Philadelphia Flyers Penalty Kill Analysis, Part 3: Evaluating Defensive Zone Play

This is part three of a four-part series by Charlie O’Connor describing and evaluating the penalty kill of the Philadelphia Flyers during the 2015-16 season. 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 manual tracked forechecking metrics. Today, we put the spotlight on Philadelphia’s penalty killing in the defensive zone.

To evaluate a penalty kill, there are two distinct areas that must be measured. The first — yesterday’s task — was determining the PK’s ability to prevent teams from entering the zone and ever getting set up in formation in the first place. Theoretically, the perfect penalty kill would stop every single rush and in turn, never allow a shot on their goaltender at all.

But obviously, that’s not realistic. Even the best penalty kill forechecks will get beaten a fair amount, and the opposing power play will have opportunities to set up on the attack in the offensive zone. This is where defensive zone shot and goal suppression comes into play — the ability of the penalty kill to prevent opponents from generating tangible attack pressure that turns into goals.

Today, we’ll evaluate the nature of the Flyers’ play in the defensive zone, specifically the distinction between passive and aggressive tactics. We’ll also use manually-tracked shot data to determine which strategies were most successful. Finally, we’ll judge how Philadelphia performed in the defensive zone on the PK on the whole, and whether it was a strength or a weakness in 2015-16.

What tactics did the Flyers use in the defensive zone?

During their first round series with the Washington Capitals, the Flyers completed eschewed pressure-based tactics in the first three games. But after being gashed for eight goals on 17 opportunities, they shifted to a more aggressive style focused on attacking the forwards on the half boards and the defenseman up high.

Whether it was caused by the tactical shift or was merely the result of an inevitable regression, the Capitals did not score another power play goal the remainder of the series. The playoffs showed that Philadelphia is capable of using both the Passive Triangle +1 and the Czech Press tactics in the defensive zone, both outlined in Part One of the series. But which did they use more often this year?

To answer this question, I tracked every time an opposing power was able to successfully set up in the offensive zone, and noted which tactic was used by the Flyers’ penalty kill to combat it. The two primary strategies were the Passive Triangle +1 and the Czech Press, though on rare occasions they did flash the Low-High Press, which I outlined back in April. However, for ease of analysis, we’ll separate out the tactics into two buckets — Passive Triangle +1, and Pressure-Based (combining the prominent Czech Press and the Low-High Press).

It’s clear by looking at this data that the Flyers primarily played passive defense in their own zone while on the penalty kill. On only one of every four opportunities did Philadelphia choose to actively pressure the puck carrier on the perimeter using a Czech Press or Low-High Press strategy. 27.8 percent is high enough that opponents couldn’t rule the possibility of pressure out entirely, but it certainly wasn’t the Flyers’ first choice in the defensive zone.

Which tactics were most successful in limiting shots?

Philadelphia may have used the Passive Triangle +1 more often than pressure-based tactics, but that says nothing about the effectiveness of each. To answer that question, I also tracked the amount of shot attempts that occurred when the Flyers employed these defensive zone strategies.

This isn’t a perfect measure of defensive zone efficiency — it doesn’t account for the quality of shots, for example. But it does give us a general idea of the amount of attack that each Flyers tactic allowed to occur.

According to this data, the Flyers did a much better job of shot suppression across the board when putting some pressure on the perimeter of an opponent’s power play. Whether it be shots on goal, unblocked attempts (Fenwick) or total shot attempts (Corsi), pressure-based tactics win out.

However, it’s fair to note that this is far from conclusive evidence that pressure-based tactics are inherently better than passive structures. For starters, it’s difficult to tease out the difference between situations where the Flyers planned in advance to use pressure (like late in the Washington series) and instances when an opposing puck carrier fumbled the puck and inspired a Flyers penalty killer to attack. Pressure-based tactics may have worked in the latter scenario because the penalty kill chose the perfect time to use them, not because they are best in all situations.

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.

It’s also possible that the simplest answer is indeed the correct one, and the Flyers should attack the puck a lot more while in the defensive zone. As I’ll propose in Part Four, it’s at least worth trying.

Did the Flyers’ tactics change over the course of the season?

As we learned yesterday, Philadelphia adjusted their neutral zone forechecking strategies on the penalty kill as the year progressed, moving more towards a Retreating Box and cutting back on their usage of the Tandem Pressure and Same-Side Press. Let’s do a similar analysis with their defensive zone tactics. Did their ratio of Pressure-to-Passive change on a month-to-month basis?

The Flyers actually used less defensive zone pressure as the season progressed. During the 2015 calendar year, it was close to a 60/40 split, but during 2016 it was more like 80/20, a large discrepancy.

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.

What we see here are four months (November, February, March and April) of the penalty kill performing at a solid level from a shot and goal prevention standpoint, two months (December and January) when things were disastrous, and one month (October) when the outcomes were weird, as shot prevention was fine but goal prevention was terrible. Regardless, there’s no real correlation between “frequency of defensive zone pressure” and a good Flyers penalty kill. This disconnect will be addressed in Part Four.

We’ll come back to this month-by-month chart later. First, we’ll delve a bit deeper into defensive zone play, and then use that to better explain these monthly efficiency numbers.

Were the Flyers a bad defensive zone shot suppression team?

So far, we’ve looked at the tactics that Philadelphia used in the defensive zone on the PK, and also evaluated their efficiency in terms of shot suppression. However, we’ve yet to address the most important question — by the numbers, were the Flyers a “good” defensive zone shot suppression team?

The best way to answer this question is to look at their performance relative to the rest of the league. Unfortunately, we do not have league-wide data on the 2015-16 season, but thanks to Corey Sznajder, we do have it for the 2013-14 season.

Corey not only tracked entries for each NHL team at even strength, but also during 5-on-4 power plays. This suits our purposes perfectly. Using this data, we can determine what was the league average outcome for three entry types during 5-on-4 power plays — controlled entries, uncontrolled entries, and offensive zone faceoffs. Once we know how many shots (on average) can be expected on each entry, we can compare that to the metrics that the Flyers posted in 2015-16 to see if they performed above or below that expected average.

When facing controlled entries, the Flyers did a worse-than-expected job at suppressing shots. But against faceoffs, Philadelphia was significantly better than the 2013-14 league average, allowing just 0.39 unblocked attempts per defensive zone faceoff situation.

But this doesn’t answer whether the Flyers on the whole were a good defensive zone shot suppression team. To determine that, we’ll need to use a new method.

We’re really asking if Philadelphia prevented more shots than would be expected considering the amount and type of entries they allowed. Corey’s data gives us the roadmap on how to weight each entry, and through my manual tracking, we have the total number of controlled entries, uncontrolled entries, and defensive zone faceoffs allowed by the Flyers in 2015-16. If we take those total counts and weight them by the 2013-14 NHL average outcomes for each entry, we should be able to get a rough gauge of the amount of unblocked attempts the Flyers “should” have allowed if each entry produced a league-average outcome.

This “Expected Fenwick” total can then be compared to the Flyers actual 4-on-5 unblocked shot attempt (Fenwick) totals. If Philadelphia’s actual Fenwick is higher, then they did a below-average job at defensive zone shot suppression. If it’s lower, then we can say they were above-average in 2015-16.

So where do the Flyers end up?

This means that if the Flyers allowed the league average unblocked shot attempts on each 4-on-5 entry in the 80 games that were manually tracked, they would have allowed 495.18 total unblocked shot attempts. And how many did they actually allow?

495. Nope, that’s not a misprint.

The Flyers’ defensive zone shot suppression on the penalty kill in 2015-16 was neither bad nor great. It was instead almost perfectly, painfully average. Just armed with the Expected Fenwick formula derived from Corey’s data, we could have finished with the same result that took weeks to track manually.

While this tells us that Philadelphia’s defensive zone play isn’t a weakness, that doesn’t mean that it couldn’t use improvement. After all, no Flyers fan really wants to see the team hover around the 13th to 17th place in penalty kill efficiency — they want the return of the glory days when the team had one of the most effective units in hockey. This data makes it clear that, while not a disaster, there’s definitely room for defensive zone shot suppression improvement.

Back to those month-by-month metrics…

Now that we have a method to evaluate defensive zone play, and ways to evaluate neutral zone results from yesterday, we can try to build a narrative regarding the Flyers’ statistical performance at 4-on-5 last season.

Let’s look at four metrics on a month-by-month basis to see if they shed any light on the situation. First, we’ll check Entries per 60, which will show us if Philadelphia was getting overwhelmed by a large amount of raw entries in the neutral zone (or opponents were forcing lots of extra faceoffs). Next, we’ll check Fenwick per Entry, which will show us if the Flyers were getting bludgeoned in the defensive zone in a given month.

We’ll also look at Expected Fenwick per Entry, which lets us know if Philadelphia was allowing higher quality entry types to opponents during certain months. And finally, we’ll check save percentage, because maybe a bad month from a goalie was the main reason for poor goals against totals.

Now we have a clear picture. As previously noted, November, February, March and April were the best months for the Philadelphia penalty kill from a statistical standpoint. And in all four of these months, the Flyers performed respectably in every aspect of penalty killing — limiting raw entries, being stingy in the defensive zone, not allowing a disproportionate amount of controlled entries, and getting good goaltending.

In the three bad months, however, there was always one glaring issue.

October saw the penalty killers do their job fine, with the exception of the most important one — the goalie. If you remember, Steve Mason got off to a horrific start on the PK last season, which was the main driver behind that poor 0.816 team save percentage.

Then, in December, the penalty kill was seemingly allowing opponents to move into their zone at will. The defensive zone shot suppression was fine, and the goaltending was passable. But the Flyers were buried by raw volume of entries.

January’s biggest issue was defensive zone shot suppression. Considering the entries they conceded, Philadelphia was expected to allow 0.55 unblocked attempts per entry. Instead, opposing power plays racked up a whopping 0.69 attempts on their entries. Not particularly nice.

So it wasn’t one issue consistently rearing its ugly head in the bad months. Instead, a problem would pop up, then seemingly would get fixed, only to see a totally new issue appear weeks later. This helps to illuminate the difficult job of coaches — sometimes, the job can be like trying to plug holes in a boat in the middle of a lake. Once you have one hole plugged, a new one appears, and to survive you need to repair the new leak while still ensuring the old breach doesn’t rip back open.

Tomorrow, we’ll take that plunge and put on our coach’s hat. The goal? Recommend adjustments for the Philadelphia penalty kill so that it takes a measurable step forward in 2016-17.