A theory of why Dave Hakstol keeps starting Nick Schultz and benches Evgeny Medvedev

At the start of 2015-16 season, it was obvious that the Philadelphia Flyers‘ primary weakness was on the blueline. It was noted in every season preview in every hockey magazine and website, and was essentially acknowledged by the organization when they drafted defenseman Ivan Provorov with the seventh overall pick in the 2015 NHL draft despite already owning the rights to four other top blueline prospects.

Flyers general manager Ron Hextall clearly understood the need for more talent on the back end. But Hextall still planned for specific players under contract to play major roles on the defense this season. Mark Streit and Michael Del Zotto were obvious – Streit led the team in ice time last year, and Del Zotto was signed to a two-year, $7.75 million extension in the offseason. They were lineup locks before camp even began.

Two others appeared destined for full-time status. Nick Schultz was also a Hextall signing, first to a low-risk one-year deal in 2014, and then a two-year, $4.5 million extension in February. Evgeny Medvedev, on the other hand, was maybe the general manager’s biggest free agent pickup in 2015. The Flyers enticed the 33-year old Russian to leave the KHL and relocate to North America with a one-year contract worth $3.0 million.

Radko Gudas, Luke Schenn, Brandon Manning and even Andrew MacDonald were not afforded the same roster security. All four have been scratched or loaned to the Lehigh Valley Phantoms in the season’s first two months. But one of the four “locks” has surprisingly joined them out of the lineup – Medvedev has been designated a healthy scratch by coach Dave Hakstol seven times this season and only recently was able to earn his way back onto the ice.

While Medvedev fell out of favor, Nick Schultz’s grip on a roster spot remained strong. This was despite minimal (even for Schultz) offensive production and cratering even strength puck possession statistics. His 5-on-5 on-ice shot attempt differential (Corsi For) was a defense-worst 43.90% entering Tuesday’s game against the Ottawa Senators.

As a result, the critics are growing louder. The advanced statistics crowd cannot understand why Hakstol would continue to start Schultz while relegating a higher-paid, more physically gifted, and statistically-superior defenseman to the bench. So why has Schultz avoided the press box while Medvedev was almost ready to reserve a personal seat? The answer may lie in the defensive zone play of the two blueliners.

Understanding neutral zone play

In my recent microstat articles, I’ve included a metric called “Zone Entry For percentage” as a way to track a player’s neutral zone effectiveness. The statistic is relatively simple – it counts all zone entries that occurred while a specific player was on the ice.

For example, let’s envision a scenario where Claude Giroux wins the opening faceoff and dumps the puck into the opponent’s zone. Then, the opposing team takes control of the puck, exits the defensive zone, and carries the puck through the neutral zone and into the Flyers’ end. Steve Mason smothers the ensuing shot, ending the shift.

Claude Giroux’s Zone Entry For percentage on that shift would be 50 percent, since the Flyers entered the offensive zone once, and so did the opponent. The statistic primarily measures raw zone entries, in order to understand which players are helping Philadelphia to move the puck in the right direction.

But there’s a problem with the metric – not all zone entries are created equal.

Through the great work of former BSH writer Eric Tulsky, we learned that controlled zone entries are far more valuable than uncontrolled entries (dump-ins). And Corey Sznajder’s ‘All Three Zones’ project in 2013-14 gave us exact metrics over a large sample size. Sznajder determined that over time, each controlled zone entry produces an average of 0.66 unblocked shot attempts, while an uncontrolled entry produces around 0.29 unblocked attempts.

Controlled entries are clearly more valuable than dump-ins. So in order to get an accurate view of neutral zone play, it makes sense to give more weight to controlled entries than uncontrolled entries, which Zone Entry For fails to do.

Enter Neutral Zone Score, an improved metric developed by Tulsky. The statistic was introduced in the Sloan Sports Conference presentation Using Zone Entry Data To Separate Offensive, Neutral, And Defensive Zone Performance (Tulsky, Detwelier, Sznajder and Spencer). Japers Rink did a great job of presenting the statistic in less technical terms last year, as well.

Now that we have a statistic that best measures neutral zone play, let’s look at Medvedev and Schultz’s respective scores.

The neutral zone statistics, even after accounting for zone entry type, do not support Nick Schultz’s cause. Medvedev’s Neutral Zone Score drops a bit from his Zone Entry For percentage, mainly due to a high amount of controlled entries against that have occurred on his watch. But it’s not enough to drop him anywhere near Schultz.

Clearly, it’s not neutral zone play that Hakstol likes about the NHL veteran.

Schultz and Medvedev in the defensive zone

When Evgeny Medvedev earned his way back into the lineup for last Friday’s game against the Nashville Predators, Dave Hakstol made a telling comment that hinted at a key reason for the extended scratching of the Russian defenseman.

Hakstol said Medvedev (paired with Ghost) needs to focus on D play in own zone

— Bill Meltzer (@billmeltzer) November 27, 2015

Throughout November, the coach had been tight-lipped on the reasoning behind Medvedev’s removal from the lineup. Now Hakstol was giving the player clear direction on what he felt needed to be improved in order for the defenseman to stay on the ice.

But was it accurate? Had Medvedev’s defensive zone play been so poor to justify the scratching? In order to answer that question, it’s necessary to find a statistic that accurately measures defensive zone play.

Plus/minus was used in the past to judge defense – a high “minus” rating would imply weak coverage. But using on-ice goal statistics to judge play is flawed, as goals happen infrequently. Good (or bad) luck can have a dramatic impact on a player’s plus/minus, making it a fairly unreliable stat.

The analytics community has turned to shot attempts as the best way to judge players, because they increase the sample size dramatically. If a player is on the ice for lots of shot attempts against, the theory is that he’s probably poor defensively, since his team is consistently pinned in their own zone when he plays.

But even looking at shot attempts has flaws, particularly if the goal is to specifically measure defensive zone play. A player could be above-average at preventing opponents from shooting the puck once they enter the offensive zone, but is so poor in the neutral zone that the opponent is racking up shots due to lots of zone entries. The coach sees the strong play in the defensive zone and is satisfied with the player, but is missing the mass amount of zone entries that occur when the player is on the ice.

Luckily, by tracking entries, we can isolate defensive zone play from neutral zone play. By counting the total number of controlled entries, uncontrolled entries, and defensive zone faceoffs (also worth 0.29 unblocked shot attempts) that occurred when a player was on the ice, we can determine what his expected on-ice shot attempt numbers should be, assuming that opponents generated the league-average production associated with each entry type.

If opponents generate more shot attempts than expected, it can be inferred that the player did a poor job of shot suppression in the defensive zone. If opponents generate less than expected, his defensive zone shot suppression (to this point) has been strong.

Let’s again analyze Medvedev and Schultz, this time looking solely at their defensive zone play.

We may have our answer. While Medvedev dominates the neutral zone, he’s allowed more shot attempts than would be expected considering the way opponents entered the offensive zone. Schultz is the opposite – his neutral zone play has been poor, but once teams have set up on the attack, he’s been able to help the Flyers suppress unblocked shot attempts.

This does fit with the prevailing narrative surrounding Nick Schultz – a reliable, stay-at-home defenseman who can be trusted in his own zone. Combined with that narrative, these statistics may help to explain how Schultz is still passing his coach’s eye test despite poor overall possession statistics. They also explain why the coach became so frustrated with Medvedev even though his neutral zone play was stellar.

Does the reasoning make sense?

If we assume that Dave Hakstol is keeping Nick Schultz in the lineup due to strong play in the defensive zone, the next question is whether that is a justifiable reason. Schultz’s defensive zone play has been strong so far, but could it just be a fluke? Is shot suppression in the defensive zone an actual skill, or just random variance?

The jury is still out on that question. In Tulsky’s original work on zone entries, his data implied that only neutral zone play was repeatable – offensive and defensive zone success looked largely random and not tied to talent. But recent work by Garik16 of Hockey-Graphs concluded that teams are able to sustain above-average play in the offensive and defensive zones, though it’s far less repeatable for players.

More work does need to be done on this topic before making concrete statements on the sustainability of strong play in each individual zone. But in answering the Schultz question, it’s largely irrelevant for this reason: even if Nick Schultz has an ability to reproduce above-average shot suppression in the defensive zone, his limitations in the other two zones have thus far negated the positive impact it provides.

Schultz’ 46.12% Neutral Zone Score is by far the worst among Flyers defensemen (second-worst is Mark Streit at 48.14%). In addition, Philadelphia has struggled mightily in the offensive zone with Schultz on the ice as well. Based on Schultz’s on-ice entries, his Fenwick For should be 214.26. Instead, the Flyers have generated only 175 unblocked shots with Schultz on the ice, a 18.32% drop from the expected total.

Any gains that Schultz makes in the defensive zone, he gives back everywhere else.

Conclusion

Nick Schultz avoided the press box in November, while fellow opening night starter Evgeny Medvedev was a healthy scratch for seven games. Schultz remained in the lineup despite posting poor possession statistics and delivering underwhelming play in the neutral zone.

Defensive zone play may be the reason why Schultz’s remains in Dave Hakstol’s good graces. When Nick Schultz has been on the ice this season, the Flyers are doing a better-than-expected job at suppressing shots, after accounting for the way opponents have entered the offensive zone. On the other hand, Evgeny Medvedev has struggled in that area, an assertion supported both by the stats and Dave Hakstol’s own comments.

If Schultz’s defensive zone play is sustainable, it would make him a good fit to play in an end-of-game, shutdown role, when the only real goal is to escape with a victory, not generate offense. It’s understandable that Hakstol would be willing to sacrifice some prowess in the other two zones in order to dress a player that he trusts with the game on the line.

But a late-game shutdown role might be the only one where Schultz can excel. His poor neutral zone and offensive zone statistics imply that he has been a liability in terms of helping the Flyers score goals. Schultz lags so far behind his teammates in two-thirds of the ice that any benefit gained from Schultz’s good defensive positional play is outweighed by his struggles everywhere else.

Dave Hakstol is clearly an intelligent coach, who does not make lineup decisions without reason. But in this case, it’s possible that Schultz’s solid play in his own end is masking struggles in the other zones. At this point, the Flyers would be best served to remove Schultz from the lineup and dress the best two-way defense possible, even if the team takes a slight hit in defensive zone coverage.

All statistics from War-On-Ice.com or manually tracked by Charlie O’Connor.