opportunity). In this example, Daniel is primary half wall distributor because he's swapped spots with his brother (Henrik is along the goal line on the right side). That's the usual Sedin twin interchange that you're probably familiar with. Anyway Kesler is in the high slot and there's no net front player at the moment (though Kesler is headed there). Note that Dan Hamhuis gets zero respect as the weakside shooter...

 

Finally we have an example from last weekend, with Tom Sestito in the net front role (where he was... oddly effective actually) in the game against Anaheim:

 

This is a really good example actually of what Vancouver's overload really looks like. Henrik Sedin will often waltz up and down the half wall, whie Kesler, Daniel and Garrison react to his positioning and try to get open for shots. Because Henrik is low, Garrison is in something of a safety-valve position up top (if Henrik were to stickhandle his way closer to the blue-line, Garrison would move to center ice). Meanwhile Daniel Sedin is in the very high slot in the above screen grab, and is about to skate right down Anaheim's diamond formation where he'll miss a Henrik Sedin pass attempt.

Ryan Kesler, in this example, has taken Dan Hamhuis and Alex Edler's slot on the left point (where he can unleash that dangerous wrist shot). To continue developing a theme: look at the degree to which Andrew Cogliano's stick and positioning suggests a preoccupation with preventing a Kesler' one-timer. Now that's what a penalty-kill respecting a shooter looks like. 

Okay, so now I think we have a firm handle on how Vancouver's first unit power-play formation behaves for the most part. Henrik Sedin quarterbacks things from the right halfwall for the most part while the other players read off of him and "overload" the weakside in anticipation of one-time attempts.

Among those players: Garrison is a mainstay on the right point where he can one-time the hell out of passes from the weakside, Kesler is mostly used in a net front role though more recently has been a third man high on the left-point (a role more commonly occupied by Edler and Hamhuis). Daniel occupies the high-slot in theory but in practice is a freelancer who will pop up high on occassion, interchange with Henrik on the half-wall on occassion, and even move into a net front position as the situation dictates. Then there's a net front guy (usually Kesler, but sometimes Daniel).

Weakside Shooters

You're not going to move Henrik Sedin - the best NHL playmaker of the past five years - off of the half-wall, even if his shot doesn't generate much respect from opponent's penalty-killing units. Fact is, that's almost an advantage since it allows any weakside players to just shoot it off the end boards under pressure and Henrik will have a good shot at retrieving the puck anyway. 

So the relevant and central question, I think is: "are Vancouver's weakside personnel the right fit for the overload system the team plays?" The likes of Garrison, Edler, Daniel and Kesler have all been extraordinarily productive goal scorers on the power-play over the past four seasons, but clearly the finish hasn't been there this year. So is it just luck, or is there something off with the calibration of Vancouver's first power-play unit?

Here's where I'll bring in some heat maps, and compare Vancouver's current personnel with their personnel back in 2009-10, 2010-11 and 2011-12, when Vancouver's power-play was legitimately among the league's best.

Right-handed Shooters

When Newell Brown was fired this summer, he blamed the personnel he was given for Vancouver's complete lack of power-play success. Brown is a pretty smart power-play coach (see: the success Phoenix has had this season), so I think his claim is somewhat credible even if it doesn't serve to explain the 2012-13 Canucks' cratered 5-on-4 shot rate. 

Over the past couple of seasons the Canucks have lost some quality right-handed shooters in Mikael Samuelsson and Sami Salo. Here's a glimpse of the shots they managed to take on the power-play between 2009-10 and 2011-12:

One thing to note is that in 2009-10, Samuelsson wasn't really a weakside shooter so much as he was just a winger on Vancouver's first unit power-play (this is before Newell Brown loaded up the first unit with Ryan Kesler). So the left-side heat map (when Samuelsson filled in on the left point when Sami Salo and Alex Edler were out with injuries) is more representative of what a right-handed shot from the point might do in Vancouver's overload system.

Let's check out how Samuelsson's 2010-11 graph compares with Sami Salo's from 2009-10 and 2011-12 (Salo, you may remember, missed most of the 2010-11 season injured):

Salo and Samuelsson both shot a phenomenally high percentage at five-on-four despite taking long distance shots. Compare their results and heat maps with what Vancouver's most common PP1 LD have managed this season and, well, it's kind of depressing:

Where Salo and Samuelsson were cashing in at an exorbitant rate from the left point on the power-play between 2009-10 and 2011-12, this season Vancouver has gotten next to nothing from the players who most commonly lineup in this slot. I'd imagine that might, in part, serve to explain some of Vancouver's 5-on-4 shooting percentage drag...

A Closer Look at Edler

Looking just at Edler's results this season probably isn't all that fair. After all, Edler has played the left point on the power-play for Vancouver an awful lot over the years (even when Sami Salo and Mikael Samuelsson were on the roster). He's mostly been effective in that spot too:

Similarly during the lockout abbreviated 2012-13 campaign, Edler played a good deal of left point on the power-play (though he also played a lot in Garrison's current RD slot because the Canucks refused to use Garrison on the first unit for some reason):

The big takeaway from looking over Edler's heatmaps over the past few seasons is that he's pretty effective playing the point on his strong-side (the left point spot occupied by Salo and Samuelsson in recent memory). Edler may not be able to one-time passes feathered to him across the seam by Henrik when he play the left point, but he's converting on a good deal of shots and doesn't seem to have any issues getting point shots off.

He's actually a pretty strong option to occupy this slot on the first power-play unit when he returns to the lineup...

Left-handed shooters

Canucks fans who bemoan Christian Ehrhoff's importance to Vancouver's power-play dominance in 2010-11 would do well to look at Ehrhoff's non-existent production in Buffalo and conclude - as we did long ago - that Ehrhoff misses the twins more than they miss the German puck-mover. That said, Ehrhoff's chemistry with Vancouver's former Art Ross winners was undeniable and certainly Vancouver employs zero defensemen who can match Ehrhoff's play off the rush, overall offensive instincts and abilities in the neutral zone. 

When set up in the offensive zone on the power-play, however, Ehrhoff pretty much brought to the Canucks power-play what Garrison brings at the moment. Here are Ehrhoff's heat maps from his two complete seasons in Vancouver:

And here are Garrison's rather similar ones (bear in mind here that Garrison spent most of 2012-13 with the second unit):

If any Canucks skater has a good claim of having been "snake bit" at 5-on-4 this season, it has to be Jason Garrison. During his last season in Florida and his first in Vancouver, Garrison managed to convert on 15.4% of power-play shots. This season he's managed just three goals in 68 drives. While that massive reversal of fortunes may be in part a function of his high usage rate this season, I'd expect some regression over the balance of the season...

Daniel Sedin

Daniel Sedin is, as we described previously, something of a freelancer in Vancouver's overload formation. He's the guy most likely to be up high, or behind the net, or in the slot, or in all three positions on any given power-play shift. Predictably, Daniel's heat maps reflect that:

Boy does Vancouver's cratered 2012-13 5-on-4 shot rate ever show up on Daniel Sedins' heat map...

Looking over these charts a couple of things are plain: first of all it's clear that Daniel's Art Ross trophy winning year was driven in large part by a massive shooting percentage spike on the power-play (though we already knew that). Secondly, it sure looks like Daniel isn't getting the same sort of looks in the slot that he was back when Vancouver's power-play was truly potent...

Daniel's unlikely to shoot 50+% on the power-play over a full seasons ever again, but figuring out a way to get Daniel looks in the slot again should perhaps be a priority for Glen Gulutzan going forward...

Ryan Kesler

Ryan Kesler is the most difficult player to peg here. On the one hand: he's Vancouver's most dangerous right-handed shooter (by a mile) and as a result would seem to be the best bet to effectively fill the Sami Salo slot on the left point with the first unit. On the other hand he's also Vancouver's most dangerous net presence forward (by a mile). So there's an opportunity cost incurred when you play him in either spot, and Vancouver just doesn't really have the personnel to replace him adequately. 

To the heat maps!

There's a couple of things worth pointing out in the above heat maps. First of all in 10-11 - his first season with the first power-play unit - Kesler was exclusively used at net front (with a slight lean to the left side). Like Daniel, Kesler similarly shot an outrageous percentage that season.

In 2011-12, Kesler remained primarily a net front presence, but the team began to look for him at the Bowman line as well (Kesler took 9 more wrist shots and 10 more slapshots in 2011-12 when compared with 2010-11). During the lockout abbreviated 2013 campaign when Kesler was mostly injured, he was barely used at the net front at all, which probably reflects Newell Brown's desperation to find a right-handed shooter for the point.

Finally this season Kesler has mostly been at net front, but has occassionally been used on the left point as well and in particular has found himself there of late.

Conclusion

We don't really know decisively what sort of an impact shot location has on shooting-percentage with the man-advantage, and until there's SportsVU installed in every NHL arena we probably won't. The location data is simply too unreliable at the moment.

While Vancouver's anemic power-play shooting percentage looks to my eyes like a mix of luck and odd personnel decisions, a cursory glance at the heat maps we pulled up today would suggest that having a right-handed shot on the left point has been critical to the club's power-play success in the recent past. Kesler is the best bet to fill that slot, but I'd imagine that having Edler or even Yannick Weber instead of Kesler in that spot represents less of a drop off in talent than replacing Kesler at the net front with one of Alex Burrows, Tom Sestito, Zack Kassian or Chris Higgins...

Hopefully when Edler returns to the lineup the club just puts their four best power-player shooters (Edler, Kesler, Garrison, Daniel) on the weak side. You can always design a set or two where Edler rotates into the high-slot, Daniel takes the net front and Kesler rotates up to the point, after all.

3136ae487fac57943f99a50e66e4d6cf
Thomas Drance lives in Toronto, eats spicy food and writes about hockey. He is an NHL News Editor at theScore, the ex-managing editor of CanucksArmy.com and an opinionated blowhard to boot. You can follow him on twitter @thomasdrance.
Avatar
#1 NM00
January 14 2014, 01:38PM
Trash it!
8
trashes
+1
5
props

There's been a lot of talk about trading a defenseman (usually Edler) to upgrade the forward group.

Considering his plummeting goal rate the last couple of years, at what point does Daniel Sedin find himself on the trading block?

Because riding shotgun with Henrik Sedin is a pretty sweet gig and surely there's a better way to spend 16% of the cap than on what can be expected from Daniel & Alex Burrows...

Avatar
#2 Cam Charron
January 14 2014, 01:47PM
Trash it!
2
trashes
+1
3
props

I maintain my skepticism in NHL shot location data, and also a skepticism that shot location is an indicator of strategy or talent.

Granted, people will want to find reasons for why this team can't shoot on the powerplay, but the important thing to me is that it looks like it should be getting more goals than it is, and you can't really teach finish. You just have to account sometimes the puck won't bounce your way.

I'd also point out that the Capitals are a high-volume shooting team on the PP, with a 59.9 SF/60. If they had six more 5-on-4 shots in their 226 minutes this year, they'd have the same rate as the Canucks. I don't think you get quality without volume, or volume without quality, as a lot of PP shots get created off of scrambles and rebounds.

Avatar
#3 Cam Charron
January 14 2014, 01:49PM
Trash it!
4
trashes
+1
12
props

@NM00

at what point does Daniel Sedin find himself on the trading block?

And... there go the chances of the comments thread staying on topic...

Avatar
#4 NM00
January 14 2014, 02:08PM
Trash it!
15
trashes
+1
3
props

@Cam Charron

People are focussed on the thread about last night's moral victory.

While I don't want to discourage Drance from keeping up with his New Year's resolution, what exactly have we learned from reading his work that we didn't know already?

Or your addendum to his post for that matter?

"the important thing to me is that it looks like it should be getting more goals than it is, and you can't really teach finish".

What has this subjectivity taught us?

But B+ for staying on topic...

Avatar
#6 wes
January 14 2014, 02:13PM
Trash it!
1
trashes
+1
1
props

illuminating and still super hard to understand.

Avatar
#7 Cam Charron
January 14 2014, 02:35PM
Trash it!
1
trashes
+1
1
props
Thomas Drance wrote:

@Cam though the Capitals' shot volume has increased this season, it was real low last season. I don't think it's a stretch to describe their success as "percentage based" rather than volume based, frankly, especially considering the past 92 games.

It is percentage-based, but I think the Caps brass understood some changes to the system were necessary to keep the PP rate high this year. I'd like to see a breakdown of their system. Their success is now a little more sustainable, but where are the extra shots coming from? Are they dump-ins? Rebounds? Their shooting percentage has dropped by 14% but their scoring has increased by 6%.

Volume does matter in the long run more than percentages.

Avatar
#8 JCDavies
January 14 2014, 02:42PM
Trash it!
3
trashes
+1
0
props

@NM00

"what exactly have we learned from reading his work that we didn't know already?"

I see what you did there...

Avatar
#10 dahim0301
January 14 2014, 05:03PM
Trash it!
1
trashes
+1
0
props

I'm no advanced stats guru, but a ran some numbers really quickly:

If our PP% over the past 47 games had been a very respectable 20% (rather than the terrible ~15% we've seen)... that would only account for an additional ~7-8 goals over the 47 games we've played so far (~.15 increase in G/G). I suppose this could have resulted in a couple more wins (depending on the games where these goals occurred). However, is this enough to make a significant difference in the Canucks prospects this season? I'm not saying it isn't, but would love to hear your thoughts on this.

P.s. I think the difference is much more significant if we could achieve a 25% PP%... but that doesn't seem very realistic.

Avatar
#12 van
January 14 2014, 05:23PM
Trash it!
2
trashes
+1
2
props

Given we know our pp formation, instead of using unreliable shot location data, why not look at who is taking the bulk of our pp shots, compared to say, the Capitals, Hawks, Blues and Sharks.

Shots/60 of players with most PP time/game (1st Unit)

Canucks: Garr 29; Kes 21; Daniel 13; Henrik 9; Edler 12 Total: 84 S/60

Hawks: Sharp 18; Kane 14; Keith 12; Toews 9; Shaw 6 59 Total: 59 S/60

Capitals: Ovechkin 21; Green 16; Backstrom 9; Brouwer 8; Johansson 6 Total: 60 S/60

Sharks: Marleau 18; Couture 17; Pavelski 14; Boyle 13; Thornton 9 Total: 71 S/60

Blues: Steen 18; Shattenkirk 14; Pietrangelo 11; Oshie 8; Backes 8 Total: 59 S/60

Double checked those Canucks numbers, which are insane. If you want to compare Garrison to an NHL team, Tampa and Jersey muster 41.5 shots/60.

Obviously having more than a third of our powerplay shooting at 4.5% is a big factor. Some of that will be bounces, but Garrison's hardly the most agile guy - goaltenders will know that's the Canucks default move and they'll be set well in advance. Hawks top 2 shooters by volume are Sharp and Kane, Caps have Ovie and the Blues have Steen. Kesler's obviously at least matching these guys in volume, but I'd guess the a big portion of his shots are rebounds and tips rather than set ups.

If we're going to run our power play through Garrison as we are, we have to be more effective around the net, which is more or less what Torts keeps going on about. Maybe adjust the formation and get another guy closer to the net to help Kesler.

Avatar
#13 NM00
January 14 2014, 06:12PM
Trash it!
10
trashes
+1
3
props
Thomas Drance wrote:

@NM00 you're right you can easily find writers taking the time to breakdown VAN's PP formation and how various shooters have performed on PP1 over the years elsewhere on the internet.

Because that's exactly what I said.

Do you want a gold star for effort?

What ground breaking discovery is being made here that can only be gleaned by internet writing as opposed to, I dunno, watching the game?

"While Vancouver's anemic power-play shooting percentage looks to my eyes like a mix of luck and odd personnel decisions, a cursory glance at the heat maps we pulled up today would suggest that having a right-handed shot on the left point has been critical to the club's power-play success in the recent past. Kesler is the best bet to fill that slot, but I'd imagine that having Edler or even Yannick Weber instead of Kesler in that spot represents less of a drop off in talent than replacing Kesler at the net front with one of Alex Burrows, Tom Sestito, Zack Kassian or Chris Higgins...

Hopefully when Edler returns to the lineup the club just puts their four best power-player shooters (Edler, Kesler, Garrison, Daniel) on the weak side. You can always design a set or two where Edler rotates into the high-slot, Daniel takes the net front and Kesler rotates up to the point, after all."

While I don't necessarily disagree, how did the elongated post supplement your subjective opinion?

That said, I certainly don't want to discourage you from writing...

Avatar
#15 NM00
January 14 2014, 08:18PM
Trash it!
6
trashes
+1
3
props

@Thomas Drance

Again, I'm not disagreeing with your suggestions.

Well, except A since I much prefer B.

And, sure, the team could use two Keslers for the PP.

But that's all pretty intuitive...

Avatar
#17 pheenster
January 14 2014, 09:09PM
Trash it!
4
trashes
+1
6
props
NM00 wrote:

Again, I'm not disagreeing with your suggestions.

Well, except A since I much prefer B.

And, sure, the team could use two Keslers for the PP.

But that's all pretty intuitive...

I see you're moving on from being merely arrogant to being an arrogant d!ck. Keep going pal, soon you'll make the varsity.

Avatar
#18 Cale
January 14 2014, 09:55PM
Trash it!
2
trashes
+1
1
props

fuck it, the markup on these comments is a pain in the ass

Avatar
#19 sean
January 14 2014, 10:31PM
Trash it!
2
trashes
+1
0
props

they should put kesler on the second unit powerplay and put kassian in front of the net, deflecting shots and banging in pucks from the weakside, with garrison and weber on the point throwing potential bombs.

I think the problem is that when you spend so much time shooting, you stop aiming, you stop looking for open seems. Torts mantra, although reasonable is to just throw as many pucks at the net as possible, and while thats great and all, its not going to work well when the shots are getting blocked or saved. Garrison has a bomb, theres no denying it, but that isnt going to stop goalies from making positional saves.

I think the boys need to go for the 1-3-1/umbrella style, with hank on the sideboards, daniel in the middle and kesler or kassian in front, weber on the left side high slot, with garrison or hamhuis quarter backing the point.

Perhaps we need to spread out the talent aswell, and go for two functioning powerplay units. I can see these lineups being able to perform given a formation of 1-3-1: PP1: Hank, Daniel, Kassian, Weber, Garrison PP2: Higgins, Kesler, Santorelli, Hamhuis, Edler

Another thing is that it seems we spend too much time heading towards the boards as soon as we gain the zone. i know that defending teams want to pin us against the boards, as we would want to pin them against the boards, but I bet if you looked back at the last couple seasons PP units, you'd likely find that they were able to work the puck to the point, retrieve it and get set up along the half boards in time to catch a PKer out of position trying to pitch in. This season it seems we go to the boards, with 3 guys down low, get the puck, try and move it to the point, where garrison gets a one timer, and either misses or the goalie controlls the rebound (either by getting a hold of the puck or by angling the shot if it gets through into the boards)

looks like now-a-days we dont set up someone directly in the middle of the O-zone, for quick one timers anymore, and our defence never pinches down the middle, and only along the boards (and while thats important to keep the teams forecheck going they should also find opportunities to fall down into the slot and possibley catch the defenders by surprise)

It just seems so stubborn, the whole situation, its looking like old dogs dont learn new tricks.

Avatar
#20 NM00
January 15 2014, 07:28AM
Trash it!
4
trashes
+1
3
props

@Thomas Drance

Perhaps a different alignment would decrease the shot volume (back to 2012-2013 levels), increase the shooting percentage and the end result would be the same...

In any case, I'm quite sure the coaching staff would love a second Kesler.

And perhaps that's the crux of the issue with the PP specifically and the goal scoring decrease since 2010-2011 in general.

Deployment/alignment/wizardry can only do so much to make up for a personnel issues...

Avatar
#21 antro
January 15 2014, 11:20AM
Trash it!
3
trashes
+1
4
props

@TD: great article, I learned a lot from it, and from most of the comments...;)

Comments are closed for this article.