This is part four of a four-part series by Charlie O'Connor describing and evaluating the penalty kill of the Philadelphia Flyers in 2015-16. In Part One, we outlined and defined the different formations that the Flyers' penalty kill used in their forechecking and defensive zone strategies. Part Two delved into the manually-tracked forechecking metrics. Part Three did the same for defensive zone metrics. Today, we use our findings to recommend potential adjustments to the Flyers' PK strategies in 2016-17.
This week's series on the shorthanded units of the Philadelphia Flyers was, first and foremost, meant to be an informational exercise. By breaking down each of the most-used forechecks and defensive zone schemes, the hope was that readers would begin to recognize these patterns while watching games in the future. Also, by providing statistics that measured the efficiency of each of those tactics, it would provide a glimpse into the thought processes of coaches, who are constantly trying to evaluate the best practices for their systems.
But this type of deep dive into tactics and statistics also gives us the opportunity to speak intelligently about how the team might go about improving the PK. Our new knowledge of the effectiveness of these tactics, combined with an understanding of the tactics themselves, makes it possible to suggest how the Flyers' penalty kill could adjust for next season to become a better shot and goal prevention unit.
Before I outline those recommendations, however, let's recap our basic findings.
What do we know now about Philadelphia's penalty kill?
First, on the forecheck:
- The Flyers' main forechecking tactics were the Passive 1-3, the Retreating Box, Same-Side Press and Tandem Pressure.
- The Passive 1-3 was used 39.42% of the time -- by far the most of any tactic.
- The Retreating Box came in second at 24.57 percent, and saw its usage skyrocket by the Flyers over the final three months of the season.
- The Tandem Pressure forecheck was the most effective in generating full breakups of rushes and in limiting opponents' ZEFR rate -- yet it was only used 11.04% of the time.
- The Retreating Box was the second-most effective in limiting ZEFR rate, while the Passive 1-3 ranked third and the Same-Side Press came in last.
- The Passive 1-3 was the most effective forecheck in preventing successful opponent entries into the offensive zone from turning into full power play formations or dangerous rushes.
And then, in the defensive zone:
- In the defensive zone, Philadelphia mostly shifted between a Passive Triangle +1 formation and a more aggressive Czech Press.
- The Flyers used the Passive Triangle +1 a whopping 72.2% of the time, and Pressure-Based defensive zone tactics just 27.8%.
- However, Pressure-Based tactics did a better job by 25-30 percent at limiting defensive zone shots allowed over the Passive Triangle +1.
- Despite this, the Flyers' overall defensive zone shot prevention metrics did not get worse over the final three months of the season, even though the penalty kill's ratio of Passive-to-Pressure tactics increased.
- After accounting for the type and number of entries (controlled, uncontrolled, and faceoffs) that the Flyers allowed in 2015-16, the penalty kill performed at basically a league-average defensive zone shot suppression rate, based on the data from Corey Sznajder's All Three Zones project.
- The Flyers' worst three penalty killing months were caused by three distinct issues -- in October goaltending, in December neutral zone entry prevention, and in January defensive zone shot suppression.
Recommendations for 2016-17
Armed with this information, it's time to make some suggestions on what the Flyers could do next year tactically in order to improve the state of their penalty kill. The goal is simple -- return the Philadelphia PK to where it was just three years ago, when it held a position as one of the league's best shot and goal prevention units.
Recommendation No. 1: More Tandem Pressure
The most glaring disconnect between our findings from the forechecking data and the Flyers' penalty kill usage percentage was the fact that the Tandem Pressure was the least used yet most effective forecheck when it came to halting opponents' rushes and preventing them from setting up in the Philadelphia zone. Employed just 11.04 percent of the time, the Tandem Pressure successfully broke up 39.29% of rushes, and finished the season with a ZEFR Against Rate of 38.10% -- another team-best.
But the Flyers seemingly jettisoned this tactic in the 2016 calendar year, as its usage dropped below 10% in January, February, March and April. Even though the Philadelphia penalty kill delivered some of its best shot and goal prevention rates in the later months of the season, that's no reason to justify the abandonment of the Tandem Pressure. The metrics are clear -- it should be given a second chance.
That's not to say that the Tandem Pressure should be used as the base of the Philadelphia penalty kill. It can't be used effectively in all situations -- it's at its best when opponents are regrouping deep in their own zone -- and it does carry risk. Our ZEFR tracking implies that risk is overblown, but there likely is a threshold of usage (or matchups against certain opponents) where teams will begin to exploit the Tandem Pressure.
Still, I find it hard to believe that a move back to the 20 percent range (which is around how much it was used at the start of 2015-16) would surpass that threshold. And even its worst-case scenario -- that only the two remaining defensemen are left in the neutral zone in the event of the forecheck failing -- isn't much different than a failed Same-Side Press, except the Tandem Pressure usually sees the two forwards backchecking furiously and therefore skating with speed to help attack the opposing power play as it tries to set up in the offensive zone.
Recommendation No. 2: Use Pressure-Based tactics more often in the defensive zone
Here's the big one. If you're looking to really move the statistical needle in terms of shot and goal prevention for the Flyers' penalty kill, this is the recommendation that would have the biggest impact.
As I noted on Wednesday, Philadelphia used the Passive Triangle +1 a whopping 72.2% of the time, leaving just 27.8% for pressure-based tactics like the Czech Press. But the stats indicate that the penalty kill gave up less shots of every type when employing aggressiveness in the defensive zone rather than sitting back.
|Defensive Zone Strategy||Times Used||Shots on Goal Per Formation||Unblocked Shots (shots on goal + missed shots) Per Formation||Shot Attempts (shots on goal + missed shots + blocked shots) Per Formation|
|Passive Triangle +1||322||0.54||0.89||1.29|
This table seems to be the smoking gun in favor of pressure-based tactics taking up a larger percentage of the mix in 2016-17. But one curious fact remains. As I found yesterday:
But now we run into a legitimate issue with our data. If the Flyers performed better in the defensive zone when using pressure, and they used less pressure over the final three months of the season, the logical conclusion would be that their penalty kill became less effective as they cut back on pressure. Interestingly enough, that wasn't the case.
[In 2015-16 there was] no real correlation between "frequency of defensive zone pressure" and a good Flyers penalty kill.
If pressure is so much better than passive defensive zone tactics, then why did the Flyers post two of their best penalty killing months when very little pressure was used in the defensive zone? Are pressure-based tactics even necessary? Did the Flyers get better at passive techniques as the year progressed?
Fortunately, we can answer these questions by looking at the shot prevention outcomes for each defensive zone tactic on a monthly basis. As it turns out, our explanation is not that the Passive Triangle +1 became significantly better at shot suppression at the tail end of the year. It's that the Flyers' pressure-based tactics, though used much less than earlier in the year, got really really good.
Aside from a terrible month of January, the Flyers generally allowed around 0.80 unblocked shots per usage of the Passive Triangle +1 strategy. Pressure-based tactics produced consistently better shot prevention outcomes, but they reached their peak effectiveness at the end of the year. If anything, the team's late-season improvement in defensive zone shot suppression can be credited to pressure-based tactics, rather than being a repudiation of their usefulness.
So what are the arguments against adding more pressure in the defensive zone to the gameplan? One is that increased pressure may help in terms of shot suppression, but it will result in the penalty kill allowing a higher percentage of dangerous chances. After all, increased aggressiveness could result in penalty killers ending up out of position more often, leaving areas like the slot and net front more open for power plays.
Let's test this theory. Luckily, Corsica.Hockey added scoring chance data over the summer, so by dividing 4-on-5 scoring chances allowed by total shot attempts (Corsi) allowed on a team level, we can determine what percentage of shot attempts permitted by the Flyers were of the highest quality. And it turns out that over the past two years, the Flyers have been near the top of the league in limiting scoring chances relative to all shot attempts allowed. Last season, they ranked sixth with just 22.42% of all attempts categorized as scoring chances, and in 2014-15 they were fifth with 22.89 percent.
But just because the Flyers have been efficient at preventing 4-on-5 scoring chances does not guarantee that passive defensive zone structures were the main cause. Nor does it even mean that preventing scoring chances relative to total shot attempts is a repeatable skill. Both of those assertions must be tested.
We'll start with the latter. If scoring chance suppression is a repeatable skill, you can expect that teams will mostly retain it year-over-year. Sure, there will be slight personnel shifts and some teams might adjust their tactics, but we should still see a relationship between team performance in year one vs. year two if it truly is a skill. Via Corsica.Hockey, we can go back as far as the 2007-08 season to test this relationship.
An R-squared of 0.0896 implies that there is something of a year-to-year relationship, but it's not particularly strong. Therefore, it's probably not prudent for a team to view strong scoring chance suppression percentages in one season as a guarantee that the next year will produce the same results.
And there's another reason why the Flyers' recent move towards passive defensive zone strategies on the PK shouldn't be credited for the team's success in suppressing scoring chances. In the two years preceding this most recent stretch (2012-13 and 2013-14) Philadelphia also ranked in the NHL top-five in scoring chances allowed relative to all shot attempts. In fact, during the 2013-14 season, no team was better in that area.
While I did not manually track penalty kill data from 2012-2015, the general consensus locally is that Philadelphia dialed back on their defensive zone pressure starting with the 2014-15 season. What this data shows is that even if you believe scoring chance suppression on the PK to be a sustainable skill, the Flyers were already successful in this area prior to cutting back on the pressure. Passive defensive zone play clearly isn't a prerequisite to limiting scoring chances, making it less risky to re-introduce more aggressive tactics.
The other argument against increasing pressure is more difficult to fully rebut -- the game theory argument.
In addition, there's a game theory element at play as well. Maybe tactics like the Czech Press worked better precisely because they were used infrequently. Teams grew accustomed to the Flyers sitting back, so when they did attack 28% of the time, it caught opponents off guard. Overuse of pressure-based tactics potentially could blunt their effectiveness.
Without league-wide manually tracked data, it's tough to fully refute this theory. My best response to this line of thinking is, "What's the harm in trying more pressure?" After all, pressure-based tactics were superior in terms of shot suppression each month last season for the Flyers, even when the distribution was closer to 60/40 in favor of passive defensive zone play. There might be a passive-to-pressure ratio threshold where pressure-based tactics stop being more effective than passive ones, but Philadelphia clearly has not found it yet.
And the rewards are potentially enormous. Move last year's 72/28 passive-to-pressure ratio to just 50/50 (assuming that the "shots per formation" averages remain stable despite the change), and that shaves 23 unblocked shot attempts off the Philadelphia total allowed for the season. Such a change would have placed the Flyers' PK sixth in the NHL in Fenwick Against per 60, sixth in Corsi Against per 60, and fifth in Shots (on Goal) Against per 60.
With that kind of reward and with the risk seemingly limited, I feel comfortable advocating for more use of pressure-based tactics from the Flyers in the 2016-17 season.
Recommendation No. 3: Stick with the Retreating Box/Passive 1-3 hybrid and scrap the Same-Side Press
The Flyers may have actually lucked into this adjustment. The Retreating Box became a larger part of the Philadelphia PK game plan in February, right around the time Sean Couturier went down with injury. Through my video study, it was obvious that Couturier was far and away the most effective forward in the F1 role in the 1-3 formation, whether that meant harassing the puck carrier as the only real pressure on the Passive 1-3, or directing traffic to the desired side of the ice on the Same-Side Press.
Possibly due to the absence of Couturier, the Flyers ramped up their usage of the Retreating Box, a forecheck that places less emphasis upon the assertive defensive instincts of the forwards involved and more upon forcing the power play to commit to their entry strategy and then reacting as quickly as possible. Surprise, surprise -- the new tactic worked beautifully. The Retreating Box finished the season with a higher breakup percentage (28.34%) and a lower ZEFR Against rate (44.39%) than any forecheck not named the Tandem Pressure.
The truth is, beyond Couturier, I'm not sure the Flyers have the personnel to run an effective standard 1-3 formation that switches between the Passive 1-3 and the Same-Side Press. Especially now that Ron Hextall seems intent on decreasing Claude Giroux's role on the PK to save wear-and-tear, it seems to me that Philadelphia lacks the dynamic two-way forwards up front necessary to extract the most value out of that strategy. Therefore, the Retreating Box appears to be the best way (for now) to proceed.
One positive of the Retreating Box is that it can be used in tandem with the Passive 1-3, especially when opponents utilize drop passes. As noted on Monday, one way to beat the Retreating Box is move just up to the extent of its range, drawing in the first line of penalty killers, before sending a drop pass to a trailing teammate.
In this scenario, the forward who did not commit to the pressure could still retreat back into the center spot at his own blue line as his fellow forward recovers, essentially creating an impromptu Passive 1-3. As we explored on Tuesday, the Passive 1-3 may allow a higher-than-ideal percentage of controlled entries, but its strength lies in its ability to attack the power play post-entry, as they are setting up in their formation. This gives the power play two options: attack the Retreating Box directly and try to beat it, or execute a drop pass and face the Passive 1-3 in the offensive zone.
As for the Same-Side Press, I don't believe it is an inherently flawed forechecking strategy. In fact, I suspect a forward corps filled with elite two-way guys with speed could execute it to perfection and overcome its weaknesses. Do the Flyers have that kind of forward depth right now? I have my doubts.
That's not to say it should be abandoned entirely. There will be matchups where the coaching staff believes that the Passive 1-3 should be the main look shown in the neutral zone, and in those cases, the Same-Side Press would be a great audible option. But based on the data, the Flyers might be better served with this forecheck solidly in the fourth position on the "Most Frequently Used" list.
Conclusion and Areas for Further Research
To all of those who took the time to read through this entire four-part project, my hope is that you gained a better understanding both of the penalty kill of the Philadelphia Flyers in 2015-16, and of PK tactics in general. Statistics remain essential in evaluating on-ice performance, but I believe that a greater knowledge of coaching strategies at the team level only serves to increase the depth of analysis.
But even a four-part series isn't enough time to analyze every element of the Flyers' penalty kill. There remain avenues for future analysis and research that I did not include in the series, and I would encourage any readers with interest in this topic to explore these areas.
To start, this series focused on team-level tactics and statistics. Adding in player-level analysis -- from individual microstatistics such as Breakups Created to on-ice metrics determining whether certain tactics are used more or less often with specific players on the ice -- could be a massive aid to teams looking to optimize their personnel. This is an avenue I hope to have time to explore in the coming weeks, but it involves additional manual tracking and data scraping capabilities that I was not able to complete before publication. I imagine the results could be eye-opening.
Second, an interesting quirk in the data appeared when tracking the Flyers' outcomes in terms of entry types. While their average number of unblocked shots allowed on dump-ins matched up with Corey's data from the All Three Zones project, the Flyers did a much worse job than expected at shot prevention on carry-ins, and a much better job than expected against defensive zone faceoffs.
This could just be noise, so it would be intriguing to check if shot prevention by entry type at the team level shows signs of being repeatable in Corey's dataset. In addition, tape study could show if the Flyers' strategies for shot prevention against carry-ins have some sort of flaw, and if their post-faceoff strategies are particularly effective for a specific reason.
Finally, a league-wide tracking project of each forecheck across the NHL would be invaluable in helping us to understand which ones might be "the best" in preventing entries and limiting ZEFR rates. Merely studying the Flyers leaves too many variables open (personnel, tactical unknowns) but a larger project could give us insight into which forecheck types are generally more effective than others over large samples and distinct personnel sets.
In any case, I hope that the information that was presented in this series helped to provide a clearer picture of the Philadelphia Flyers' penalty kill. And in a little over a month's time, we'll see if Ian Laperriere and Dave Hakstol largely stick with their 2015-16 blueprint, or adjust in the hopes of better outcomes. Armed with this data, we'll now be there every step of the way to keep track of their progress.