Some Things to Know about One-Goal Games

Last season, as hockey analysts struggled to explain how a possession-dominant Kings team failed to make the playoffs while Anaheim and Vancouver topped 100 points, there was a lot of discussion surrounding the role of one-goal games in the standings. LA’s disappointing season was largely dismissed as bad luck, with an argument that went something like this: the outcome of a one-goal game is effectively random, and the Kings’ 13-9-15 record in these games (against the 33-1-7 and 22-4-5 performances of the Ducks and Canucks, respectively) was the difference in keeping them out of the postseason. I wasn’t entirely convinced by this, but it got me thinking about the randomness of close contests. How random are one-goal games, and how significant a problem is this for people trying to use numbers to understand why some teams win and others don’t?

Photo credit: Flickr user pointnshoot. Use of this image does not imply endorsement.

One of these teams will probably win by one goal.

Primarily using WAR on Ice, I gathered game-level data from the nine 82-game NHL seasons from 2005-06 to 2014-15 (because I wanted to look at single-season counts of one-goal games, I excluded the short 2012-13 season), and characterized games as non-1GGs, regulation one-goal games, and OT/SO games. Except in unusual cases where they turned out to be game-winners, I dropped empty-net goals from final scores and characterized the game accordingly. So, for example, a 4-2 game where Team A’s fourth goal was an ENG would still be considered a regulation 1GG. In no particular order, here are five things I found out about one-goal games.

1. One-goal games are increasingly common.

There are a lot of reasons for it (the Bettman Point and decreased scoring league-wide are two big ones), but the NHL of today has more close games than ever. Back in 2006-07 and 2007-08, NHL fans were treated to 667 and 668 one-goal games, respectively; in 2013-14, we were up to 709 such games, and last season, there were 730. If you’re not a fan of increasingly random factors determining who gets into the NHL playoffs, this is something that should bother you.

2. Stronger teams win non-one-goal games pretty consistently, but regulation one-goal games aren’t entirely random.

If you’re the sort of person that enjoys using numbers to understand the results you see in the NHL (or, you know, you just gamble a lot), you should love non-one-goal games. For one thing, the year-on-year correlation in an NHL team’s non-1GG winning percentage is 0.5; put another way, about 25% of the variation in teams’ win % in non-close games can be explained by last year’s win %. And, as the following scatterplots show, that win % correlates closely to most measures of “this team gets good results”. (Note: These are full-season measures of GF%, CF% and PDO.) The GF% result isn’t surprising (the best way to end up with a big positive goal differential is winning a lot of blowout games, and vice versa), but possession and PDO each have a strong role in driving these outcomes.

1GG_fig 1

Now, if one-goal games are random, you’d expect zero autocorrelation from season to season, and you’d expect the graphs below to look like clouds with a flat trend line. And when you look at overtime and shootout games, that’s exactly what you see. But one-goal games ending in regulation fall somewhere between random and predictable.

1GG_fig 2

The autocorrelation in teams’ win percentages in these games is about 0.22; not terribly strong, obviously, but also pretty far from zero. What this all suggests is that 5v5-based hockey analytics are awesome in games that aren’t close, useless in OT/SO situations, and marginally useful in regulation one-goal games. It also suggests an interesting endogeneity question (how much of observed shot differentials is attributable to inferior teams playing conservatively to force overtime?) that I’ll let go for now. Given that non-close games are becoming less common, of course, none of this is great news for analytic game prediction.

3. In a lopsided match-up, the stronger team still has an excellent chance of winning a close game.

While the above graphs imply that the outcome of a regulation one-goal game is fairly unpredictable, there are situations where analytics still have considerable explanatory power. When one team’s score-adjusted Corsi % is 5 percentage points or more higher than their opponent’s, they win 68% of non-one-goal games, but they also win 60% of regulation 1GGs. When one team’s large-sample even-strength GF% is 5 percentage points or more better than their opponent’s, they win 71% of non-close games and 60% of regulation one-goal contests. In other words, so long as you have a strong sense of the relative underlying strengths of each team, you can be fairly certain that a much stronger team will win any game that doesn’t go to overtime.

4. Home ice advantage and back-to-back games don’t really matter in close games.

For someone who’s spent a ton of time writing about home-ice advantage and back-to-back effects, I was a little surprised to see that these effects are almost entirely confined to non-close games. That is, home teams win 59% of all non-1GGs, and teams on a back-to-back win just 40% of the time. In regulation one-goal games, however, the home-ice advantage is just 53% (close to coin-flip territory), and teams on back-to-backs win 48% of the time (both factors are irrelevant in OT and shootout games).

5. It’s largely impossible to guess how many one-goal games a team will play each season.

Of course, all of this would be super-useful information if there was a way to predict which games were likely to be one-goal and non-one-goal games. Unfortunately, this isn’t the case. On average, about 42.5% of games are non-1GGs; the only scenario in which the probability of a non-1GG differs much from this is a back-to-back situation, when about 45% of games are non-1GGs. More generally, as shown in the graphs below, teams that drive either a very good or a very bad goal differential will tend to play more non-one-goal games, but the effect is not as strong as you might think.

1GG_fig 3

So, what are we left with? On one hand, hockey games are a lot less predictable when the score is tight, and close games are becoming more and more common, but it’s largely impossible to anticipate which games will be close and which teams might play a lot of them. On the other hand, it looks like regulation one-goal contests are less random than some people think. (Shootouts are either a crapshoot, or they’re not.) These results can help us to make educated guesses about how often specific teams might face each type of game scenario, and how they might fare. For example, non-one-goal games might be more common in the Eastern Conference, where back-to-backs are more frequent. And, to the extent that forecasts of teams’ possession differential and 5-on-5 goal differential are becoming increasingly refined, this information may help in guessing how specific match-ups might play out. And, on the assumption that PDO will tend to regress strongly to average for most teams, it stands to reason that teams getting crushed in possession might find themselves on the wrong end of a lot of non-1GGs. Still, more work needs to be done before these can be much more than guesses.

Posted in Uncategorized | Leave a comment

Don’t Panic

In the NHL, as in the music of The Smiths, there’s panic in the streets of Los Angeles. What Morrissey and Johnny Marr might not be aware of, though, is the panic in the streets of Anaheim, Pittsburgh, Columbus, Boston, and Edmonton. It’s late October, and each team in the league has played about six games, so naturally we’re starting to hear which teams need to take drastic measures like firing coaches and trading core players.

Since hockey analysts are often accused of being condescending jerks who enjoy crushing the dreams of regular fans, let me use my spreadsheets to make a peace offering. Over the nine 82-game seasons in the salary-cap era, NHL teams have averaged 92 points. The year-on-year correlation in standings points is about 0.51. So, if you want to make a simple prediction of how a team will finish based on last year’s standings, all you have to do is pick the number of points halfway between 92 and wherever they finished last year.

When you do that, you can calculate a standard deviation between where teams have been predicted to finish, and where they’ve actually finished. While that doesn’t sound super-interesting, you can also use teams’ results through past seasons to figure out where the standard deviation of in-season points got smaller than the standard deviation of expected points. In other words, how many games did that team have to play before their record was more predictive of their finish than how they played the season before?

The answer to this question, over the seven 82-game seasons where the prediction is possible, is 55 games, on average. So, unless you have reason to believe that your team is much better, or much worse, than you were expecting them to be, you should relax before drawing any conclusions. There’s a lot of variation associated with that estimate of 55 games, but at no point in the past nine years has a team’s performance through six games been a better predictor of their finish than how they did the year before. In 210 team-seasons over that time, a team’s record through 20 games has only out-predicted their prior-season play six times, and all of the changes in question were very, very dramatic (i.e., Philadelphia going from 101 points in 2005-06 to 56 points in 2006-07). So, unless your team is embarking on a spectacularly good (or bad) season in 2015-16, they should probably sit tight before making any big decisions.

In other words, my advice to NHL fans at this early date is that nothing is decided yet; it really is a long, long season. Even if you think your team might be on the verge of a terribly disappointing season, it’s still way too early to know for sure. If your team is still looking bad come February, it’s probably okay to throw in the towel, but sit tight for now (and if you’re a Coyotes or Canucks fan, you know, good luck with that). But unless they’re doing something really unpredictable, you can probably wait until around 55 games before drawing any conclusions.

Posted in Uncategorized | Tagged , , , , , , , , | Leave a comment

Why Predicting the NHL Playoffs is Harder than You Think

It goes without saying that anyone who starts a site called Puck Prediction is pretty interested in finding ways to forecast the success or failure of NHL teams. But while prediction in the regular season can be interesting, let’s face it: predicting the winner of the Stanley Cup is all anyone really cares about. In the days before the playoffs begin every spring, the hockey media is flooded with series and Cup predictions, and the NHL now sponsors its own bracket challenge. What’s more, many in the analytics community have explored the predictive effects of various metrics, and some (like me) have developed models for picking series. Yet no matter what the approach taken, a way of picking the winners that’s better than some idiot choosing by home-ice advantage remains frustratingly out of reach. And I wonder if the exercise isn’t more hopeless than I’ve realized until now.

Much has been made of the role of one-goal games in the Pacific Division race this past season: despite the best puck-possession numbers in the league, an ugly 13-9-15 record in close contests bumped the 2014 Cup champion Kings out of the postseason, while a 33-1-7 record in one-goal games propelled Anaheim to the best record in the Western Conference despite the worst goal differential of all 16 playoff teams. More to the point, in games decided by two or more goals, LA had a healthy 27-18 record, while the Ducks were just 18-23 in such games. The implication here is that the relationship between possession metrics and winning is much clearer in games decided by more than one goal. And sure enough, when I looked at data on win percentages in non-shootout one-goal games and non-1GGs from the last 10 regular seasons, this appears to be the case.


Or, if you prefer that in puck-possession terms:


What the above tells us is that the correlation between winning percentage and goal differential or puck possession is pretty weak in one-goal games. As such, in close match-ups where a one-goal game is likely, we’re likely to find that our metrics aren’t very predictive, as random bounces assume a greater role in deciding the outcome. In my grand single-game-prediction experiment in 2013-14, the accuracy of my model was 9 percentage points lower in games with an expected a goal differential of one or less. This suggests that statistics are likely to struggle in any small-sample situation where the match-up is close. Which brings us to the playoffs. Replicating the analyses in the graphs above for postseason games yields similar but weaker results; the correlation between goal differential and playoff 1GG win % is just 0.13, while the correlation with score-adjusted Corsi is 0.09. In other words, with respect to the metrics analysts use to predict games, one-goal playoff games are effectively flips of the coin.

Over the past 10 seasons, postseason series have gone an average of six games; on average, three of those games have been one-goal contests. Across the four playoff rounds, series-winning teams have won two-thirds of these games. So let’s work out the math: if two of an advancing team’s four wins were coin-flips that could easily have gone the other way, the average Cup winner will have eight 1GG wins in a short space of time, and the average Finalist will have seven. A deeper dive into the past reveals some interestingly lucky (and unlucky) teams:

  • Of the 16 wins racked up by the 2006 champion Hurricanes, 10 were one-goal victories.
  • In a sign of things to come, the 2007 Ducks had just a 4-3 record in non-1GG contests in the playoffs, but skated away with the Cup on the strength of a 12-2 record in one-goal games.
  • I know the Blackhawks are a good team and all, but it’s probably worth mentioning that they’ve had a lot of luck in close playoff games during the Quenneville era. In the 2010 playoffs, Chicago was 6-1 in one-goal games, including three such wins in the Final (in fairness, they were 10-5 in their other games). In the 2013 playoffs, the Hawks had a 9-2 record in close games, including two one-goal victories to come back from 3-1 down in the second round. And this past postseason, 11 of Chicago’s playoff wins were by one goal. Over the last six playoff years, Chicago’s record in one goal games is a massive 35-17.
  • We all remember the 2011 Canucks for falling short in the Final against Boston, but Vancouver actually had a losing record in non-1GGs in that postseason. Their three one-goal wins in the finale gave them an 11-4 record in close games in those playoffs.
  • In the “we feel bad that St. Louis has never won a Cup” category: the Blues entered the 2013 playoffs with a decent chance at winning it all, yet lost four straight one-goal games in the first round.
  • Close contests were not kind to the Detroit Red Wings after their back-to-back Finals appearances in 2008 and 2009. In 2010, thanks to four one-goal losses to San Jose, they achieved the rare feat of losing a five-game series despite outscoring their opponent. In 2011, four more one-goal losses to the Sharks eliminated them, and in 2012, they lost three one-goal games in the opening round against Nashville. While they did benefit from four one-goal wins over Anaheim in the first round in 2013, two one-goal losses in Games 6 and 7 kept them from eliminating the eventual champions in the second round.
  • One more sign of how dominant the Red Wings were from 2005 through 2009: their playoff record in one-goal games in those four postseasons was just 16-17. In non-close games, they were 27-7.

So, the way to think about playoff series prediction is probably as follows. In any reasonably close match-up, we’d assume a six-game series, with three one-goal games and three non-1GG contests. We can expect the favored team, on average, to take two of the three non-1GG games. Assuming the other three games are effectively random, however, implies the following: in one in every eight series, or about twice every postseason, the underdog team will win all three games; also about twice every playoff year, the favorite will win all three one-goal contests. In three of every eight series (i.e., six times per postseason), the favored team will win two of the one-goal games and take the series in six. But six more times each postseason (again, on average), the underdog will take two of three, and the series will go to seven. The seventh game, of course, has a 50% chance of being another one-goal game. So, roughly a third of the time, we can expect that an inferior team will win a playoff series simply because of the randomness of one-goal games. This, of course, is completely agnostic to the effects of injuries, shooting luck, or hot goalies.

And one last note: I eagerly await the first person to try to convince me that Chicago’s success in one-goal playoff games is a repeatable skill, because, I don’t know, Corsi and Scotty Bowman and stuff. The point being that, if we’re going to doubt teams like Anaheim because they succeed on searing-hot play in one-goal games, we have to be willing to extend the same criticism to teams with gaudy possession statistics who also ride close-game wins to great success. I don’t doubt that the Blackhawks are an excellent team, and have been for a few seasons now, but it’s very easy to look at a team with strong fundamentals and Cup wins and assume that the Cups are primarily the result of the fundamentals. The problem, of course, is all the counterfactual examples of teams that had strong fundamentals and didn’t win anything because they didn’t have a 26-7 record in one-goal contests in three playoff runs. Such is the danger of assuming that random luck doesn’t play a massive role in playoff success, and of focusing too closely on results when evaluating how good a team is. Put simply, even when they’re steam-rolling through the postseason, few teams are as dominant as they appear.

Posted in Uncategorized | Tagged , , , | 9 Comments

How Well Does Score-Adjusted Fenwick Predict Playoff Series?

Back in the 2014 postseason, I started seeing analytics-inclined writers on Twitter repeating a fairly startling claim: they argued that someone could predict the outcomes of NHL playoff series with 70% accuracy just by comparing teams’ score-adjusted Fenwick (SAF) differential. I poked around in the data a bit, and even as Los Angeles lent support to this idea by winning the Stanley Cup behind the league’s best regular-season SAF, I couldn’t help but be skeptical. Insofar as teams that consistently outshoot their opponents win games more often than not, it stands to reason that – in the aggregate – teams with better possession numbers will win more series than they lose. But 70% is an awfully high rate of accuracy, especially coming from a single statistic; though I had a fun postseason in 2015, with my multiple-component “mGF%” model going 12-3, even my model is only about 66% accurate over the past seven seasons, and Steve Burtch’s expected-goals model is comparable to that. It’s also fair to point out that the 70% number was most likely based on the six seasons between 2007-08 and 2012-13 (we didn’t have WAR On Ice back then, so SAF data from 2005-07 weren’t easily available), or just 90 playoff series. As Michael Lopez showed in this terrific piece on the NHL’s SAP Matchup Analysis Model, predictive models based on tiny samples can yield very unreliable results. This made me wonder to what extent the 70% figure was a small-samples artifact (i.e., is SAF really that predictive, or are we just confirming that the Red Wings and Blackhawks were good in those seasons)? I also wondered about the extent to which the SAF effect was confounded by home-ice advantage. On average, we’d expect that teams with better regular-season possession numbers will end up finishing higher in the standings; as such, how much of the advantage that those teams enjoy in the playoffs arises from playing more games at home?


Now that the 2015 postseason is done, and after SAF had a rough time predicting series this year*, I decided to revisit this question. With ten seasons’ worth of regular-season SAF data, we can now look at a larger sample of 150 playoff series. To measure team possession, I used full-season team-level score-adjusted Fenwick-For % from WAR On Ice. I stayed away from the partial-season FF% and CF% statistics that I see others use, for reasons I’ve explained elsewhere. Apart from trying to confirm the 70% number, my analysis here focuses on how the predictive accuracy of SAF varies by home-ice advantage and match-up, but also includes some sensitivity analyses. First, I’ve always been skeptical of the validity and comparability of the SAF data from the 48-game 2012-13 season, so I’ve looked at SAF’s predictive accuracy with those 15 series removed. Another analysis excluded series involving the 2005-09 Red Wings and the 2009-10 Blackhawks, for reasons of comparability. My thinking here is that the institution of the salary cap and the elimination of front-loaded, cap-circumventing contracts by the current CBA will make it next to impossible to construct super-dominant teams like these in the future. I’ve presented the results with and without these  assumptions in place, so it’s fine if you don’t agree with me.

Unfortunately, the 70% number falls apart once we expand to the past 10 seasons: over the full sample of 150 series, SAF correctly predicted just 61.3% of playoff series. Without the 2013 playoff data, this falls to 59.3%; without the Wings/Blackhawks “superteams”, SAF is 59% accurate, and with both the superteams and 2013 excluded, SAF is just 56.3% accurate.

So, what about confounding and home-ice advantage? In the full sample, home ice is 55.3% accurate in predicting playoff series. Over these 150 series, the team with the higher SAF opened on home ice 81 times (54%). Teams with home ice and better regular-season possession numbers won 65.4% of their series. Teams with superior score-adjusted Fenwick numbers who opened on the road, however, won just 56.5% of the time. When we dropped the 2013 data, these win probabilities shifted to 63% and 54.8%, respectively. When we dropped the superteams, these probabilities were 62.1% and 55.9%, respectively, and with both the superteams and 2013 out of the analysis, they were 58.6% and 54.1%, respectively.

Something else that interested me was whether the predictive accuracy of possession varied by the difference in SAF between the teams involved. Presumably, a strong possession team matching up against a weak one would have a sizable advantage, while a more even match-up (e.g., the 2014 Western Conference Final, which pitted a 56.4% SAF team against a 56.1% squad) might not give a clear advantage to the team with the higher SAF. In the full sample, over 57% of series featured a match-up of teams separated by 4 or more percentage points of SAF; in these series, the team with the better SAF won 62.1% of the time. Oddly, in the 22% of series in which the teams were separated by a 1%-3% percentage-point difference in SAF, possession was even more accurate in predicting winners (66.7%). When the teams were separated by less than a percentage-point difference in SAF, possession was close to coin-flip accuracy (51.6%). This pattern is unchanged when we consider the same subsets of series as above: in close match-ups, SAF is basically worthless as a predictor, and it performs best when the disparity between the teams is moderate (suggesting an artifact of the small samples involved).

So, to wrap up:

  • Score-adjusted Fenwick is almost certainly not 70% accurate in predicting playoff series. Over the past ten seasons, SAF is 61% accurate, not far off from the 55% accuracy of home-ice advantage. The lesson in this: if you’re going to make bold claims, you should probably wait until you have a large sample of data to base them on.
  • To emphasize this last point, 150 series are still a pretty small sample, and I’m not convinced that 61% is a much more reliable estimate of SAF’s predictive accuracy than the 70% number.
  • The accuracy of SAF for predicting playoff series is confounded by home-ice advantage. In the full sample of 150 series, SAF’s accuracy dropped by 9 percentage points if the team with better possession numbers lacked home-ice advantage. This disparity shrinks when we remove the 2013 data and the superteams from our analysis, but this is mostly due to possession becoming less predictive among the home-ice teams.
  • In a close possession match-up, SAF is (unsurprisingly) almost useless for predicting the winner. If, like me, you think that changes to the NHL’s CBA will make ultra-dominant possession teams like the 2007-08 Red Wings and 2009-10 Blackhawks extremely difficult to assemble going forward, this doesn’t bode well for SAF’s future as a predictive metric.
  • Not addressed here, of course, are reasonable assumptions surrounding team Sh% and Sv%. In the long run, we’d assume that PDO will be close to league average in this analysis, but presumably analysts could easily improve on the accuracy of SAF by incorporating properly-regressed estimates of shooting and goaltending into their models.
  • This analysis also doesn’t dig into a critical assumption analysts make when they use SAF as a predictor; namely, that the team with superior possession numbers entering a series will actually control possession in that series, and will actually win because of it. In reality, this assumption is rarely born out on the ice: in my analysis last year, teams underperformed their expected possession differential in 69% of series.
  • My final piece of advice: any time someone tells you that they’ve boiled a complex game like hockey down to a single number that explains everything, you should maintain your skepticism. More generally, try to remember that puck-possession differential is just a statistic. It measures a set of underlying processes like effective neutral-zone play, crisp exits from the defensive zone, and the ability to maintain possession on the attack, but there’s nothing magical about the number itself. The sound, consistent execution of the above processes is what ultimately drives both good possession numbers and winning; the stats themselves don’t drive anything.

* The image above was my playoff bracket, not one defined by picking series with SAF. Puck possession went 7-8 in this year’s playoffs thanks to an ugly 2-6 first round.

Posted in Uncategorized | Tagged , , , , | 1 Comment

Interpreting Teams’ Puck Possession During a Streak

During the course of the NHL season, and especially into the postseason, many analysts look at partial-season estimates of puck possession as a way to get at the quality of teams’ underlying play. When we see a team race out to a fast start, we’ll often look at their early Corsi-For percentages to get a sense of whether that success is sustainable (sometimes, as in the case of the 2011-12 Wild or the 2013-14 Maple Leafs, our conclusions are entertaining). If a team labeled a contender hits an unexpected losing streak, we’ll look at their possession numbers during the streak to try to understand whether the rough patch is bad luck or something worse. And many in the hockey stats community handicap playoff series predictions by referring to a team’s CF% in the games leading up to a given series (as far as I can tell, the “last-25-games” measure is most popular).

Still, this has always kind of bugged me. Part of my concern is the sampling issue: how reliable are these CF% estimates if you only have a handful of games’ worth of Corsi events to measure? This is particularly true when it comes to using late-season CF% to predict playoff series: not only are you throwing out an immense amount of useful data, but the exercise is confounded by teams tanking their seasons, by key players being more likely to be injured, and by teams trading away important players at the deadline. (Another issue worth mentioning: imbalances between teams in strength of schedule in a subset of the season make it almost impossible to compare a team’s metrics to those of other teams, or to the same squad’s numbers at another point in the season.) But another significant concern involves teams’ records during a given span of games: if possession differential is correlated with winning, it makes sense that two teams with identical underlying fundamentals might have very different CF% measures over a short span if one is on a win streak and the other is losing games. I constructed a simulation study to explore (a) how much a team’s score-adjusted CF% over a certain number of games changes depending on their record, and (b) how likely we are to misjudge the strength of a team’s possession game based on a short-term streak.

I assumed that the best estimate of a team’s “true” possession differential was their 82-game score-adjusted Corsi-For %. This is informed by work by Micah Blake McCurdy, though I stuck with the original score-adjustment formula from Eric Tulsky and didn’t incorporate Micah’s venue adjustments. I used WAR On Ice to grab game-level data on every team’s even-strength Corsi events while tied, trailing, and leading, for the nine 82-game seasons between 2005-06 and 2014-15. (I didn’t feel that the “true possession” estimates from the 48-game 2012-13 season were comparable, so this season was excluded.) I divided the data by teams’ 82-game score-adjusted CF%. For simplicity, I used three categories: poor possession teams (those with SACF% lower than 48%; average SACF% 46.1%), neutral possession teams (SACF% 48% to 51.9%; average 50.1%), and good possession teams (SACF% 52% or higher; average 54.1%). Within each category, I divided the data again by game result, focusing only on regulation wins and regulation losses. For the analysis, I set up a Monte Carlo process that simulated 10-, 20-, and 50-game streaks for hypothetical poor, neutral, and good possession teams, by randomly drawing single-game results from won and lost games for each type of team. The idea being that the 10- and 20-game estimates represent streaks, while the 50-game estimates allow us to see how SACF% would converge in a larger sample of games. Within each streak length, I assumed winning percentages of 0.100, 0.300, 0.500, 0.700, and 0.900. Each simulation consisted of 10,000 draws.


The above table describes how SACF% changes depending on teams’ records; these results didn’t change by streak length. As you can see, a team’s score-adjusted Corsi can vary by about 2 percentage points depending on whether they’re winning or losing. A bad team on a winning streak will tend to have a score-adjusted Corsi not far below that of a neutral team, and a strong possession team that can’t buy a win will tend to have possession numbers closer to those of an average team. Keep in mind, of course, that this analysis is completely agnostic to things like goaltending or a run of hot shooting. As the number of games gets large, this relationship doesn’t really change, but, of course, it’s basically impossible to maintain 0.100 or 0.900 play for a long period of time. So, if we have a strong possession team that has one ugly 2-18 stretch, but plays 0.500 hockey the rest of the season, on average we’d expect them to finish the year with a SACF% around 53.5% (i.e., a weighted average of the 52.7% during the skid and the 53.7% in the remaining games).


Where this gets interesting is in the estimated probabilities of a team’s possession play in a streak representing who they really are. Based on this analysis, it’s unlikely that a strong possession team will be mistaken for a weak one, or vice versa. Beyond that, though, all bets are off. A bad possession team on a 9-1-0 streak will have the Corsi of a decent squad 36% of the time, while a bad puck-control team with a 0.700 win percentage in 20 games will look like an average team a quarter of the time. Similarly, a strong possession team on a 3-7-0 streak will have the Corsi of a mediocre team in 29% of cases. Good and bad possession teams playing 0.500 hockey have a roughly 24% probability of being mistaken for an average team. As far as neutral possession teams, it’s easy to mistake them for a team in one of the other categories, depending on whether they’re winning or losing. A team like this that goes 1-9-0 will be mistaken for a lousy team 30% of the time, but if they go 9-1-0, they’ll be mistaken for a strong possession team almost a third of the time.

Some takeaways:

  • In the setup of the Monte Carlo, we assume that we know how good a team truly is at controlling possession. In reality, we usually don’t know this. As such, any time we try to make judgments about a team’s fundamental play based on a small number of games, we’re likely to misjudge them if we don’t at least look at their record during those games.
  • Insofar as 82-game score-adjusted possession estimates are the closest we’re going to get to estimating a team’s true quality, we should use these whenever we can. When trying to use possession measures in evaluating playoff match-ups, it’s probably wise to use full seasons of data rather than a sample of recent games. Unless two teams enter a series with similar recent records and comparably tough schedules, we may have difficulty comparing CF% numbers from those games.
  • To draw some practical insights from these analyses, last season’s Columbus Blue Jackets won 16 of their final 19 games after struggling through the season’s first five months. Their SACF% prior to the streak was a miserable 45.9%, but they finished the campaign playing 49.9% hockey. These results suggest that the Jackets’ late-season play may not have been representative of their true quality; rather, they may have just seen the typical Corsi bump of a winning team. In contrast, the Pittsburgh Penguins won just four of their final 15 games, yet played 54% possession hockey over that stretch; my analysis implies that 54% may have underestimated how well they played last season.
Posted in Uncategorized | Tagged , , , , | 1 Comment

2014-15 NHL Season Review: Pacific Division, Playoff Teams

It’s time to wrap up my look back at the seasons of all 30 NHL teams. If you missed them, check out my reviews of the Metropolitan (playoff and non-playoff teams), Atlantic (playoff and non-playoff teams), and Central (playoff and non-playoff teams) Divisions, as well as my look at the non-playoff Pacific squads. Today, we finish with the Pacific playoff teams.

Anaheim Ducks

For those of us interested in using numbers to try to figure out the NHL, no team is honestly more interesting than the Ducks. The conventional wisdom about winning squads in the analytics community goes something like this: shooting and save percentages tend to regress to the mean over time, so any success that doesn’t derive from effectively driving puck possession will tend to be fleeting. Of course, though the conventional wisdom is in general true, it doesn’t follow that it should be necessarily true for every team in every season. And after winning their third straight Pacific crown, pacing the Western Conference in the regular season, and advancing to Game 7 of the Conference Finals, all with fairly mediocre possession numbers, it’s safe to say that Anaheim is a pretty dramatic exception to the rule.

I think the explanation for how the Ducks have been so successful in the three full seasons under Bruce Boudreau is some combination of the following:

  1. Luck: I can concede that the Ducks may have been fortunate to win the Pacific in the 48-game 2012-13 season. They were a poor shot-creation team that benefitted hugely from 8.6% shooting at even strength, and tremendous 0.930 goaltending from Jonas Hiller and Viktor Fasth helped them to finish ninth-best in goals against. More generally, given what we know about the respective roles of luck and talent in the NHL regular season, 48 games is just too small a sample for us to know if Anaheim was the best team in the Pacific that year. I can also concede the utter strangeness of the Ducks’ 2014-15 season, in which they finished with a goal differential of just +10 (worst among all 16 playoff teams), yet reached 109 points due to an absurd 33-1-7 record in one-goal games. What’s harder to concede is their division title in 2013-14, and the fact that they’ve strung together three such successful years. Following 2012-13 with another title suggests that their result in the short season wasn’t a fluke, and at some point, warnings that the Ducks’ shooting percentage would regress to average have sounded increasingly silly: over the past three seasons, Anaheim has shot 9% at 5-on-5 as a team, and if that’s puck luck, 212 games worth is an awfully improbable amount of good bounces. While I’ll agree that the one-goal game performance this past season was insane, if you believe the Ducks have been successful because they’ve done a lot of things right, you can even make the argument that Anaheim in 2014-15 was an essentially solid team whose luck in close games offset their struggles in goal (0.919 5-on-5 Sv%) following Hiller’s offseason departure.
  2. They Know What They’re Doing: The luck-vs-talent work referenced above strongly suggests that it’s incredibly unlikely for a team to lead its conference in standings points over a 212-game span on luck alone. Another set of data worth considering: in the full seasons in which he’s stood behind an NHL bench, Boudreau’s results are as follows: 1st in the Southeast, 1st in the Southeast, 1st in the NHL, 1st in the Eastern Conference, 1st in the Pacific, 1st in the Western Conference, 1st in the Western Conference. So, you know, it’s possible that he knows what he’s doing. There’s a case to be made that Anaheim wins by having elite talents who can drive above-average on-ice percentages, while having passable (if not dominant) underlying numbers. Some analysts have tended to lump the Ducks in with teams like the Maple Leafs, Avalanche, and Flames (i.e., teams that have won despite awful fundamental play), but this is unfairly negative. Over the past three seasons, the Ducks’ score-adjusted Fenwick is 50.8%; not a dominant number, but far from a terrible one. Over the same period, the Ducks sit squarely in the middle of the league in both shot-creation and shot-suppression: again, not great, but not poor enough to suggest a problem. The organization’s talent for finding good young goaltenders has consistently given them an above-average team Sv%, and it’s probably safe to assume that any team with a healthy Ryan Getzlaf and Corey Perry will score on an above-average percentage of their shots. As such, rather than follow the blueprint for success that analytics often lays out (i.e., assume league-average percentages and try to drive a good goal differential through possession), the Ducks appear to assume that they’ll come out ahead in the percentages, and that “good enough” in the fundamentals is enough. And thus far, they’ve been right.
  3. My Comparative Advantage Theory is Awesome: Recently, I posited that higher-scoring teams tend to have greater success in low-scoring NHL seasons, and vice versa. The idea being that offensive ability is hard to come by in low-scoring eras, and as such will set a small number of teams apart from the pack. Data throughout the league’s history are generally supportive of my theory, and in the past three seasons, only two teams have scored more goals than Anaheim (if you’re intrigued, those two teams were this season’s Cup Finalists). If my idea is right, you have to give Boudreau credit for choosing his jobs wisely: the top-scoring team from 2007-08 through 2010-11 was his Capitals. This would suggest that, as long as their two offensive stars stay healthy, and assuming a bounceback year from their goalies, the Ducks can be expected to keep on winning.

If fans of the other six teams in the Pacific needed any more bad news, Anaheim also has one of the deepest prospect pools in the league, and has tons of cap space heading into the offseason. After a strong playoff run, unrestricted free agent Matt Beleskey will likely be seeking big money; at just 26, and following a 22-goal season with solid two-way numbers (4.3% Corsi Rel), there’s a case to be made that he’s worth retaining, though probably not for the money he’ll sign for. The same probably can’t be said for 34-year-old free-agent blueliner Francois Beauchemin, who will almost certainly be offered more than he’s worth at this point. Jakob Silfverberg and Carl Hagelin (a superb draft-day acquisition from the Rangers) will both presumably get new deals as RFAs. The greater challenge for GM Bob Murray is likely to come in 2016, when Richard Rakell, Jiri Sekac, Simon Despres, Sami Vatanen, Hampus Lindholm, and goaltenders Frederik Andersen and John Gibson are due new deals, all as RFAs. At 30 years old, Ryan Kesler is a UFA a year from now; if he figures into Anaheim’s future plans, he could be offered an extension soon. The Ducks’ strange dissatisfaction with deadline acquisition James Wisniewski ended on draft day, as they shipped him to Carolina for additional goaltending depth in Anton Khudobin. Following the desultory finish to the Ducks’ season (two straight 5-2 losses to Chicago), there were briefly rumors that Boudreau may be sent packing. Since Anaheim wasn’t that stupid, though, it’s probably safe to pencil them in for another strong season in 2015-16.

Vancouver Canucks

Now we’re entering the “how the f— did this team make the playoffs?” section of this post. After a heartbreaking loss in Game 7 of the 2011 Finals and first-round losses in 2012 and 2013 (three division titles and two President’s Trophies in those seasons notwithstanding), ownership jettisoned long-time head coach Alain Vigneault and goaltender-of-the-future Cory Schneider in the 2013 offseason. And as Vancouver tumbled in the standings in the second half of 2013-14, months of uncertainty surrounding the status of franchise goalie Roberto Luongo came to an end, as he was dealt to Florida at the trade deadline. After seeing the Canucks miss the playoffs, fire GM Mike Gillis and head coach John Tortorella, and replace Torts with Willie Desjardins (a successful WHL and AHL coach), it appeared that a rebuild was underway in British Columbia. Yet these moves weren’t followed by further efforts to get younger or shed bad contracts, and after taking a look at their aging, declining roster, it was easy to write Vancouver off as a directionless squad with little chance in 2014-15.

And yet, here we are: for all the negatives, the Canucks racked up 101 points, good for fifth-best in the Western Conference, and began the playoffs on home ice while the last three teams to eliminate them from the postseason all watched from home. In their first season under Desjardins, Vancouver’s fundamentals were solid, if unspectacular: their 53.5 Corsi attempts for per 60 5-on-5 minutes were the ninth-lowest in the league, and their rate of shot prevention (54.5 Corsi against per 60) was mediocre. With a 50.5% score-adjusted Fenwick, their possession game was effective, but far from elite. And with lackluster 7.7% even-strength shooting and 0.917 goaltending, it wasn’t a surprise that Vancouver’s goal differential at 5-on-5 was negative. These numbers suggest a fairly mediocre team, and in a down year for the West, fifth in the conference was probably representative of the top of the playoff bubble. For anyone who followed the travails of the Kings and Ducks this season, it’s no surprise how the Canucks ended up on the right side of the bubble: their 22-4-5 record in one-goal games was second only to Anaheim’s.

Assuming that Los Angeles and San Jose work their way back into the playoff hunt next season, you can probably guess that I’m not bullish on the Canucks in 2015-16. Vancouver has some intriguing prospects working their way into regular NHL duty, including Hunter Shinkaruk, Bo Horvat, and Linden Vey, but many key contributors from 2014-15 are near or over the wrong side of 30, and the team’s cap space is extremely limited. Offseason signing Radim Vrbata led the Canucks with 31 goals, but the team may be content to watch his contract year play out rather than offer an extension, as Vrbata will be 34 at the start of next season. Dan Hamhuis, on the other hand, is entering the final year of his deal, and with a 1.4% Corsi Rel, was one of Vancouver’s better defensemen last season. Given the financial constraints the team is facing, the new deals for Derek Dorsett (7 goals and an ugly -7.8% Corsi Rel) and Luca Sbisa (-3.1% Corsi Rel) greatly complicate the picture. Finally, the team’s goaltending situation is (once again) a bit of a mess. After wading through an unpleasant goalie controversy for two seasons (and losing two excellent netminders in the process), Vancouver immediately took on another one: after young Eddie Lack delivered a solid 0.925 in 41 games in 2013-14, the Canucks supplanted him as the team’s future by signing free agent Ryan Miller to a lucrative three-year deal. Aside from the problematic implications of his $6M cap hit, Miller’s career 0.922 Sv% also suggests that he’s, you know, nowhere near good enough to backstop a team with neutral possession numbers to contention. Miller’s 0.913 work in 45 games was a big reason why Vancouver ranked 19th in goals against this season, and the team entered the postseason without a clear starter in net. Yet when presented with a chance to deal one of their goalies at the draft, they traded Lack to Carolina for picks. This will leave them entering 2015-16 with an unappealing tandem of Miller and RFA Jacob Markstrom (an intriguing young goalie who’s struggled badly at the NHL level). So, while the Canucks exceeded my expectations this past season, they still look to me like a team in decline, and I’m not sure their management is capable of guiding them through what has already been a difficult transition.

Calgary Flames

Okay, now we’re really there: seriously, how the f— did this team make the playoffs?

Before we dive into any of the numbers, it’s probably worth noting a few points for context. First off, the 2014-15 Flames weren’t really comparable to teams like the 2007-08 Canadiens or the 2013-14 Avalanche, who legitimately crushed their regular seasons despite weak fundamentals. Calgary finished last season 16th in the NHL, and only clinched a playoff berth in the campaign’s final week. And though they surprised many (but not all) observers by triumphing over Vancouver in six first-round games, if you’ve read the rest of this post, you understand that the Canucks were far from a strong opponent. So, while it’s fair to say that Calgary defied everyone’s expectations in 2014-15, their season is hardly a death knell for possession-based analytics. Even when having what many considered to be a miraculously lucky campaign, the Flames were at best a bubble team, and could easily have found themselves on the outside of the playoff picture.

The reality for this team is that it’s still working to rebuild after failing to deliver a Stanley Cup champion around Jarome Iginla and Miikka Kiprusoff, and waiting too long before beginning that rebuild. After Iginla and Jay Bouwmeester were dealt at the 2013 trade deadline, Calgary cratered to 13th in the West in both 2012-13 and 2013-14. But the returns on those painful seasons are already in evidence. Forwards Johnny Gaudreau and Sean Monahan, just 21 and 20 years old (respectively), combined for 56 goals and 126 points last season. At the back, 24-year-old T.J. Brodie (41 points, 1.8% Corsi Rel) has joined Norris-quality veteran Mark Giordano (5.7% Corsi Rel) to form a very solid anchor D pairing. At this year’s draft, the Flames set the league abuzz by snatching coveted RFA Dougie Hamilton away from Boston. And many more promising youngsters, including Sam Bennett (who missed much of last season with shoulder surgery, but took part in their playoff run), Josh Jooris, and Markus Granlund, are on the way to the NHL roster.

The danger, of course, is that Calgary looks past their wretched underlying numbers and assumes that the rebuild is further along than it actually is. And make no mistake, the underlying numbers suggest a team that’s very much a work in progress. For context, the Flames’ 45.6% score-adjusted Fenwick is the 23rd-worst possession season in the last 10 years; only the 2012-13 Maple Leafs and the 2010-11 Ducks (is there a connection between those teams?) made the playoffs with poorer numbers. Calgary’s rate of shot creation (just 49.9 Corsi for per 60 5-on-5 minutes) was fourth-worst in the league, and only Buffalo allowed shots against at a higher rate than the Flames’ 62.4 Corsi per 60. What’s worse, the goaltending they got from Jonas Hiller (0.927 5-on-5 Sv%) and Kari Rammo (0.917) wasn’t especially good; Calgary finished a dismal 17th in goals allowed, and without one of the league’s better penalty kills, it would’ve been worse. What saved the Flames’ bacon was 8.9% shooting at even strength; despite their lousy shot creation, they finished 6th in the league in scoring. It’s difficult to say how sustainable that Sh% is – given the dramatic turnover in their roster in recent years, it’s tough to tell what their baseline should look like – but it’s probably a safe bet to guess that 31-year-old Jiri Hudler won’t hit 31 goals or 76 points again next season.

The smart play for Brian Burke and GM Brad Treliving, then, is to use the team’s oceans of cap space to resign young prospects and important veterans, while resisting the temptation of more aggressive moves that set back the development of the Flames’ promising core. This may, of course, be a lot to ask of the current management group, which burned a year of Bennett’s entry-level contract for three playoff games in a series they were losing 0-2, and that inexplicably signed Brandon Bollig and Deryk Engelland to multi-year free-agent deals last year. New RFA contracts are due to Hamilton, Jooris, Mikael Backlund, and Michael Ferland, with Monahan, Gaudreau, Granlund, and D prospect Tyler Wotherspoon up next year. It wouldn’t be at all surprising to see a lengthy extension for Giordano this offseason, and if Hudler figures into Calgary’s plans after next season, now would be the time to get a deal done. With Ramo an unrestricted free agent, the goaltending situation is a bit uncertain; while he played reasonably well in the second round against Anaheim, Ramo’s age (28) and career 0.915 Sv% in the NHL suggest that the team should probably move on. Still, with Hiller only under contract for one more season (and having lost the team’s confidence during the playoffs), and few top netminding prospects in the system (Joni Ortio is probably the most NHL-ready, but that’s not saying much), the Flames would probably like to have another experienced goalie in the wings. If Calgary plays it safe and conservative next year, it might well mean a 2015-16 without playoff hockey, but in the bigger picture, that might not be a bad thing.

Posted in Uncategorized | Tagged , , , , , | Leave a comment

2014-15 NHL Season Review: Central Division, Playoff Teams

Now that the 2014-15 season is in the books, I’m taking a look back at the seasons of all 30 teams. I’ve wrapped up my look at the Eastern Conference, including the Metropolitan playoff and non-playoff teams and the Atlantic playoff and non-playoff teams, and we’ve previously checked out the non-playoff squads in the Central and the Pacific. Today, we check in on the playoff teams in the West, starting with the best teams in a very competitive Central Division, including the Stanley Cup champion Chicago Blackhawks.

St. Louis Blues

For all the moaning of fans in, well, nearly every fan base not in Chicago or Tampa, few NHL franchises have a star-crossed, demoralizing back story to rival that of St. Louis. Brought into the league during the 1967 expansion, the Blues have a proud and illustrious history, including three Cup Final appearances under the leadership of Scotty Bowman, the Hall of Fame career of Bernie Federko, the 1986 Monday Night Miracle, many successful seasons with Brett Hull, Adam Oates, Al MacInnis, Chris Pronger, and (briefly) Wayne Gretzky, and most recently, four seasons as an elite defensive team under Ken Hitchcock. In their 47 seasons in existence, the Blues have missed the playoffs just eight times, yet they have the same number of Stanley Cup championships as the proposed expansion team in Las Vegas. In the four years under Hitchcock, only the Kings and Blackhawks have had better possession numbers than St. Louis’s 54.1% score-adjusted Fenwick, and only Los Angeles, New Jersey, and Detroit have done a better job suppressing shots against. If you want to nitpick, you could point out that the Blues have ranked just 14th in goals scored in that time, and argue that it’s hard to be a dominant team in today’s low-scoring NHL without elite offensive firepower. But it’s hard to nitpick the results: since 2011, St. Louis has finished with either the second- or third-best record in the Western Conference, and won the 2011-12 Jennings Trophy with an absurdly low 165 goals against.

In 2014-15, it was more of the same.  With 109 points, the Blues once again captured the Central title, in a year in which theirs was the most competitive division in hockey. Behind a lethal power play and 8.2% shooting at even strength, St. Louis finished fifth in the league in scoring, and their superb shot prevention work tied them with L.A. for the fourth-lowest goals allowed. 23-year-old Vladimir Tarasenko dazzled with 37 goals and 73 points. Alex Steen lived up to his new contract with 24 goals and 64 points, captain David Backes scored 26 times, and young Jaden Schwartz broke out with 28 goals and 63 points. Key free-agent signing Paul Stastny added just 46 points, but with a 3.4% Corsi Rel, his work as a two-way center came as advertised. In an injury-shortened campaign, Kevin Shattenkirk had a particularly strong 4.8% Corsi Rel, but other big-minutes defensemen struggled a bit to drive play, including Alex Pietrangelo (-2.5% Corsi Rel) and Jay Bouwmeester (-2.8%). Still, for all the positives, it’s impossible to avoid mentioning their first-round series loss at the hands of Minnesota. St. Louis did control play effectively in the series, but apart from their six-goal explosion in Game 4, they managed only four goals at 5-on-5 against the Wild, and with the series tied at 2-2, goalies Brian Elliott and Jake Allen conceded five even-strength goals on just 34 shots in Games 5 and 6. In short, when they needed some bounces most, their puck luck was awful.

St. Louis is clearly a team built to win now, and they’ll very likely look the same come next season. Tarasenko will receive a raise as an RFA, as will Allen, but the team has plenty of cap space to make those moves work. Given the manner in which their playoff year unfolded, I wouldn’t be shocked if the Blues went after another goaltender. But honestly, Allen is the future in net for this team, and it would be arguably counterproductive to add another obstacle between him and NHL experience. If St. Louis has a real need, it’s probably for another scoring forward. Whether such an addition would fit into their cap picture is a different question, but regardless, there probably won’t be a better time for bold action in the interest of winning.

Nashville Predators

A lot of unexpected things happened during the 2014-15 regular season, and I’ll admit up front that the year the Predators just had came completely out of left field for me. While they’ve made regular appearances in the postseason over the past decade, much of the Barry Trotz era in Nashville was characterized by teams that (a) got badly outshot and (b) had elite goaltending from either Tomas Vokoun or Pekka Rinne to cover for their defending. Between 2009-10 and 2011-12 (all playoff years), the Preds’ possession game deteriorated badly, and in 2012, Nashville was one of the worst possession teams (47.4% score-adjusted Fenwick) in the past decade to open the postseason on home ice. In 2012-13, the crash finally arrived, as Nashville’s SAF dropped to an ugly 46.5%, and despite a 0.929 Sv% season from Rinne, the Predators finished a lowly 27th in the league. The following year, Trotz was able to tighten Nashville’s systems, as they delivered stronger defensive work and an improved 49.1% SAF, but with Rinne absent or struggling due to hip ailments, the team’s goaltending cratered to 0.911 at even strength, and they finished just 19th. Trotz was fired in the offseason, and given the Preds’ history of meager goal-scoring and weak possession play, I had low expectations for them in 2014-15.

So, naturally, they completely surprised me. Rinne had a superb comeback season, with a 0.937 Sv% in 64 games; overall, Nashville enjoyed the fourth-best goaltending in the NHL this year. Even more interesting, though, was the overhaul in systems engineered by new head coach Peter Laviolette. The Predators’ 53.3% score-adjusted Fenwick was fifth-best in the league, a remarkable turnaround for a team that has historically struggled to drive play. Despite Trotz’s reputation as a defense-first bench boss, Nashville’s rate of shots against actually dropped in year one under Laviolette, from 54.3 Corsi against/60 5-on-5 minutes to 52.2 per 60. What’s more, the team’s offensive productivity soared from an ugly 51 Corsi for per 60 to 58.4, the sixth-highest rate of shot-creation in the league, and the Preds became the NHL’s sixth-highest scoring team at 5-on-5 this year. Calder nomineesnub Filip Forsberg led the team with 26 goals and 63 points; at just 20 years old, Forsberg is only going to get better, and the 2013 trade that sent him to Nashville from Washington is only looking worse with time. Offseason acquisition James Neal chipped in 23 goals in 67 games. Mike Ribeiro had a strong season, and appears likely to sign a multi-year deal to stay in Tennessee, but given his age (35) and history of unpleasant off-ice behavior, it’s not clear that this is a great idea. Also breaking out were 24-year-old Colin Wilson, who scored 20 goals and contributed solid two-way minutes, and second-year defenseman Seth Jones (1.8% Corsi Rel), who was one of the team’s better blueliners. On the down side, brilliant goaltending disguised the struggles of defensemen Roman Josi (-4.4% Corsi Rel) and Shea Weber (-4.1%).

Long-time GM David Poile has a healthy amount of cap space to work with this offseason, but also has a fair number of decisions to make. The aforementioned extension for Ribeiro is likely to be costly, as he’s coming off a 62-point season. Wilson, center Craig Smith (who scored 23 goals and 44 points), and young Calle Jarnkrok are all due new deals as RFAs, and long-time Pred Mike Fisher is newly signed to a two-year extension. Forsberg and Jones will hit RFA a year from now, and will certainly merit raises. On the other hand, an ugly 0.971 on-ice PDO disguised strong two-way play by trade acquisition Cody Franson, and it appears the team will pass on signing the free-agent blueliner, along with the aging and declining Anton Volchenkov. In goal, Rinne is signed at a hefty $7M cap hit for four more years, but young Marek Mazanek continues to develop as the heir apparent, and is still a year away from RFA. As such, while it’s easy to wonder if a team’s sudden success will be transitory, it appears that the Predators have both the systems and the talent in place to compete in the Central next year.

Chicago Blackhawks

Moving on from franchises who disappointed this postseason, now we come to a team whose recent playoff fortunes have been almost unimaginably rosy. After a tough regular season that saw them post their worst possession numbers in the Joel Quenneville era (granted, a still-solid 52.8% score-adjusted Fenwick), that saw their defensive numbers tumble from the league’s elite, that saw Patrick Sharp’s poor season and Patrick Kane’s mid-season injury drop them all the way to 17th in scoring, and that saw them finish just two points ahead of the wild card spots in a tough Central Division, everything came together for the Blackhawks the moment the postseason started. Opening the playoffs on the road in Nashville, unheralded backup goaltender Scott Darling took over for the shaky Corey Crawford, and proceeded to outduel Vezina nominee Pekka Rinne as Chicago took the series in six. In the second round against Minnesota, Crawford turned back into a wall, and the Hawks completed their trip through the Central playoffs with a surprising sweep. In the Conference Final against Anaheim, they fought off two elimination games and a ton of score effects, and came back from 2-1 down in the Final to triumph over Tampa Bay in six games. With the victory, Chicago captured their third Stanley Cup in six seasons, a remarkable accomplishment in an era constrained by the salary cap.

Still, given the decline in the team’s underlying numbers and the salary-cap crunch that’s now arrived in Chicago, it’s hard to escape the sense that this was a last hurrah of sorts for this group of players. The matching $10.5M extensions for Kane and Jonathan Toews kick in next season, and GM Stan Bowman has a lot of work to get done despite very little cap space. The world of hockey commentary is flooded these days with proposed trades that could solve the Blackhawks’ cap crunch, but the game theorist in me questions Bowman’s ability to make good deals given his limited leverage. (I mean, if you’re one of the league’s other 29 GMs, are you going to take on bad contracts to help the most successful franchise in the cap era keep their lineup together?) Gifted winger Brandon Saad is due a sizable raise as an RFA; other important RFAs include defenseman David Rundblad and center Marcus Kruger. A year from now, key defense prospects Stephen Johns and Trevor van Riemsdyk will also be RFAs. All-Star defenseman Brent Seabrook is entering a contract year, and at 30, is almost certainly hoping for a massive multi-year deal. Unfortunately for Bowman, very little money is moving off the books either now or next season, and many of the trades that have been proposed to provide cap relief will leave significant holes to fill. With Johnny Oduya and Michal Rozsival both unrestricted free agents and Kimmo Timonen retiring, Chicago may have a serious shortage of NHL defensemen next season; if Seabrook leaves during the next year, the Hawks as constituted will be left with Duncan Keith, Niklas Hjalmarsson, and a slew of promising but unproven blueliners (led by Rundblad). Sharp’s $5.9M cap hit (due for two more seasons) has been brought up frequently in trade rumors, but after a disappointing 16-goal season, the market for the 33-year-old winger may not be strong. Marian Hossa could be an intriguing trade candidate, though such a deal would make Chicago a weaker team overnight. The Slovak winger is a future Hall of Famer and still a very effective player, and at this point, he’s only owed about $16M over the next six years; a team needing to reach the cap floor would likely see a lot to like in acquiring Hossa. Of course, six years of a $5.25M cap hit is a lot to commit to a player who will turn 37 next year. Crawford has also been mentioned as a trade possibility. The Hawks goalie played well in the regular season and in the final three rounds of the postseason, and it’s not clear who would start 65 or so games a year for Chicago if Crawford were dealt. While no one doubts the competence of the front office running the Blackhawks, the juggling act required to maintain the level of on-ice talent we’re used to here may be untenable. In short, while Hawks fans will no doubt spend the summer savoring the thrill of victory, it should be a very interesting few months for this franchise.

Minnesota Wild

A few years ago, if you’d asked me where I’d find an NHL team committed to a mercenary strategy of buying a Stanley Cup through high-priced free agents, I don’t think I would’ve guessed Minnesota. Yet ever since owner Craig Leipold committed almost $200M to the past performances of Zach Parise and Ryan Suter in 2012 (and then complained about runaway inflation in player salaries while leading the charge for the lockout later that year . . . but I digress), this has been the plan of action in St. Paul. Need a solid two-way center? Acquire then-30-year-old Jason Pominville at the 2013 deadline and commit to five years and $28M. Need a scoring forward? Sign up for a $6.5M cap hit with Thomas Vanek. Need a random third-line grinder? Give $5M to Matt Cooke (and then buy him out two years later when you need cap space). And, series victories over Colorado in 2014 and St. Louis in 2015 aside, the results haven’t really been there: for all the big names brought in to fill out the Wild’s roster, they’ve only managed to slot into the eighth Western Conference seed in 2013 and wild-card slots in both 2014 and 2015.

More to the point, in 2012-13 and 2013-14, Minnesota’s middling 49.1% score-adjusted Fenwick suggested a team whose ceiling wasn’t much higher than the playoff bubble. And in 2014-15, though their possession numbers improved to a solid 52.4% SAF, they might not have seen the postseason if not for an unlikely savior. Midway through January, the Wild were floundering at 18-19-5, well outside the playoff picture. Importantly, they had the NHL’s worst goaltending, with an 0.895 Sv% at even strength, and looking for veteran depth, they acquired Devan Dubnyk in a trade with Arizona. Insofar as Dubnyk had had a disastrous 2013-14, beginning the year as Edmonton’s starter and finishing it playing for Montreal’s affiliate in the AHL, and insofar as his career 0.918 Sv% at 5-on-5 didn’t suggest a diamond in the rough, there was no reason to suspect that Minnesota’s season was about to turn around. Yet behind Dubnyk’s 0.940 goaltending (and, it must be acknowledged, 9.4% team shooting), the Wild stormed to a 28-9-3 finish, and upset the heavily favored Blues in round one before being swept by the eventual champions.

With Josh Harding unfortunately (but understandably) set to retire from hockey, and 37-year-old Niklas Backstrom both ineffective and unable to stay healthy, it’s unsurprising that Leipold’s massive dump truck of money has made a stop at Dubnyk’s house. While a six-year deal seems like an awful lot for, essentially, a half-season of brilliant work, Dubnyk is probably the best option the Wild have for a starting goaltender in 2015-16. 27-year-old winger Chris Stewart, acquired at the trade deadline, could also see a long-term deal from Minnesota. Youngsters Erik Haula and Mikael Granlund will likely see new deals as RFAs. What’s less clear, though, is whether the big deals signed in the past few years will hamstring the team as players like Parise, Suter, Pominville and Mikko Koivu enter their mid-30s. The time for this group to win is, by necessity, very soon, yet it’s not clear that they’re good enough to get there.

Winnipeg Jets

If Nashville’s rise to the third-best record in the West was hard to see coming, just as improbable was the Winnipeg Jets’ return to the playoffs, their first postseason appearance since relocating to Manitoba from Atlanta in 2011. In the three seasons prior to 2014-15, of course, the Jets weren’t as horrendous as their poor results would suggest: their score-adjusted Fenwick over those seasons was a respectable 50.3%, and they ranked 13th in goal-scoring over that frame. Their struggles were actually pretty simple to explain: the Jets were a high-event team playing porous defense – their 56 Corsi against per 60 at 5-on-5 put them 19th in the league from 2011-2014 – and their 0.917 even-strength Sv% ranked 7th-worst in the NHL. As such, only six teams allowed more goals than Winnipeg’s 614.

This season, however, was a very different story. With a Corsi-against rate of 50.1 events per 60, the Jets were actually the fifth-best defensive team in the NHL, and their 0.928 team Sv% was ninth-best in the league. In the four seasons prior to 2014-15, Ondrej Pavelec’s 0.919 Sv% had earned him a reputation as the NHL’s worst starting goaltender, but a 0.930 this year matched his career best, and backup Michael Hutchinson posted a solid 0.924 campaign. As such, Winnipeg tied for the second-fewest goals allowed at 5-on-5 in 2014-15, and ranked 9th in overall goals-against. On offense, the picture was a bit more mixed. The Jets’ rate of shot creation dipped to 55.4 Corsi for per 60 this season after being 8th-highest in the league through 2011-14, and they finished just 16th in scoring (20th at even strength). This lack of scoring punch hurt them in the postseason against Anaheim, as, apart from a four-goal Game 3, they managed just five goals against the Ducks. When Pavelec faltered to a 0.902 Sv% in the series, the sweep was on.

Heading into the offseason, the Jets have a ton of cap space, and are in the fortunate position of having many of their key contributors either signed long-term or on cost-controlled deals. The excellent Dustin Byfuglien and leading scorer Andrew Ladd are heading into contract years, and should be priorities for extensions. Similarly, Grant Clitsome is a year away from UFA, and given his solid 3.6% Corsi Rel, could also be extended now to keep the cost of his contract reasonable. Forward Drew Stafford, a key piece arriving from Buffalo in the Kane deal, is an unrestricted free agent; he did produce 19 points in 26 games with Winnipeg, but his poor two-way numbers (-3.4% Corsi Rel) and age (he will turn 30 early next season) suggest that the cost to sign him on the open market might not be worth it. To my thinking, GM Kevin Cheveldayoff could put that money to better use on a scoring option that’s either younger or more potent than Stafford. Other UFAs include 26-year-old Michael Frolik (who had a far better season driving possession than Stafford, and will likely be cheaper to retain), deadline acquisition Jiri Tlusty (who played poorly in 24 games in Winnipeg, but is still young and talented), and depth forward Lee Stempniak. Next season, key players including Mark Scheifele, Jacob Trouba, Adam Lowry, and Hutchinson will be due new deals, so the flexibility Cheveldayoff enjoys now might be short-lived. The 2016 offseason may also feature a decision on an extension for Pavelec, who will be 28; I’ve been wrong about goalies before (see: Mason, Steve), but assuming the Czech’s play in 2015-16 reverts toward his career averages, and with the highly-regarded Connor Hellebuyck waiting in the wings, Winnipeg may elect to let his deal expire. More generally, I think the Jets have to assume that their goaltending will be closer to league average next season, and that even if coach Paul Maurice is able to maintain the stellar defensive work of the 2014-15 team, they’ll probably allow more goals than they did this year. Where the Winnipeg lineup could really use work, honestly, is on offense. One wonders, of course, whether the trade of winger Evander Kane will haunt the Jets down the road. Although Kane’s situation in Winnipeg had become untenable, its handling reflected just as poorly on the team as it did on the player, and one wonders how much the Jets will regret having traded a 23-year-old with 222 points in 361 career NHL games. But what’s done is done. If the cost of Winnipeg’s newfound commitment to defense is a lower rate of shot creation, so be it, but they’d be well-served to explore the market for a true scoring threat.

Posted in Uncategorized | Tagged , , , , , , , | 1 Comment