/cdn.vox-cdn.com/uploads/chorus_image/image/62956493/1085263936.jpg.0.jpg)
The first week of November we examined the Islanders’ 8-4-1 opening, comparing it to the hot starts of the Colorado Avalanche, New Jersey Devils, and Vegas Golden Knights of the 2017-18 season.
At the time, the Isles’ goaltending appeared unsustainable (.963 Fenwick save % at 5v5) and the shot-share was discouraging (44.2%, adjusted) even compared to COL/NJD/VGK, who each started below 50% in their first three weeks.
The most significant reason for hope: the Isles managed to squirrel away 17 points in 13 games while spending 8-of-13 on the road against tough competition, facing Nashville and San Jose twice each, while winning all six against Metro foes, including two-a-piece vs Pittsburgh and Carolina.
:no_upscale()/cdn.vox-cdn.com/uploads/chorus_asset/file/13719475/Screen_Shot_2018_11_04_at_7.14.37_PM.png)
Today we will consider how the Islanders have performed in November, December, and January compared to the trio of surprise teams last season. Note that a handful of games overlap the stats table above, such as NYI’s home wins against NJD and PIT the first week of November.
Though many hockey experts north of the border held onto the belief that “Since November First” stats are flawed for examining the play of the Islanders, I think we can all finally agree that this Islanders team since 11/1 is much different than they were the first three weeks of October, as the players adapted to the 5v5 system of new head coach Barry Trotz.
If this looks like a lot of information at once, it is! We’ll break down the table into smaller portions (more readable on phones) as we go, so stick with me here. The rankings below are relative to their respective seasons for the three-month period: 1st through 31st in the NHL for each stat.
As we can see, the Isles’ 23-11-4 record is far superior to NJD’s 17-14-8. It is better than COL’s 21-13-4 run, but not quite as impressive as VGK’s 25-9-4 campaign for the corresponding three months last season.
:no_upscale()/cdn.vox-cdn.com/uploads/chorus_asset/file/13719550/Screen_Shot_2019_01_27_at_5.32.38a_PM.jpg)
Remember that COL is blue, while NYI is the last column, in orange. We’ll focus mostly on the rankings (in bold).
5v5 Play
We can see, below, that on the surface the Islanders appear to have a stronger 5v5 game than any of COL/NJD/VGK from last season, for these past three months. They rank 5th for both expected goals-for* and expected goals-against rates. None of the other three teams ranked in the top seven for either stat.
* Expected goals stats consider the quantity of shots as well as quality measurements, such as shot location and shot type.
Colorado ranked 2nd for goals-for rate, but 28th for expected. In other words, like the Islanders of October this season, many expected the Aves’ 5v5 goal rate to level off (if not crash down) the final 33 games of the season, after some “lucky shooting.” In fact, it did: the Aves finished 2.08 (27th) from February 1st onward, very near their expected stat of 2.19 (30th).
New Jersey were a respectable 13th for expected goals-against rate, but 27th for actual. Vegas were similar, to a lesser degree: 9th for expected, 17th for actual. In each case, this indicates goaltending was not great. In contrast, while the Islanders’ 5th place ranking for expected goals-against is strong, the actual ranking (2nd) is even better, thanks to excellent goaltending. (More on that later.)
The Islanders weren’t able to score as much as they would have liked, but the 16th best rate at 5v5 is not bad for a team focusing on playing a conservative style in today’s wide-open NHL.
The Takeaway: The Islanders appear the clear winner here, though Vegas’ numbers are solid. New Jersey’s underlying numbers are encouraging (despite losing 22 of 39 games), while Colorado’s are concerning (despite winning 21 of 38).
:no_upscale()/cdn.vox-cdn.com/uploads/chorus_asset/file/13719621/Screen_Shot_2019_01_27_at_5.44.33_PM.png)
In terms of catch-all stats for 5v5 play, below, Colorado struggled in terms of both shot-share and expected goals-for percentage, while New Jersey performed better for high-quality chances (as 52.3% x GF% indicates) than for all shots (48.9%). The Islanders outperformed New Jersey’s 2017-18 three-month stretch for each stat, while losing the penalties drawn/taken battle by a margin similar to that of the Devils.
When penalties are taken into consideration, it is very arguable that Vegas have better 5v5 stats than the Islanders do here. (It depends how much one weighs shot share, compared to expected goals.)
Keep in mind that these stats do not consider goaltending (or shooting) skill / luck.
The Islanders are 27th in the league for penalties drawn over this stretch. That is perhaps a consequence of playing their conservative 5v5 style. The Isles rarely throw numbers forward on the rush, though Mathew Barzal’s skating ability (combined with his other skills) certainly helps to draw penalties.
The Takeaway: Without considering goaltending or shooting, the Islanders and Golden Knights appear more solid at 5v5 over these three months than the Devils, while the Avalanche are clearly fourth best.
:no_upscale()/cdn.vox-cdn.com/uploads/chorus_asset/file/13719717/Screen_Shot_2019_01_27_at_5.44.56_PM.png)
The following 5v5 stats are all Fenwick-based. In other words, this takes into consideration all the shot attempts (goals, saves, and misses) except shots that are blocked by opposing skaters before they reach the goalie. This is a useful stat for those who view shot-blocking as a helpful skill.
All four teams are near-average (13th through 19th) in terms of expected save percentage, meaning in terms of the quality of shot their respective goalies didn’t face particularly difficult nor particularly easy shots, on average.
Remember that the Islanders were 5th for expected goals-against rate. (The 3rd table in this article.) Since they are 19th for expected save percentage, we can deduce that Isles likely limited expected goals by keeping the shot total low, since quality is near-average.... And it is: 3rd lowest 5v5 unblocked shot rate (38.4) over the past three months, only behind this season’s SJS (37.3) and STL (38.2).
While Vegas’ and Colorado’s actual save percentage largely lines up with the expected rankings, New Jersey’s (28th, compared to 13th) is disappointing. The Islanders’ (2nd, compared to 19th) is remarkably good.
This isn’t the surreal .963 (Fenwick-based) save percentage from the first 13 games, but the .954 effort by Robin Lehner and Thomas Greiss provided a huge advantage at 5v5 for Isles—an already solid 5v5 team— compared to this season’s Metro rivals Pittsburgh (.944) Washington (.943), Columbus (.940), Carolina (.939), who are each ranked between 10th and 19th in the stat over the last three months.
Just as remarkably, the Islanders are middle-of-the pack in actual shooting percentage at 5v5, but they are 1st for expected shooting percentage, meaning that opposing goalies are facing shots that are theoretically the most difficult to save in the league (on average), at 5v5.
It has been noted that the Isles do not have the shooting talent of Toronto or Pittsburgh or Tampa Bay, but (cross-referencing with Natural Stat Trick now) the Islanders are 2nd in the league (to CAR) over the past three months in terms of 5v5 high-danger shot attempts rate, while they are a pedestrian 13th in terms of the rate of all attempts. (Furthermore, Isles’ “high-danger” shooting percentage is 24th in the NHL over this stretch.)
If the Isles can continue to take a high rate of their shots from the crease and lower slot area at 5v5, they may be able to creep into the top-10 for actual shooting percentage at 5v5, even without exemplary shooting skill, with a little more shooting luck.
The Takeaway: Even if the goaltending continues to level off for Isles, their shooting percentage may be more likely to increase than decrease, from 12th in the NHL over this past stretch. (This is a potential pitfall of glossing over PDO stats and coming to a conclusion without digging deeper into shot-quality indications: not all styles of play forecast the same shooting percentage.)
We do need to remember that goaltending is very volatile, though. It is possible the Isles’ goaltending falls into the bottom-third of the league the remaining 30 or so games. (Not likely, but significantly possible.)
:no_upscale()/cdn.vox-cdn.com/uploads/chorus_asset/file/13719848/Screen_Shot_2019_01_27_at_5.45.20_PM.png)
5v4 Power Play
While the Islanders often look like a mess at 5v4, the underlying stats are actually encouraging: 6th for expected goals rate, 2nd for expected goals-against rate, and 1st (just like 5v5) for expected shooting percentage of unblocked shots.
My theory is that the Islanders’ power play has been both (a) unlucky at times and (b) too predictable. I know league-wide teams have been criticized for a formulaic copy-cat approach to the 5v4 power play, but the Islanders’ power play in particular appears stagnant. (I bet some of you have faced more dynamic power plays in your latest bubble hockey tournament.)
While the Islanders are often shooting from the crease and lower slot area when they do release the puck at 5v4 (they are 29th for shot attempt rate the past three months, but 8th for high-danger attempts, according to Natural Stat Trick), it often is too predictable, as the goalie is already set.
Anders Lee and Barzal are threats whenever they have the puck, but neither Josh Bailey nor Brock Nelson has been lethal from the right wing spot. Furthermore, the Isles have difficulty finding loose rebounds, as a team, after Lee creates net-front chaos.
The Islanders have underlying (expected) numbers similar to Vegas, below, but their scoring rate has been the least successful of these four surprise teams.
The Takeaway: While the Islanders have significant power play weapons, it may make sense to prioritize adding a power play specialist— particularly a left-handed sniper who can combine with Barzal and Lee— over aiming to improve an already-solid team at 5v5.
If the Islanders play well the next month and GM Lou Lamoriello decides to add a rental, helping improve special teams may be the way to go.
:no_upscale()/cdn.vox-cdn.com/uploads/chorus_asset/file/13720029/Screen_Shot_2019_01_27_at_5.45.46_PM.png)
4v5 Penalty Kill
I have held off trying to analyze the Islanders’ penalty kill with stats— largely because it is difficult to compare players of the same team, since the PK1 skaters start with a defensive zone draw, typically against the best power play opponents, who are often rested.
It is safe to say now: the penalty kill has been terrible. While Valtteri Filppula and Leo Komarov perhaps helped to stabilize the 4v5 penalty kill in October, it is painfully obvious that the duo simply is not mobile enough to skate from point ‘A’ to point ‘B’ as a top-PK tandem.
Filppula and Komarov can be the most positionally-sound penalty killers in the league (and most shifts their positioning looks great), but in today’s NHL the PK forwards need to be able to limit time-space and win some loose pucks in all areas of the defensive zone.
While Filppula and Komarov help to clamp down around the slot area at times, the Islanders are still 27th for goals-against rate the past three months (29th expected) despite good underlying numbers for Lehner at 4v5.
Furthermore, the Isles are 31st in the league for expected goals-for on the PK over this stretch. Opponents need not be concerned about a counter-attack on a regular basis.
Of the four surprise teams, below, Isles are by far the worst in terms of 4v5 PK stats.
The Takeaway: The Isles should strongly consider using Barzal as a PK2 or PK3 forward. Not only is his skating a huge weapon for winning loose pucks and attacking— and likely drawing some hooking penalties in the process— but the first time the Islanders clear the puck (even after 20 or 30 seconds) the opposition need to think, “Is Barzal stepping onto the ice?” Just that momentary distraction alone may be enough for Isles to prevent an entry, even if Barzal remains on the bench. (I realize his defensive awareness is not always the best.)
At the very least, the Isles should consider putting Barzal out there the last 20 or 30 seconds of the penalty kill, when the opportunity arises. Barzal played a total of 54 seconds of 4v5 the past 38 games.
If Isles do not intend to try Barzal on the penalty kill, perhaps they can target a PK specialist at the deadline. Additionally, we may see improvement from the relatively inexperienced Pulock and Mayfield on defense for 4v5.
:no_upscale()/cdn.vox-cdn.com/uploads/chorus_asset/file/13720100/Screen_Shot_2019_01_27_at_5.46.16_PM.png)
The main takeaway: While the Islanders’ 5v5 play and goaltending have been far superior to New Jersey and Colorado of this same three-month stretch last season, the special teams need to improve if the Isles hopes to make noise in the playoffs, as the Golden Knights did a year ago.
This is a good hockey team, and the Islanders have room to improve.