With the recent popularity in the United States surrounding the World Cup, many people have brought up how it seems like the better team does not always win. Well as hockey fans, and certainly as Devils fans, we have known this to be true for a while. Seemingly on many nights, the New Jersey Devils would dominate the possession game only to find a way to lose 2-1. This offseason, the team has looked to address this issue by going out and signing forwards who can help to score goals. Hopefully, the additions of Mike Cammalleri and Marty Havlat will help with that. Instead of looking at scoring (or the lack thereof), however, another interesting angle on the team has to do with luck. Was our favorite team simply unlucky last year, with the dice weighted against them, or was luck not a major factor? Let's see what we can uncover.
How to Track Luck Using Statistics
While it may seem like luck is a subjective quantity that cannot really be quantified, in fact there is data out there that can make it somewhat more tangible. Some of the best information on the impact of luck can be found in the book Hockey Abstract, written by Rob Vollman (Twitter: @robvollmanNHL). There are other studies out there that can be found for free if you google something along the lines of "luck statistics in hockey" or "PDO in hockey," but I feel that the chapter dedicated to luck in Hockey Abstract is written well and makes understanding the statistics and impact of luck quite easy.
In the book, Vollman notes that there are 5 major, measurable areas in hockey where luck can have a definitive impact. These are:
-Shooting and Save Percentages (PDO)
-Special Teams (STI)
-One-Goal Game Record
Before delving into these, it must first be noted that not all of these are strictly luck-based. A team may have a great save percentage because they have a great goalie, and may have a great STI because they have an excellent power play or a shutdown penalty kill. However, there is also some part of each of these that has luck involved in them, making them useful for this study.
The first of these is the combination of shooting and save percentages. Sometimes, bounces and other fluke plays can influence these numbers, so not all of it is strictly skill. The easy-to-understand stat which breaks down the luck within these numbers is known as PDO. To calculate PDO, simply add together a team’s shooting percentage and save percentage. If the number is over 1000, the team was lucky. Last season, according to Extra Skater, the Devils had a team PDO of 996, with an approximate shooting percentage of 9.0% and a save percentage of 90.7%. This is very close to the break-even 1000. However, it should also be noted that if you only look at 5 versus 5 time-on-ice, the team’s PDO drops to an unlucky 986 thanks to a rather poor 7.1% shooting percentage. So clearly, the team was unlucky in 5v5 play, but made up for that luck in special teams situations.
The next area is indeed the special teams index. Since hockey is generally not played on a power play or penalty kill, their respective percentages can be influenced by luck. A team can get a fluke goal here or there which can boost up their power play percentage, despite having what looks like a pretty poor attack with the man advantage. The stat dedicated to this is the special teams index, or STI. It works similarly to PDO: add up a team’s power play and penalty kill percentages. If the number is over 100, they are lucky. For the Devils, according to nhl.com, the team had a power play % of 19.5 and a penalty kill % of 86.4. This creates an STI of 105.9, which is quite high. The team definitely was lucky on special teams.
Next up is injuries. While some will point to players being injury prone or to teams having better conditioning as to why injuries occur, in the end luck is a major driving force behind them. The stat to assess a team’s luck based on injuries is the CHIP stat, or the Cap Hit of Injured Players. This stat works better than to simply tally up the total numbers of games players lose to injury, as some players are more important to teams and have bigger impacts when injured. Thanks to the Springing Malik blog, there is a nice and easy chart available to see the effect of the CHIP stat this past season. The Devils had a CHIP of $8.345 million, which made them the 11th unluckiest team in the NHL in this regard. That is not incredibly unlucky, but considering that 19 teams were luckier than them with injuries, it was not exactly lucky either.
Fourth is the team’s post-regulation record. Obviously, shootouts are incredibly based on luck, and since an overtime only lasts for 5 minutes, many overtime game endings can be luck-based as well. Therefore, a team’s record in post-regulation games is largely luck-based. A team with neutral luck in these games should end up at around a .500 win percentage. New Jersey’s record, obviously, was miserable. The team, according to sportingcharts.com, was a successful 9-5 in the 5 minute overtime, but as we all know, went a putrid 0-13 in the shootout. This puts the team’s post-regulation record at 9-18, which equates to a very unlucky .333 win percentage.
Finally we have the one-goal game record. As I mentioned at the start of the article, a team can dominate in possession and control the flow of play, but lose 1-0 or 2-1 in regulation thanks to unlucky bounces. New Jersey has had its share of heartbreaks this way. Therefore, one-goal games can be decided by luck many times, especially if that game winning goal was fluky. Neutral luck would give a team around a .500 win percentage in one goal games that end in regulation. For the Devils, they played in 23 one-goal games that ended in regulation. Their record in these games was 13-10, which equates to a .565 winning percentage. This means that the team had a little luck in one goal games, although it was minimal. (Note: Combining post-regulation and one-goal regulation games, the Devils played in 50 games that were decided by only one goal. That accounts for over 60% of the entire season. Clearly, the team’s low event game strategy lends itself to a multitude of close games. I guess that is where many of my gray hairs come from…)
So, Was New Jersey Lucky or Unlucky?
To break it down again quickly, here is how the Devils performed in all 5 luck categories:
-PDO of 996 is neither lucky nor unlucky, but a 5v5 PDO of 986 is definitely unlucky.
-STI of 105.9 means the team was very lucky in special teams play.
-CHIP of $8.345 million is somewhat unlucky for injuries, but not overly so.
-Post-regulation record of 9-18 is very unlucky, thanks to the 0-13 shootout record.
-One-goal regulation record of 13-10 is somewhat lucky, but not overly so.
So we have three of the 5 statistics where the team was unlucky, with only two statistics that show some luck in the team’s favor. Again, it should be noted that these statistics are not entirely based on luck. A high STI can be partially attributed to a real strong penalty kill, which the Devils certainly do have. Also, one can possibly argue that the Devils should be better in one-goal games, considering they play so many. Following this logic, their record in them should be better than .500. In truth, there is no ideal way to perfectly calculate the impact luck had on a team’s performance. This is just one way that attempts to quantify it using different statistics that all have some luck involved in them.
To help showcase luck, Rob Vollman actually posts a chart online that depicts all 5 luck-based stats. It is free to look at, and can be found here. Looking at it initially, the Devils were indeed unlucky. The chart has NJ as the 5th unluckiest team last season. However, the real interesting thing about this chart is that you can weigh each one of the 5 stats differently, with either a full, heavy, medium, light, or ignore weight. Vollman initially has CHIP and OT-Shootout record with a full weight, PDO and One Goal Game record with a medium weight, and STI with a light weight attached to it. If you equalize all of the stats, the Devils become the 9th unluckiest team in hockey. Look at it yourself, and you can come to your own determination of how unlucky our favorite team really was.
In the end, it is pretty clear that the New Jersey Devils were an unlucky team this past season. While most any fan would have likely said that from the get go, the numbers do actually prove that sentiment to be largely true. The team lost Ryane Clowe for a significant portion of the season, as well as spending a decent amount of time without the likes of Patrik Elias and Damien Brunner. Also, the team’s shootout record was as unlucky as it gets, and if the team actually had neutral luck in the shootout it would have made the playoffs. While the team was lucky on special teams, with what I believe to be a somewhat bloated power play percentage, the luck gained there did not make up for the unluckiness seen elsewhere. Hopefully, if the team finds some improved luck next season, along with improved scoring and goaltending, it can find its way into the playoffs.
Now what do you think? Do you agree with the statistics that prove the Devils were unlucky? Or do you think that this is all smoke and mirrors, and luck cannot really be quantified? Is there any other way to determine how lucky or unlucky a team was? Or should we not discuss luck at all, but instead only look at a team’s skill and performance? Please leave your comments in the section below, and thanks for reading.