So far this week I’ve introduced some of our newer metrics using our data on the Toronto Maple Leafs. We’ve looked at general shot contribution and on-ice data as well as network and linkup data. Today ,we’re going to look at something new that may help us understand more about how teams generate offense and where teams fail to defend the opposition.
Understanding where the offense originates while preparing for a specific opponent can provide great value. If I know which lane and zone a player is likely to linkup with another, perhaps I can scheme for such a situation. If a LW-LD combination is getting overrun down their side of the ice, yet the LD has decent left lane numbers apart from that LW and the LW’s terrible numbers persist irrespective of who is behind him, I know there’s either a communication breakdown between those players, or that the LW is more likely being propped up by the LD.
Digging deeper into how the game can be analyzed with new data is the first step in how we’re going to answer some of these questions. All data is 5v5 unless otherwise specified, and is through games completed as of 12/4. This represents 13 of the Leafs first 26 games this season.
Passing Lane Corsi
Passing Lane Corsi is something new I’ve created this year. We are going to take a look at which sides of the ice teams attack the Leafs, and where the Leafs favorite path to attack lies.
These shot assists are the total number of passes completed that led to shot attempts and the lane in which they were attempted. It tells us how a team will move and attack down a certain path of the ice, and it will tell us how a team will also defend these lanes. Here, we see Toronto with a slight advantage down the left side of the ice, but at a disadvantage centrally and down their right side of defense. An example of how this looks in a game is from their 11/12 game against the Nashville Predators. Here are the charts from that game, followed by the breakdown of lane Corsi.
An important thing to understand about these charts is that, again, left and right are from the offensive team’s point of view.
The Predators were busy everywhere. They also were effective and exiting the zone on the left and then breaking across through the neutral zone at times.
The Leafs were able to complete some passes down the right on their breakout, but very little down the center of the ice or down the left. Nashville was especially dominant down the middle. Here’s a look at the numbers.
The right side of Toronto’s defense and wingers were overrun, along with their centers. Only on the left side were they a match in terms of sustained possession. Naturally, the question is, “Why are teams able to move across the center and right lanes against Toronto?” Well, let’s have a look at this game and see if we can’t figure something out.
We’ll start with Dion Phaneuf, who had the worst Right Lane Passing Corsi (2 for, 7 against) of the Leafs defensemen. Let’s take a look at a sequence in which he probably could have done better.
This starts after a chance for Colin Wilson and the missed shot goes to the corner where Mike Ribiero retrieves it. All five Leafs are down low, essentially playing 5-on-3 below the tops of the faceoff circles. It’s obvious Ribiero will go back to the point to sustain offense and force the Leafs to come out a bit.
Barret Jackman collects Ribiero’s pass and promptly sends it back down the left wing.
Ribiero collects and the Leafs have opened up a bit. Craig Smith has position on Nick Spaling down low, and Colin Wilson will drive to the net here. Gardiner is probably out a bit too far here and it will get him in trouble as Wilson is already moving and Gardiner will have to turn and go from a static position.
Which is exactly what happens. Ribiero sends it down to Smith, who has position on Spaling to collect, turn, and send a pass out front to Wilson. Gardiner fails to front Wilson and a chance is yielded.
This sequence could have been avoided at a number of phases: 1) Phaneuf could have closed down Ribiero quicker on either pass to the point or down low to Smith; 2) Spaling could have checked Smith to prevent the pass out front; and 3) Gardiner could have prevented Wilson from getting body position for the chance.
After the chance, Wilson crashes into the goal and play is halted. Phaneuf misses it because he still thinks Ribiero is on the half-boards.
We pick up this next sequence only moments later, but both teams have changed personnel. Mattias Ekholm collects a puck and passes it across to Ryan Ellis.
Ellis surveys the ice and this is where off-the-puck movement can dictate defensive coverage and manipulate where the passing lanes and soft ice are. Here, Colton Sissons will drive across the center lane, pulling Shawn Matthias with him.
This opens up a lane for Ellis to send a pass to Austin Watson. Notice how much room he has on the left (Toronto’s right) side of the ice.
Here, Watson fires a shot that is saved and does not amount to much danger. Had Watson been left-handed, he likely could have had an easier time picking out the stick of Gabriel Borque who is fronting Roman Polak. At the same time, Nashville missed a great opportunity here. Had Ellis skated down into the top of the house, he likely could have presented a great option for Watson to pass it to. Bozak isn’t in position to recover had Ellis followed his pass.
Moving to second period and some transitional play. This is just after a Nashville powerplay has expired and they are regrouping in the neutral zone. Roman Josi sends the puck over to Ryan Ellis at the near boards.
Ellis sends a quick pass on to Craig Smith as he’s moving with speed through center ice. Bozak us unable to check Ellis to prevent the pass.
Smith dishes to Ribierio and the duo enter the zone. Smith drives to the net, pushing the defense back and creating space for Ribiero.
Ribiero is a great passer and is able to thread one past P.A. Parenteau and over to Miikka Salomaki on the opposite side of the zone.
You can see here that Phaneuf has again found himself on the wrong side of good positioning as Smith now has excellent position and presents his stick blade for a possible tap-in. Ideally, you’d want Smith and Salomaki to switch positions due to their handedness, but if a good pass finds its way to Smith, this can still be a dangerous chance as it would be a bang-bang play for James Reimer. Luckily for Toronto, Salomaki simply shoots it and Reimer covers.
All in all, here’s how the defensemen performed lane-wise in this game.
Now, you’ll see these be quite similar for each pairing since it’s just a single game, but, as we’ve seen above in the pictures, quantifying how and where the puck is advanced can provide insight into whose responsibility it might be for that particular shot sequence. Let’s have a look at the forwards.
Apart from the Leafs top line of Kadri, JVR, and Komarov, it was not a pretty night. Players are either responsible for a man or for an area of the ice while defending. Quantifying where teams create from, where the origin of the end result (shot attempts) begins, can inform how we evaluate defenders, which is something we can’t quite do at present.
It’s one thing to say, Phaneuf had a 39.5% Corsi and Gardiner had a 24.2% Corsi, but it’s another to review how each chance was generated and be able to determine that perhaps Phaneuf could have done more to interrupt possession for Nashville down his side of the ice, or that Gardiner could have done better positionally to support Phaneuf. These sorts of things are what we should be doing to build metrics that are more accurate of player performance.
Let’s move from one game to the totals we’ve recorded across these thirteen games.
Passing Lane Player Totals
We’ll begin by looking at the Leafs defense by lane. These are the total number of completed passes down each lane that led to a shot attempt. Starting with a look at the left lane, this includes all passes originating in the left lane of the ice for and against for Toronto.
The blue bars are the passes for Toronto and the red are those completed against Toronto. The green line shows the possession percentage. I’ve included both so you can which percentages are over a smaller or larger sample. So, over the course of these thirteen games, we see that Phaneuf and Gardiner do well via the left lane, more so than any other defender. We can also see that team are most successful at moving down Toronto’s left wing with Reilly, Polak, and Hunwick on the ice. Let’s move to the center lane.
When we move to look at the center lane events, we see an uptick across the board. Only Hunwick is below 50%. Basically, most of the Leafs do okay without Hunwick on the ice, but tend to get caved in with him. Hunwick is the only Leafs defender to have a negative Center Lane Relative Passing Percentage.
And now the troubles begin. Even though Phaneuf and Gardiner are still above 50%, the rest of the blue line takes a major dive. Rielly and Hunwick take the brunt of it. We know Phaneuf typically plays the right side and Gardiner the left, so we can attempt to quantify what to expect when both of them are on the ice.
Let’s move to the forwards.
And here is where you start to see a greater ranger in the percentages. Lots of forwards will mix into the lineup with various line mates. For forwards that primarily occupy the left wing, this metric will mean more to them than it will a forward primarily playing on the right wing. Granted, there will be certain situations that result in forwards playing on either side of the ice.
Moving to the center lane.
There’s fewer peaks and valleys here than there are on the wings.
And now to the right lane.
This is the first metric that shows Kadri under 50%, so it might simply be a poor run of games against teams with strong left wingers and left defensemen (remember, the Leafs right side will go up against the opposition’s left side since tracking is done from the offense’s point of view). That’s the fascinating aspect of new data – the questions that arise from it.
This will best be answered in tomorrow’s post, but knowing how things happen or how teams want to make things happen can provide a significant opportunity on how to combat the opposition. If a coach is presented with a shot chart (like those on the fantastic war-on-ice or hockeystats after a game or a run of games, they will see where the opposition has been shooting from. Naturally, a coach or video analyst will want to prevent those shots. IN order to do that, you have to prevent the chance from being created. When and where that happen are far more important than the end result. Using lane possession metrics offer more insight into where the offense is coming from, while also giving us an on-ice possession metric that tells us a little more than overall Corsi. Using data to pinpoint where the team excels or needs to improve is where we need to be moving towards. You likely have heard me talk about how to improve hockey analytics and make it coach-friendly. I believe with the data we’re collecting (and will be released on Friday), this is a small step towards that.
Conclusions/What’s Still to Come
One of the obvious first questions that comes up with this data is how much blame/credit should we attribute to the left winger and left defensemen for a completed pass? It may be something we can isolate further by looking at left, right, and center passes by zone (which we can do with our data).
As much of our work remains in the early stages for this season, numbers will fluctuate until they settle. This week will offer a few new ways to analyze players with more granular data. Next up, I’ll take this concept of Lane Corsi and apply it to opposition analysis of a particular opponent. I’ll also go over the numbers for that specific game and deconstruct what happened and why.
Have any questions? Hit me up on Twitter @RK_Stimp