Getting analytical to break down the Panthers - Islanders series
Looking at the advanced metrics to give insight into the Cats-Isles match-up
During the course of this season I have spent some time being mighty critical of some pundits in the hockey press who jumped to early conclusions based on a very limited number of advanced statistical categories. What I am referring to, of course, are those who jumped to the conclusion that the Florida Panthers were a flash in the pan based on poor shot-differential results, reflected in Corsi. My response has always been that, while useful, Corsi is something that (in today's NHL) must be used in conjunction with a number of other metrics and observations to truly reflect a team's success and failure.
In some small part, vindication came the Cats way, as analytics devotees were forced to dig deeper this season to figure out what was happening in the NHL. In a massively entertaining year where playoff races were not decided until the last game of the season, the vaunted Corsi-For % statistic had some difficulty in the predictive area, as Eastern Conference darlings Washington finished 14th in the league in that stat. Detroit, who struggled to make the playoffs, and ultimately backed in, finished 9th, Montreal 10th, Carolina 11th, Winnipeg 12th, and Toronto 13th, while Florida won the Atlantic Division. The experts were forced to start looking for other answers, and many of them came (begrudgingly) to the conclusion that some mia culpa's were warranted.
What truly happened is that the league-wide parity tightened Corsi-For % up. The league best (in that stat) Kings rang up a 56.4 Corsi-For %, while 2nd place Pittsburgh finished the regular season at 52.7% The Panthers finished 20th with a 48.7%, only 4% lower than the Penguins, yet a whopping 18 ranks lower on the chart. It was a tight year, but within that parity, many teams started playing systems that were less concerned about pure numbers of shots, and more concerned about scoring chances. This season showed that to really understand who was good, and who was not, an analyst was going to have to dig a lot deeper than Corsi and Fenwick.
Regardless, we will take a look at a range of numbers to see if there are any indications, on paper, as to who has an advantage in this series between the Panthers and the New York Islanders. Let's take a look at where the Isles and Cats landed in a variety of statistics to give us some predictive power over what we may see in the series.
Its still a valuable statistic, but its weakness is those who rely on it without inclusion of a variety of other metrics. Nonetheless, we will start here. The Islanders finished the regular season with a 49.5% Corsi-For %, while the Panthers finished with a 48.7% Give the Islanders a slight (less than 1%) advantage in the shot-differential department.
Shots-for per game
The Islanders were 18th in the NHL regular season in shots-for-per game this season, with 29.4. The Panthers finished 24th, with 28.8. Give the Islanders a slight (less than shot per game) advantage in the shots-for per game category.
Shots-Against per game
The Panthers finished 13th in the league in this defensive category, with an average of 29.5 shots-against per game. Once again showing some of the parity we saw this season, the Islanders finished 24th in the league in this category, with 30.4, almost one shot per game difference (yet 7 spots lower in rankings).
This one is important - it is heavy defense that wins playoff match-ups.
Staying on the defensive side of things, the Panthers were 5th in the league with a .917 team save-percentage. The Islanders finished 9th, with a .915 (a mere .002 difference).
Looking to the probable game one starters, Roberto Luongo posted a .922 save-percentage over 62 games played. Thomas Greiss played 41 games and posted a .925 save percentage. The two goalies posted almost identical goals-against-average, with Lu at 2.35 and Greiss at 2.36. Whether Greiss could have maintained that rate over a full season like Luongo played is the big, open question, and the 10,000 lb. elephant in the room.
Goals-Against per game
The Panthers were 7th in the league, with 2.44 goals-against per game, while the Islanders finished 11th with 2.57 goals-against per game. Once again, the metrics are tight, with the Cats holding a slight defensive edge.
Goals-For per game
The Panthers were 8th in the league, with 2.83 goals-for per game, while the Islanders were 11th with 2.77 goals-per game. This is somewhat non-reflective of of the shots-for totals, where the Isles took slightly more shots-per game. That means we need to look at shooting percentage.
On-Ice Shooting Percentage
Florida posted an 8.8% shooting percentage, while the Islanders posted an 8.1%. Obviously a tight differential, but despite the fewer shots-for per game by the Panthers, and more goals-for per game, it's not necessarily reflecting puck luck.
PDO is often referred to as a puck luck stat, and the Panthers are posting a 102.1 PDO, while the Islanders have a more average 100.7. This is where things really start to get interesting. PDO combines save percentage and shooting percentage, and being well above the mean suggests that a team is getting a lot of luck. But the issues seem more complicated this season, as Roberto Luongo posted numbers similar to last season, and not well above average. Nonetheless, the Panthers' team save percentage (as shown above) was one of the tops in the league. ESPN hockey analyst John Buccigross made a great deal of sense earlier this season when he tweeted that the Panthers were obviously playing great team defense because both of their goalies were posting top save percentages. There is some debate about how much defense can effect save percentage, but the War-On-Ice shot location charts showed that the Panthers were absolutely preventing high percentage scoring chances, and that stat has recently reflected that.
That said, it is also evident that the Panthers are not taking a lot of poor shots. Their shot selection has been focused on higher probability areas of the ice. So PDO normally reflects luck, but the Panthers do seem to be improving their statistical chances on both sides of the rink by playing the averages.
Where do the Islanders stand on that metric?
The shooting percentages between the two teams are tight (8.8 to 8.1%), as are the save percentages. The Panthers scored 232 goals this season, the Islanders 227. Roberto Luongo faced an average of 29.04 shots-against per game, while Thomas Greiss saw an average of 29.19.
I went back and tracked five late season games for the Islanders to see where the shots for and against were coming from (what some have called the "JC stats"). On April 9th in an overtime loss to Buffalo, the Isles gave up more than 25 shots in the house (high probability scoring area out from the net through the high slot). Of course, that was the game that had Rangers fans crying foul at the Isles trying to lose to face the Panthers.
On April 4th the Islanders beat Tampa Bay 5-2, and gave up 18 shots in the house, 11 of which their goaltender had to make saves on (unblocked, on net). On April 2nd in a 5-0 loss to Pittsburgh, they gave up 21 shots in the house, 13 of which had to be saved (or were not saved). On March 29th in a 2-1 win over Carolina, they let the 'Canes shoot 15 from the house, and 10 had to be saved. On March 21st, in a loss to the Flyers, the Islanders gave up 19 shots in the house, 13 of which turned into goals or saves.
Looking at games played by both teams since December 1st, 2015, the Islanders are giving up a lot more scoring chances against-per game than the Panthers are, and that is being reflected in the shot location charts. Florida is giving up anywhere from 15-18 scoring chances, while the Islanders are giving up in the range of 21.5 to 24 scoring chances per game in that time period.
The Islanders have the 4th ranked penalty kill in the NHL, with an 84.5% efficiency rate. Their kill will match a Panther power play that is 23rd in the league, clipping along at 16.9%. That does not bode well for the Panthers' man-advantage opportunities. The Cats aren't likely to get many power plays either, as the Isles take only an average of 8.76 penalty minutes per game, good for 25th (or 5th best) in the league.
Flip this around and there is not anything mind boggling either. The Islanders power play is 17th best in the league, with an 18.3% success rate. The Panthers penalty kill, however, is 24th in the league, with a 79.5% success rate. The two teams are tight in this category, with the Islanders holding a small edge. The Panthers took the 10th most penalty minutes in the league, with 849, an average of 10.35 minutes per game (10th in the league).
There is no question that the Panthers need to stay out of the box.
In three games this season, you get what you'd expect with teams matched this closely in the metrics, even games. On November 27, 2015, Florida beat the Islanders 3-2 in a shootout, in a game that the Islanders out-possessed the Cats in. The Islanders also dominated the house portions of the ice. Of course, that was was not during a good stretch of play for Florida The Islanders won more than 50% of face-offs, and shot 8.7%, or close to their average on the year. The Panthers came away with the shootout victory.
On December 15, 2015, the Cats beat the Isles 5-1, but the Isles still dominated possession, but only shot 4.2%, which was a testament to the Panthers (at that time) new defensive system. It was clear on the shot location chart that the Cats were now using the system they would follow for the rest of the season. Neither team had many opportunities in the house areas of the ice. The Islanders won 51.3% of face-offs.
On March 14, 2016, the Panthers suffered a 3-2 loss, in a game they laid an egg in. The Cats ripped off a series of wins following that game, so it is hard to take much away from it. Nonetheless, both teams loaded up on chances in the house, with the Islanders getting a corresponding 12% shooting rate (higher probability areas of the ice). Florida won the face-off battle in that game.
Once again, we see two rather evenly matched teams.
Where a difference does start to appear is when we look at depth. The Islanders had four players score twenty or more goals this season, one of whom scored 20, and two who had 25 or more. The Panthers had five players score 20 or more goals, one of whom scored 20, and three who had 25 or more.
The Islanders had three players with 50 or more points this season, while the Panthers had six. The Cats also had far more offensive production from their defense.
It appears that just as predicted, depth and injuries will be the difference-makers in this series, with scoring chances as the biggest difference between the two teams....but keep an eye on those special teams.