Against the Grain: Breaking Down Breakouts

Last time, I took a look at some of the early forechecking data for the New Jersey Devils. This was done using data collected from a new project led by myself and several others to dig into forechecking and breakouts in the same way zone entries and passing have been interrogated. In the first piece, our data informed us that the Devils were generating fewer forechecking attempts in third periods compared to the first two, across all score states except for when they trailed. In this post, I’ll look at the other side of the equation: breakouts. Specifically, how to improve them.

All data in this piece is from 5v5 situations and from formation situations - the forechecking team was in a 2-1-2, 1-2-2, or other formation and does not include situations after rebounds, faceoffs, or turnovers. We track those situations, but early analysis suggests how teams breakout against a proper forecheck (in formation) may be more indicative of skill than the latter situations, which are more prone to 50/50 battles. For now, I separate them.

The Difference Between What Happened and How it Happened

Analytics is really just about using data in various forms from multiple sources to make the best decisions possible. Much of my recent work has been using data to try and improve upon the tactical preparation and on-ice decision making of teams. If there are optimal breakout plays to make, teams should exercise more of these. Granted, there are circumstances where only one play is available, but over the long run, keying in on what makes a play successful and incorporating that into your tactics should pay dividends. It’s one thing to acquire players that perform well based on shot metrics, it’s another entirely to improve your roster based on tactical decisions.

The analytics community excels at identifying good players. Where we are still limited by the public data of our time is identifying how and why things happen. What is it specifically about this player or team that generates offense? Suppresses it? And, most importantly, what are ways that data can inform optimal ways to play? We know we want to create shots and limit those of our opponent, but are there specific plays that do that? We know zone entries are one way to generate offense. I wrote about how effective pressure play is a trait of solid defensive teams and limiting shot assists, but what plays work best in your own zone?

Breaking the Forecheck

What plays am I talking about? This is the basic list we work from on the project.

Now, the analysis you’ll see is from ten games that have been tracked for the Devils. This is all descriptive analysis and, like any small sample size, will fluctuate as we add more games. As I will show through both linear regression (all regressions have a p-value of <0.05 unless stated) and video analysis, the conclusion I’m positing is one that makes sense. Certain breakout plays will regularly lead to higher zone exit possession rates and correlate to lower shots against rates. That’s all we’re doing here: quantifying the plays made to break out of the zone, and then tying those to what we care about - shots against.

Hockey, like many sports, is about manipulating space. When you have the puck you want to create space and force the opposition to defend as much of the ice as possible. By developing and working on structured plays and movements, teams can dictate to the opposition where they move and how they will react to the puck. Forcing the opposition to cover multiple options and lots of space is the optimal way to do this. Again, this is about where teammates get open for passes, but just as important is how their movement forces the opposition to go with them, opening lanes that weren’t there before.

This is the opposite of what the team without the puck wants to do. They should want to squeeze play and force the team with the puck to make plays in small spaces - it’s why conservative forechecking and defensive zone play are terrible ideas.

Apart from Power Plays, there’s very little of the game that lends itself to structured play and easy analysis. Where teams can find an edge is in this space. So, I started by thinking like a coach and came up with two questions to try and solve by using data: 1)“How can we reduce the number of shots we face?”; and 2) “How can we break out of our zone more effectively?”

I had to see what metrics we could devise from the data that correlated well with explaining our shot rates against. Luckily enough, logic is our friend.

The rate at which a team breaks out of the zone with possession against qualified forechecks (see project intro article linked above) correlates well with the total number of shots against they face. As you exit your zone with possession more often, you face a lower number of shots. This makes complete sense to anyone who has ever watched a hockey game.

And this is where we start to get a little deeper. Breakouts with possession are nice, but how do we create breakouts with possession? Is there an optimal way to do that? This is what coaches want to know: how can the nerds optimize my tactics and my team’s on-ice performance?

After running regressions on the various plays (I won’t take up the entire article with charts, I promise), it turns out that looking at the rate of “over” and “reverse” plays attempted by the breakout team will explain seventy percent of their possession breakouts in our sample thus far. But why is this the case? Why are these plays better than others? It’s all in how they break the forecheck by going against the grain. They aren’t the only plays that do this, but they do it most often. Furthermore, by drilling down to this level in analysis, qualitative things like “patience” and “hockey IQ” reveal themselves. Let me show you.

The first clip you’ll see is an example of what not to do, but something that players do all too often.

Damon Severson goes behind his net to retrieve the puck and simply shoots it up the boards. Adam Henrique is over there, but so is an Anaheim player. The puck comes right back on goal. Severson’s mistake is playing into the hands of the Anaheim forecheck. It’s easy to see Anaheim wants him to go that way as they have two forwards applying pressure, another marking Henrique on the boards, and the two defensemen watching the points, but all are moving towards the right side of the ice, as Serverson sees it.

Severson can’t do much and this is why it’s difficult laying blame at his feet. Watch it again and notice Kyle Palmieri swinging through in front of goal. He has to provide support here in the event that Henrique can make a play on the puck and get it to Palmieri, thereby creating an outlet through the center lane. However, the real fault in this breakout lies with Yohann Auvitu.

Severson has time to take a peek, and then has another second to look up when he reaches the puck. Auvitu is in a position that doesn’t allow Severson to make a good first pass. Auvitu should be skating into the corner to the left of Severson to allow for a preferred, against the grain outlet. Watch it again. Auvitu stars off that way, but ends up gliding and swinging right behind Palmieri.

As I mentioned above, when you have the puck you want your players to maximize the area your opponent has to defend. Anaheim would prefer that Auvitu skate alongside Palmieri as it’s one less option to mark. If he skates into the corner with speed, Severson can reverse the puck to him and then Auvitu can easily hit Palmieri up the middle or skate it up himself.

Here’s the same game a minute later and an example of exactly what I mean.

It’s a similar setup as Anaheim applies pressure with two forecheckers on the puck as Pavel Zacha retrieves it behind the net. There’s a forward on the right hand side boards as well - the side Severson sent it up. John Moore goes to the corner behind Zacha. A simple reverse and the forecheck is beaten. When you make a play like this the forechecking team has to gain depth in their pursuit angles over to the strong side of the ice - they can’t cut across laterally as the puck carrier would just blow by them. Had Anaheim’s defense activated to pressure Moore, then they would be able to be more aggressive in their pursuit angles. Teams should be more aggressive and flexible in this manner, but that’s a topic for another day.

The difference between the two examples is where the support is. The simplest way to explain this to players is “be an option.” Auvitu was not an option in the first situation. Moore was an option in the second. That’s oversimplifying it, but it rings true.

This last clip from the game against Anaheim earlier this season is a lengthy one.

The key theme you’ll see in the above video, and throughout many of these clips, is patience. The Devils’ first attempt to get the puck out of the zone by Ben Lovejoy is a terrible decision. Corey Perry is already peeling away to the left to take away that passing option and Ryan Getzlaf is already turning up the near-side boards because Lovejoy has telegraphed where he’s going with the puck. If Lovejoy simply waits a half-second and gets his head up, he’ll see he has support in the center and on the backside.

As the play develops, the puck comes back down to Lovejoy. Vernon Fiddler is shoved into him by Getzlaf, taking himself out of the play as an option. Perry has the pass option behind the net covered, so this is a tough spot for Lovejoy. As is much of hockey, 50/50 puck battles are part of the game. Lovejoy sends the puck up the boards and I can’t fault him too much on this one. The only recommendation I would make is for him to turn back inside and see if he can lay off a pass to Moore who has stepped out in front of goal.

The final attempt of the series is Lovejoy’s worst. The puck comes back down to Perry who gets off a shot. Lovejoy picks up the puck and skates back up the wall. Here is where he should have had some patience. The entire Anaheim team is flowing his way, squeezing the space with which he has to work. If Lovejoy pulls up and reverses course, but opens up his stance here, he has targets at varying levels he can make a pass to. This series is also evidence that playing defense on your off side can severely limit your effectiveness.

On to the glory that is Taylor Hall.

Hall simply takes this puck from behind the net, then reverses to lay the puck back down to a defensemen. You can see by that one quick motion and pass, Chicago’s forwards lose contain and have to circle back, allowing the Devils to take their time pick out an option in the neutral zone.

Here is another situation where getting your head up and having some patience would have been helpful. Lovejoy finds himself in possession of the puck and nudges it right to Devante Smith-Pelley who has a Blackhawk on him. Had Lovejoy been alert, he would have been able to nudge it to Henrique all alone in front of the net. If the option through the center is there, teams should take it each time.

Here is another example of not using your partner or being patient enough. Moore does look, but with Chicago’s forward bearing down on him, panics and sends this out of the zone. As his teammates are swinging through the zone, Moore could take a step inside and fire a pass over to his partner, coming open behind the forward swinging through. Chicago’s Nick Schmaltz (8) is concerned with taking away the play up the boards here, so Moore would have a second to move inside and make a pass.

It wasn’t just over and reverse plays that beat a forecheck in this fashion. Before posting this, I finished up a game from last month and simply made notes on how often teams break out with possession by going against the grain of the forecheck, no matter what play was used. While this a bit more subjective - which is why we track only the specific plays as outlined above - I counted fifteen plays that went against the grain of the opposing forecheck: eleven resulted in controlled exits.

Here is a simply turn by Jaccob Slavin that allows him to get position on Travis Zajac. The rest of the Devils flow that direction and are beaten when he passes over to his partner.

Here there are couple of plays made to break this forecheck. Moore reverses to Palmieri, who then chips it over to Kyle Quincey. Quincey looks up and passes to Zacha. Zacha then gets his head up and fires a pass over to Moore as Carolina has squeezed this side of the ice with Justin Falk jumping into the play. Sometimes it takes more than one good play to beat an aggressive forecheck, but by having options all over the ice, Carolina can only shut them down for so long. Being an option in support for your teammates and recognizing where go against the grain of the forecheck always gives you an outlet.

Much of breakout failures shown above are simply players throwing the puck away, trying to be “safe.” What we’re seeing in our data so far are these plays that routinely, not always, but routinely, counteract the aggression level of the opposition forecheck, it allows for plays to be made to get out of the zone with possession. Playing into your opponent’s hands is not safe. Being patient, going against the grain, and being an option are safe plays.

Conclusions

Hockey is played in shifts. The longer a shift gets, players will start to look for opportunities to change, so it is not always possible to sustain offense/forecheck and change players simultaneously. In order to give themselves a temporary reprieve, they will dump the puck out, change, but never get any sort of sustained pressure and so it’s back to running around in their own end. This is hockey’s version of a “3-and-out” series in football.

In order to break this cycle, data like this can isolate scenarios for each breakout play and corresponding forecheck. After all, when you should use a specific breakout play is just as important as which one you decide to use. The breakout possession rate for “Over” and “Reverse” plays is twice as high against a 1-2-2 than a 2-1-2, which makes sense given than a 2-1-2 will often take away that D-to-D pass more easily. All of this is valuable in the opposition analysis phase leading up to the game. A well-trained team should be able to recognize the situation and get into position - each player reads off of their teammates and the opposition.

The data forces you to look at something and that is where the trend is revealed: breaking against the grain of the opposition forecheck is how you can break your opponent. Data analysis simply expedites the coaching process if it is done correctly. After this step, the coaching staff can then tweak the situational tactics for his team: placement of players, the progression reads, and, thinking ahead, where the optimal zone entry plays are to be made based on these new breakout tactics. Each phase should feed into the other.

Your Thoughts?

What do you all think? What are some things you’d like to see going forward? What questions do you have on this? If anyone wants to help out tracking this season, please don’t hesitate to reach out to me on Twitter as I’d love to get some more data to work with. You’d all benefit!