Statistics are fascinating stuff…in the sports world. They have exposed utility and futility in players and systems that was previously hidden from all but the most professional observer, and even those paid observers often struggled to see what the numbers have revealed. Bar stool discussions about hockey often now include a smattering of Corsi or possession metrics, as fans embrace the new measures of their favorite team’s value. Especially here on LBC, we debate what the numbers may show vs. what the “eye-test” indicates about a player like Derek MacKenzie, for instance. Our resident metrics guru, Shane, has spent much of this summer providing us just this type of analysis to feast on, and his work has gone deep into more useful metrics than just Corsi or other such measures. Those metrics may prove useful predictive tools for what comes this season. I…. am not that guy. When the numbers start cracking, my brain runs, screaming bloody hell, in the other direction.
Yet, one thing has struck me this summer as I have been doing my lazy hockey reading and trying to ignore the fact that I am paying a cable tv bill for something I don’t turn on when hockey (or football) isn’t on, and that is this: the greatest metric of them all, and the one that cannot possibly be quantified, is luck. Now, before we get going here, let me acknowledge that “YES” hockey statistics DO try to account for luck, such as: PDO, a stat that ends up measuring “puck luck.” But I am talking about something different and more all-encompassing. Our own Shane O’Donnell, stats guru, has stated: “We see crazy things happen all of the time despite slim odds; there are forces at work in the universe that people just don’t have control over.” (you can find the excellent article this excerpt came from here)
I want to look at luck and other hard to quantify variables, and I am prepared to frighten you with this conclusion: an NHL season is a total crapshoot.
Lets start with this: the Eastern Conference is tightly packed with playoff contenders, and some good teams, at least as good as some teams that will make the playoffs, will miss the playoffs by razor thin margins again this season. This writer from the hockey writers.com nailed this fact here:
Whether or not [the Panthers] make the playoffs has more to do with the teams around them. The Eastern Conference, not just the Atlantic Division will be very tight this year. The Panthers, more than everyone else, will certainly need to have a good record against their divisional opponents.
How very true this is. The Panthers missed the playoffs with a final tally of 91 points last year. It took 98 points to get into the playoffs in the Eastern Conference in 2014-15. In 2013-14, a team in the East could’ve gotten in with 93 points, and back in 2011-12, it took only 92 points to make the postseason dance. How good, and lucky, the competition is will impact the Cats chances greatly. So who misses out this season, and what leads to that miss? There is more to that discussion than simply who has the “best” players.
Even at the top of the conference, questions abound: Will Tampa Bay’s kids remain on their tear or fall back into sophomore-like slumps? Can Montreal stay at the top with limited goal scoring, playing Carey Price 1000 games per season, and with questionable center depth? Will the Rangers be able to score more goals, and can the Swedish King maintain his health and play as he ages another year? Can Washington put it all together, or will they miss Joel Ward up front or Mike Green on defense? Will Pittsburgh score 2 million goals this season….but give up 2 million and one? Is a now healthy Columbus for real? Somebody explain the Islanders to me (I believed, dammit!). Should I listen to Boston’s management, that the team is better, or to every pundit, as well as my own brain telling me that team looks like a non-playoff contender. And is Detroit sliding or not? Is Ottawa finished with their run of improbable “holy craps”, and did Buffalo really get that much better? Questions, questions.
What then to make of our Panthers. Metrics will tell us a lot, but they will not be the whole story. The team got close last year, missing by seven points. The unquantified variables of luck had a major impact on the team and those seven points in ways that could not be predicted. Here then, a look at some of those variables headed into 2015-16:
It can be darn near impossible to predict whether an NHL rookie will produce or not:
A good NHL draft may yield one or two NHL players out of nine or more picks. That is success, and that is what an incredible shot in the dark most drafting is. Detroit, legendary for finding European talent in late draft rounds, does not manage to do so every draft year or in every round. Even the mighty Red Wings scouts fail on picks. Taking a look at several studies of the NHL draft, and how many players drafted end up playing an NHL game, or 100 NHL games, the numbers are striking. Josh Weissbock, at www.canucksarmy.com, has done some revealing studies of drafted CHL players and how many go on to play in the NHL. By his calculations, only 30.2% of CHL players play more than one AHL game, and only 8% of CHL players go on to become NHL regulars. Oh my…
Another study of the NHL draft from 1990-1999 determined that of 2,600 players drafted in that time-frame, only 19% of them played a minimum of 200 NHL games. That means that only 494 of those 2,600 players hit that mark, and more striking is that of those 494, 160 were first round picks. That same study found that in that 1990’s sample, a 1st round pick had a 63% chance of being a career NHL player, a 2nd rounder a 25% chance, and 3rd rounders and beyond only a 12% chance of becoming NHL regulars. Clearly, at a minimum, these numbers suggest that a team should expect their first round picks to find NHL success, and that it may be foolhardy to count too heavily on later round picks.
Adam Gretz, at www.CBSSports.com, had another interesting study of NHL draft success, in June. According to his study, the numbers of players drafted who appear in 100 or more NHL games breaks down as follows:
– Top 5 picks: 96.3%
– 6-10 picks: 78.1%
– Rest of 1st round picks: 63%
– 2nd Round: 31.1%
– 3rd Round: 27.9%
– 4th Round: 18.7%
– 5th Round: 14.2%
– 6th Round: 14.3%
– 7th – 9th Round: 11.6%
This matters to Panther fans, or rather, it should. Even without the rookies, there is youth throughout the roster. The Detroit Free Press took a look at the Panthers back on July 28th, and had this to say about some of the youngsters:
Outside of Edmonton, the team from Sunrise, Fla., has the best collection of up-and-coming star power. Ekblad is a 19-year-old defenseman who stands 6-feet-4 and tops 215 pounds. He just won the Calder Trophy as best rookie. He’s a franchise defenseman who turns heads every game. Ekblad, Huberdeau, Bjugstad, Barkov and Jussi Jokinen provide a well-fed offense.
The first variable that we are going to deal with going into, and during this season, is draft luck, or put another way: “Did we mine any other talent aside from those named by the Detroit Free Press that can be added to that bunch?” The team is going younger, and let me be perfectly clear that I 100% support that move (since we are using percentages of course). Between one and up to perhaps three spots on the final roster will be filled by NHL rookies. This could be Rocco Grimaldi, it could be Garrett Wilson, maybe Lawson Crouse, maybe Quinton Howden. Any of those players could be on the roster, and the big question is: are any of these players NHL caliber players? And more importantly, are any of these players going to be productive NHL players? Because, we really need one or two of them to be all-that…..
For Crouse, we have only potential to work from to answer that question. As a number 11 draft pick, Gretz’s study suggests there is a 63% chance of Crouse playing more than 100 games in the NHL, and an 88.6% chance he will appear in the NHL. Those are obviously good indicators that Crouse will be an NHL player, but here comes into play yet another variable that cannot be accounted for, and that is how Crouse is developed by the team.
Some players develop well at the faster paced NHL level, while others are best served with additional years in junior. That determination is very easy get wrong – due in large part to NHL contractual issues: a team gets a 9-game look at a player before they may be returned to juniors. That’s not a lot to work with, and if the team gets it wrong, the player is typically stuck at the NHL level for an entire season, as they cannot be sent to the AHL and the NHL club is loathe to pay a player on an entry level deal while he is playing for a junior team. So count on Crouse to make magic this early at your own risk, or be happily surprised if he turns into another Ekblad. This question will only be solved if he makes the team and then as he progresses.
Moving on to the other players we may see in rookie seasons this year, we have a few other factors to work with to try and predict how they will turn out at the NHL level. Grimaldi was a second round draft choice, meaning purely from the standpoint of draft position, and Gretz’s study of relative success derived from draft position, Rocco has a 31.1% chance of playing 100 or more NHL games, and a 65.7% chance of appearing in NHL games (which he has already accomplished). But Grimaldi played last season in the AHL, and Josh Weissbock, of the aforementioned Canucksarmy.com, has done a study of success of AHL players at the NHL level.
Weissbock has measured levels of success by position, age, and points-per-game at the AHL level. Grimaldi is 22-years-old and a forward. He played in 64 AHL games last season and put up 42 points for a .65 points-per-game average. For forwards aged 19-21, NHL success is likely if they produced .7 ppg at the AHL level, by Weissbock’s study. Rocco was 21-years-old for much of last season, so we use the 19-21 year old numbers. By that category, Rocco was slightly off pace in the likelihood of NHL success, but very close. Forwards at age 21 who score at .7 ppg rates in the AHL, according to the study, have a 22.22% chance of NHL success.
While that may seem like a small chance of success, things do get worse in the prediction department, as Quinton Howden played two seasons in the AHL in the 19-21 year old range. He played a total in those two seasons of 116 AHL games, and put up .49 ppg, well below the .7 ppg standard. Last season, at age 22-23, he played 33 games with 18 points, or .54 ppg. At age 22, Weissbock determined that an AHL forward would need to put up 1.0 ppg to have a 18.58% chance of NHL success.
Similarly, Garrett Wilson played the last two AHL seasons as a 22 and 23-year-old forward, meaning by the Weissbock analysis, the likelihood of NHL success would be likely if he were producing at one ppg. But Wilson was producing these last two seasons at a .47 ppg rate instead.
If you are interested in what a bona fide top-6 forward looks like in the AHL, look at Gustav Nyquist of the Red Wings. At age 25, he played in 15 AHL games and produced points at a rate of 1.4 ppg, at the age of 24 and 23 he was producing at 1.03 ppg. he has, of course, been equally effective at the NHL level.
The rate of success for CHL players who play in the AHL, to make it in the NHL is roughly 30% (again according to Weissbock). By the majority of these studies, things look rather grim for a rookie to step in and be a productive NHL player next season for the Panthers. Crouse and Grimaldi, based on studies of historical performance by similarly situated players would appear to have the best chance of being that productive NHL player. But that said, there is the “x-factor” of what roles players are assigned to on their AHL teams, whether their teams were good or bad, and whether the systems used by that team were meant to produce large offensive outputs, or prevent goals by taking fewer offensive risks. All of these factors are why it is typically very difficult to predict how a player will turn out at the NHL level.
Simply stated, if any of the four prospects makes the Panthers and becomes productive, the team will be more likely to succeed, while if these players fail to produce, the Cats will have a gaping hole to fill in the roster. We need a bit of the draft luck we salivated over in Shane’s summer series on “what could have been” draft picks. Chalk up our first unquantifiable variable that will remain an unknown until the season kicks off and gets rolling.
It is darn near impossible to determine when a veteran player will hit “the wall”:
Jaromir Jagr is a super hero. He has defied all conventional wisdom to this point as a 43-year-old who remains a highly effective NHL player. I am nearly the same age and wake up every day feeling as if I did an overnight triathlon. He is the epitome of a first ballot NHL Hall-of-Famer. Yet, he is also a statistical outlier. For most players, the early-30’s might as well be the early-60’s for the rest of us (career-wise). Effectiveness slides quickly for most players when they enter their 30’s, and the only question is how fast they fall. Olli Jokinen (as a well noted example) fell off the NHL planet last season.
Looking at EricT.’s fabulous little article on the effects of age on scoring, we can see that a player at the age of 29 retains approximately 90% of their even strength scoring rate, but that a mere two years later, at age 31 that drops to 80%, and at ages 32-33 that drops to 70%, and finally at age 35, drops to 60%. Panther fans are perhaps the most aware of any fan base in the NHL of this tendency, as we have had more than our share of older players signed to contracts who hit the proverbial wall. The names Ed Jovanovski and Steven Reinprecht are two names, among many, that standout on this front.
The trouble is trying to predict when the steepest part of the player’s slide is going to occur, and that is what leads to this unpredictable variable that can have a massive impact on the team. For the 2015-16 Panthers, these players will have at the very least a small question mark due to age:
Jaromir Jagr– age 43
Roberto Luongo– age 36
Willie Mitchell- age 38
Al Montoya– age 30
Brian Campbell– age 36
Shawn Thornton– age 38
Derek MacKenzie- age 34
Jussi Jokinen- age 32
If all goes right, and as is hoped for, all these 30-plus year-old players will have continued production for the team. The uncontrollable variable of suddenly feeling their age, or injury due to age, could also cripple the team if it’s to one of the more important veterans like Luongo, Jokinen, Jagr, Mitchell, or Campbell.
The goalie issue is, without question, the biggest, as Luongo is a 36-year-old future Hall-of-Famer who played in 61 games last season, and did deal with injuries. Historically speaking, four goalies have played 60 or more games as a 36-year old in the NHL: Tony Esposito (in 1979-80), Dominik Hasek (in 2001-02), Patrick Roy (in 2001-02), and Ed Belfour (in 2001-02). Two other goalies have played more than 55 games as 36-year-olds: Tim Thomas (in 2010-11), and Johnny Bower (in 1960-61). There was one 36-year old goalie in the league last season: Niklas Backstrom, who played 19 games.
The Panthers older players are, for the most part, not being counted on for large amounts of goal scoring. This is, of course, more true of players like Thornton, MacKenzie, Mitchell and the two goalies. But they are being counted on in massive part to stop goals from being scored.
The Cats have some otherworldly veterans on the roster who have defied the odds and have remained highly effective (looking at you Jagr, Luongo, and Campbell). The issue that remains as unpredictable and elusive as ever, is: do any of these large contributors hit the Jovanovski wall this season? As we found in our look at rookies, that question can’t be answered until the season gets rolling. This age factor brings us to our next unpredictable, yet related variable:
Who gets injured, how badly, and when?
Injuries in the NHL are like…..well, ….. rear-ends, everyone’s got em’ but some are bigger than others. One need look only as far back as last season to see what a critical injury at a critical time of the year can do to a team’s playoff chances. In the game against the Maple Leafs that would be one of the biggest debacles of the season, both Roberto Luongo and Al Montoya were injured. The injuries would combine to cost the team a critical win against a down-and-out opponent, and remove Luongo from the team down the playoff stretch.
The Panthers only lost 127 man-games to injury last season, which is nowhere near Columbus and their 383, or Colorado and their 325. It is also nowhere near the 336 man-games lost to injury the Panthers suffered through in 2011-12. But there are more than just totals at work here, as the nature of the player hurt can have a far more significant effect on the team. Take for instance Nick Bjugstad going down with his back injury for the stretch run last season (and playing through the injury prior to that). A second line center, leading the team in scoring, and having a breakout season is a costly player to have fall victim to injury, as is an All-Star goaltender.
It goes without saying that injuries cannot be predicted, but their impact on a team’s playoff chances is huge. Willie Mitchell’s injury last season had a substantial impact on a defense that looked more scattered and disorganized without him in the lineup. Of greater importance still, was that the Panthers injuries mostly came during the final portion of the season with the team in playoff contention, when wins were critically needed. Panther fans must hope that the team stays healthy, or a season of hope can quickly turn into a lottery pick disaster. Few teams Corsi’s, Fenwick’s or other advanced stats can stay in the top half of the league when important players, who have the largest effect on those stats, fall. A little luck on the injury front will go a long way.
Speaking of luck…..
A well-timed hot stretch of high PDO and Save Percentage can save a season:
PDO and Save Percentage are two stats that typically regress to the mean throughout the course of a season. Unusually high PDO or save percentage are typically unsustainable over long stretches, and higher than normal numbers in either category are usually considered the result of luck and/or a hot hand. Here is an excellent definition of this concept from www.theleafsnation.com:
Basically, if a team is playing with a PDO number way higher than 1.000, they’re producing above their expected output. If a team is playing with a PDO number below 1.000, they’re producing below their expected output. Over the course of a long season, the number will generally correct itself.
PDO can get crazy at times though. Note this (from the same article) from the 2013 season:
After 40 games or so games, we should expect just 5% of the teams in the league to be outside 1.025 or .975 (great work here by Snark SD). The actual number is 23%, as this is officially a silly season and has made an absolute mess of things.
This was seen last season as well, as the Ottawa Senators went on an unbelievable PDO and save-percentage tear at the end of the season to get into the playoffs. For approximately the final three months of last season’s regular season, the Senators were running a PDO of 102.1. Ottawa was out of a playoff spot by 14-points earlier in the campaign and still came back and got into the postseason. These things happen, and for an excellent explanation of why they can’t be sustained for long periods of time, see this article from Yahoo Sports Ryan Lambert. But for illustrative purposes, the Rangers finished the season with the league’s highest PDO, at 101.9, while Ottawa fell back to a season-long average of 101 (for 7th best in the league). Florida was 22nd in the league at 99.6.
But what also drove Ottawa was an incredible (and unsustainable) save-percentage by goalie Andrew Hammond. In 24 games last season, Hammond put up a .941 save-percentage. More importantly, he put together a .956 run over his first 13 NHL games, and put up a .960 run from February 18, 2015 through March 17, 2015. For comparison’s (and sanity) sake, Carey Price led the league in save percentage last season with a .933 in 66 games, which was remarkable. That should give you an idea of how unsustainable Hammond’s numbers would have been over a longer stretch, and how incredible (and lucky) his run was last season.
In 61 games last season, Roberto Luongo finished with a 2.35 goals-against-average and a .921 save-percentage, both exceptional numbers. However, in a stretch of losses from roughly January 13, 2015 to February 2, 2015, Lou’s relative save-percentage dipped below his average as he ran through a tough stretch. His save percentage in those games: .800, .893, .941 (against Oilers), .913, .706, .939 (against Columbus), .846, and .865. You may recall that period, where the team went 1-6-1 and left us scratching our heads and lashing out in any direction the compass could point us. Shane correctly pointed out, at that time, that it was a save-percentage dip that would correct itself, and it did.
A metrics destroying run of luck on the PDO and save-percentage charts at valuable times of the season (down the stretch or prior to Thanksgiving) can make magic happen in a playoff race. Like our other variables, there is no way to predict a PDO or save-percentage run, but when they happen it can make many problems disappear. On the other hand, avoiding a save-percentage or PDO dip of any longevity can also save a season in a parity-filled league.
How do you predict a sophomore slump?
The answer, of course, is that you don’t, and you hope you don’t have any of the dreaded second year slumps on the team. That said, there is no reason to think that any of Florida’s second year players, such as Araon Ekblad, Vincent Trocheck, or Alex Petrovic will fall victim to a slump, but there is also no way to predict whether they will. One more uncontrollable variable that will only let itself be known when the season gets rolling.
Outside Variables:
As stated earlier in this article, what happens to division and conference rivals can have a significant, and unpredictable outcome on the Panther’s season. Many hockey writers felt that the Panthers got into the 2011-12 NHL playoffs not just because of “loser points” in overtime, but also because the Southeast Division was weak that particular year. That may have been, but there is no way to predict whether (as an example) injuries will cripple another team like they did Columbus last season, or whether another team pulls an Ottawa-type run of luck. These things happen and they effect division and conference rivals. All of the unquantifiable variables we have explored for the Panthers exist for other teams as well.
If the Eastern Conference is as tightly grouped as it’s expected to be, a small point margin will mean the difference between success and failure. A division rival like Detroit may not have the services of Pavel Datsyuk for the start of next season, and at 37, he faces questions similar to the Panthers older players. Also, how the schedule unfolds, and what the schedule is during the time-period a key-player is out of the lineup will have an unpredictable impact on a team or opponent.
Many pundits look at the standings at Thanksgiving. Those teams in a playoff position at Thanksgiving are typically the same as those that end up making the postseason, with a few exceptions. Thus, a quick start to the season is often times essential to making the playoffs (and this served the Panthers very well in 2011-12). Staying with Detroit, and their lack of Pavel Datsyuk to start the season, the Wings open the season with back-to-back games, first at home against the Leafs, and than immediately they go on the road to Carolina for a game the following night. Three days later they play Tampa (in Detroit), followed by Carolina again three days after that. Immediately after the Carolina game they hit the road for a game the very next night in Montreal. Three days later the Red Wings head out on a Western road swing through Edmonton, Calgary, and Vancouver.
All teams have these stretches, and back-to-back games, and Detroit’s schedule includes 5-games in an 11-game October against non-playoff teams from last season. By contrast, Florida plays 11-games in October, but only one back-to back series (against Boston and Washington on October 30-31). But Florida also runs through the murderer’s row of California teams in early November, with back-to-back games against San Jose and Anaheim. How teams respond to their respective schedules, and what their opponents look like when they meet can have massive impacts on a team and their chances. For example, playing the Penguins when Sidney Crosby is out of the line-up is a much different proposition than playing a Pittsburgh with Crosby prowling the offensive zone. None of this is predictable, but all of it has an impact on playoff chances.
Teams with more back-to-back games and travel may be exposed to more injuries or illness. Teams traveling through the west coast early last season ended up having players exposed to the mumps. While that is statistically unlikely to occur again, when a flu bug hits a team can also have an unpredictable, yet important effect on a team’s won-loss columns. Losing Crosby to the mumps was something nobody could have predicted, but the benefits were certainly felt by the Penguins opponents during the stretch he missed with the illness.
Staying within the topic of “strange circumstances that can effect a team’s wins and losses,” something as weird as when a team plays an opponent can have an effect: on occasion the NHL may follow the circus at a venue, or on the heels of a basketball game and ice conditions may suffer as a result. This may work to a team’s advantage (Boston, for instance likes to grind games down and force opposition to play slow, and Madison Square Garden is renowned for poor ice), or disadvantage. Are the referees assigned to a game more apt to make mistakes? Referee Tim Peel has achieved Yahoo Sports hockey notoriety for the mistakes he has made in games.
Something as unpredictable and unquantifiable as the value of the Canadian dollar could also effect Canadian team’s efforts to sign free agents or make trades. Will a team be able to spend what it wished to at the deadline if the team is in contention if the budget is tight with a weak Canadian dollar. Far-fetched? Absolutely, but something to be wary of all the same.
Most of these last items border on the ridiculous, but wrap them within the greater issue and they take on some meaning: in a conference where a mere point or two can separate a team from being in or out of the playoffs, there is a lot of unquantifiable luck that goes into a final conclusion. Ultimately, it is normally the teams with the best metrics that end up in the playoffs, meaning the best teams, with the best players will succeed. But when you talk about the last few teams to get in, or be knocked out of the playoffs, the margins get fuzzy, and the separation in points, as well as metrics, becomes razor thin. That is where luck will have its greatest impact, and fans must hope that the roll of the dice breaks your way.
There is no greater variable than a “hot” goalie:
Of all the variables discussed here, there is simply nothing that can change a team’s fortunes more drastically than a goalie who plays above normal averages. Carey Price and his .933 save-percentage in 66 games was astounding last season. Equally amazing was Devan Dubnyk and his .929 in 58 games (with nary a night off during that stretch). During a critical stretch with Henrik Lundqvist out with an injury, Cam Talbot played 36 games for the Rangers and put up a .926 save-percentage. Could he have sustained that over the course of a season? Could Andrew Hammond have sustained a .941 over a season? Time may answer both of those answers, but both of these goalies hot stretches saved and/or propelled their teams into the playoffs.
For reference’s sake, in 2013-14, Tuukka Rask was the only goalie in the NHL who played more than 50 games who got to a .930 (in 58 games). That was a crazy season for Rask, and for Semyon Varlamov and his Avalanche team, as he played 63 games and posted a .927 save-percentage that propelled Colorado into an unlikely playoff appearance. Unexpected, exceptional, above-average goalie play is the most impactful of all the unpredictable variables that can propel a team into a playoff run.
A lot of things have to come together for a team to appear in the playoffs. One need simply speak with a fan in Ohio’s capital city to know what it feels like to have a team that has all the players with great metrics on paper to make the playoffs, only to see it all come apart, and be forced to sit through a season of misery due to unpredicted injuries. We here in Florida look at this Panthers team as one that may finally be ready to compete, but there are a lot of potholes out there that could just as easily blow it all up. Here’s hoping for some luck.