Balls and Sticks Podcast

Blog

Balls and Sticks

Posts tagged hockey
NHL Early Career Progression

The Max Pacioretty trade (1:00), Nate Schmidt's suspension (6:03), Blake Wheeler's extension (8:12), Noah Hanifin's contract (10:00), the biggest question facing each team in the Western Conference (11:26), early career progression (27:42).

(You can listen to us discuss this topic in the episode above.)

Here in early September, hockey optimism is at its zenith. Rookie tournaments are happening, training camps are about to open, and soon we'll get to hear about how every player is "in the best shape of my life, for sure." And a large part of this optimism centers around young players, early in their careers. The refrains are familiar: “he had a tough rookie season, but he’ll be better next year” or “the sophomore slump was rough, but he’ll bounce back.”

But how often is this actually the case? How often do players successfully make the jump and improve throughout the first three years of their career?

The population of players I studied here included the 320 forwards who played their first three full seasons (with a season defined as at least 400 minutes of 5v5 play) between 2007-08 and 2017-18. Their points/60 at 5v5 were ranked via percentile, as compared to the other players in that season (e.g., a player’s first year points/60 was compared to all other players in the cohort in their first season, regardless of when that first season was). Lastly, the players were grouped by quartiles in each year.

(Eternal thanks to Corsica for the raw data.)

r = 0.37

r = 0.37

r = 0.41

r = 0.41

r = 0.45

r = 0.45

The three graphs above show the relationships among points/60 during the seasons.

Overall, there was definitely movement among the quartiles during the first three seasons. Only 17% of players stayed in the same quartile for all three seasons, with the bulk of those coming from players who stayed in the bottom quartile. Just 36% of players stayed in the same quartile from their first season to their second, and 39% stayed in the same quartile from their second season to their third. There was slightly more consistency within the top quartile: 39% of players in the top quartile in their first season stayed there for their second, and 47% of players in the top quartile in their second season stayed there for their third. Only 15 players remained in the top quartile for all three seasons, and the list contains the usual suspects: Connor McDavid, Jamie Benn, Brad Marchand, Artemi Panarin, Jonathan Toews, etc.

Specific player data is available in the linked data visualization (if it’s not showing up at the bottom of this page, refresh or click here), with the ability to search by path and by player. Shown below are a few highlights:

2-3-4.png

Some players are able to make continuous progress, from the second quartile in their first season to the third and then the top. This group of players includes Jack Eichel, Aleksander Barkov, and Sean Monahan. (There’s also a pretty decent group of players who jumped from the second quartile in their first season up to the top and then stayed there: John Tavares, Tyler Seguin, Patrice Bergeron, and Evgeny Kuznetsov, among others.)

4-1-4.png

Shown above is the path of players who experienced the sophomore slump: they spent their first season in the top quartile, dropped to the bottom, and then jumped back up to the top. This group of players includes Anders Lee, Dylan Larkin, and Jason Zucker.

Explore more in the visualization below, and feel free to reach out on Twitter with any questions or comments.

Meghan Hallhockey
Individual Power Play Units in 17-18

The losers (2:18) and winners (10:40) of the offseason (featuring Vancouver, Washington, Ottawa, Montreal, Edmonton, New York, Buffalo, Toronto, and St. Louis), an analysis of last year’s top power play units (15:28).

#KarlssonWatch is back (1:05), our picks for next year's breakout players (3:53): Andreas Johnsson (4:30), Casey Mittelstadt (5:15), Dylan Strome (5:45), Travis Konecny (8:19), Alex Galchenyuk (10:08), Ty Rattie (11:52), Jordan Eberle (12:52), more on Mittelstadt because we just can't resist (14:22), our excitement over the season starting (16:11), more power play analysis (17:20).

(You can listen to us discuss this topic in these two episodes.)

My interest in looking at power plays initially spawned from a comment on Biscuits, Dave Lozo and Sean McIndoe's podcast, about how it'd be more intuitive to express power play success in terms of time (e.g., this team scored every eight minutes on the PP) than simply success percentage. Success percentage, the standard measure that most hockey media uses to determine the value of a power play, doesn't take into account how long a team actually spent on the power play. Rate stats (e.g., points per 60) are often better than stats that depend on a variable opportunity denominator (e.g., points per game) since they level the playing field, so to speak, and looking at minutes per goal (or seconds per shot) is just a new way of interpreting the standard goals per 60.

Teams.png

This was fairly easy to calculate on a team level (shown above, 5-on-4 regular season play only from 2017-18), and then I went further to identify the top power play units, those who logged at least 20 minutes together and netted at least one goal. That ended up being the top 98 units, and you can see the details of each by looking at the interactive visualization (or see the bottom of this page).

Units.png

The top 10 units, in terms of time on ice, are shown above, sorted by minutes per goal. Toronto’s first power play unit topped this list, which wasn’t particularly surprising as someone who watched a lot of that power play last year.

(Also of interest to note: all of the top units in 2017-18 used four forwards and one defenseman. In fact, of the top 98 units, fewer than 25 percent used the three forwards-two defensemen structure. Matt Cane has written extensively about the benefit of using 4F-1D over 3F-2D. In summary, it's true that those units tend to allow more shots, but they also generate more shots, and have a higher shooting percentage, which puts the differential in their favor. This seems normal now, but it was under 10 years ago that 4F-1D on the power play was relatively rare.)

StructureIndex.png

Speaking of Matt Cane, my next aspect of power play analysis uses his Power Play Structure Index, which is a weighted average of how far each player tends to shoot from their average location. Please read his work to see the full explanation, but in summary, a lower structure index indicates that a power play operates with a tighter formation, in that each player tends to take shots from a specific location. He also found that it’s a repeatable skill and nearly as predictive of future goal scoring as shot generation. He looked at this metric on the team level and found that over the past few years, the Capitals have performed best on this metric (i.e., had the lowest structure index). I calculated this metric for the individual power play units (those with at least 75 shots), and you can see those data in the interactive visualization (or see the bottom of this page).. The top 10 are shown above, and it shouldn’t be surprising that the Capitals rank high on this list as well.

WSH.gif

And thanks to provided shot locations, we can construct shot maps in order to visualize the structure index, like with Washington's top power play above. Carlson’s on the blue line, Oshie’s in front of the net, Backstrom and Kuznetsov are a little bit all over the place, and Ovechkin doesn’t stray far from the Ovechkin Spot™.

TOR.gif

The same is true of Toronto's top power play unit. You can clearly see Reilly along the blue line, Marner and Bozak mostly on the sides, and Kadri and Van Riemsdyk in front of the net.

CAR.gif

Now contrast those first two gifs with Carolina’s second power play unit, which had the highest (and therefore worst) structure index among the qualified units. There’s a little bit of structure, but as you can see from how spread out the shots are, it’s not nearly as clear-cut as the other two units.

(If the visualization isn't showing up below, refresh your page or click here.)

Meghan Hallhockey
Episode 14: Winner Winner Buffalo Chicken Dinner

The losers (2:18) and winners (10:40) of the offseason (featuring Vancouver, Washington, Ottawa, Montreal, Edmonton, New York, Buffalo, Toronto, and St. Louis), an analysis of last year’s top power play units (15:28).

In this podcast episode, we discuss the top power play units, by team and at the individual line level, from last year. This idea spawned from a comment on Biscuits, Dave Lozo and Sean McIndoe's podcast, about how it'd be more intuitive to express power play success in terms of time (e.g., this team scored every eight minutes on the PP) than simply success percentage. This was fairly easy to calculate on a team level, and then I went further to identify the top ~100 power play units from last year, those who logged at least 20 minutes together and netted at least one goal. (This was my first time using play-by-play data, so please let me know if something looks wonky!)

In the visualization below, you can see each team's scoring efficiency on the power play, in terms of minutes per goal and seconds per shot, along with the top 10 PP units by ice time and the top units by team. (If the visualization isn't showing up, refresh your page or click here.)

Meghan Hallhockey
Episode 13: The Young Guns

Meg's hockey-heavy weekend (0:32), an update on the recent NHL headlines (3:53), incoming rookies who could compete for the 2019 Calder (9:50), a debate over which of last year's rookies will score the most points next year (24:04).

In today's episode, which you can listen to above, we attempted to answer the following question, posted last month on NHL's Instagram:

NHL.png

My instinctive answer, without looking at any underlying data, was Calder winner Mat Barzal. Barzal led this group in overall points, at 85, with Keller second at 65 (though Boeser was on pace for 73 points if he hadn't missed 20 games due to injury). Connor had the most goals with 31, though again, Boeser was on pace for 38—which would have put him at close to top 10 in the league.

But let's take a look at the numbers from last year, starting with their performance at even-strength.

Data from Rob Vollman.  All data apart from SH% and IPP% are per 60 minutes.

Data from Rob Vollman. All data apart from SH% and IPP% are per 60 minutes.

Shown above are points, primary assists, goals, expected goals (shot attempts weighted by location), unblocked shots, shooting percentage, and individual points percentage (the percentage of on-ice goals to which the player contributed a goal or assist).

DeBrincat, Connor, and Boeser were the strongest in terms of getting off unblocked shots, and unsurprisingly, those three also led the pack in goals and shooting percentage. The average shooting percentage for forwards in the league usually hovers around 10 percent, though some players are capable of sustaining an above-average percentage over their career (exhibit A: Steven Stamkos). I am very curious to see what Boeser's shooting percentage looks like next year.

Data from Rob Vollman.  All data apart from SH% and IPP% are per 60 minutes.

Data from Rob Vollman. All data apart from SH% and IPP% are per 60 minutes.

Shown above are the same metrics as before, but this time on the power play. Barzal, Keller, and Boeser spent the most time on the power play, in terms of average TOI per game. (For context, both Barzal and Keller logged approximately 250 power play minutes.) Boeser was the most successful, with over three goals per 60, and his 7.6 points per 60 on the power play was good enough for 11th in the league, among skaters who played at least 100 minutes. He scored over 40 percent of his points last year on the power play, and he spent nearly 80% of those minutes playing alongside Daniel and Henrik Sedin.

Shown above are percentiles for several pass-based metrics, from Ryan Stimson's Passing Project. Some definitions:

xPrP60: Weighting of primary shot contributions on likelihood of becoming goals/assists
ixA60: Likelihood of an assist accounting for preshot movement
PSC60: Primary shot contributions (shots + shot assists, which are passes that lead to a shot)

Barzal looks pretty good across the board here. Boeser is lower on the assist metric, which makes sense given his lower assists per 60. (Boeser collected over half his points on goals, compared to around a quarter for Barzal. Boeser ranked in the top 15 in the league on that measure, among skaters who scored at least 50 points.)

To sum it up, I believe it'll be close, but I would still go with Barzal (you can hear Hannah's pick by listening to our episode above!), even though his upcoming season will certainly be affected by the departure of John Tavares. They didn't play on the same line last year at even-strength (Tavares centered the first line with Josh Bailey and Anders Lee, Barzal the second with Jordan Eberle and Anthony Beauvillier/Andrew Ladd), but they did share time on the power play—nearly 90% of Barzal's PP minutes were with Tavares. Barzal will likely get more minutes in the first-line center role, although he averaged only about a minute less than Tavares last year, but he might also face tougher competition.

All of that said: it wouldn't surprise me at all if Boeser edges out Barzal. He will likely be playing with the same linemates as he did for most of last year (Bo Horvat and Sven Baertschi), and if he stays healthy all year, keeps that shooting percentage above average, and finds a home on a productive power play, he should have an impressive year.

Meghan Hallhockey
Episode 10: Hit Me Baby One More Time

(You can hear us talk about this topic in our July 24 episode, at the 11:16 mark.)

When Hannah and I decided that we were going to talk about fighting and physical play on a hockey episode of our podcast, I knew that I wanted to look into “hits” as a statistic. Hits don’t garner much respect (as a statistic, that is—as a style of play, that’s a whole different story, depending on who you're talking to), which is for two main reasons.

  • People often consider hits to be a positive thing (such thoughts are usually accompanied by a lot of vague qualifiers like “physicality!” and “gritty, high-energy play!”), but that’s not really true. Teams often do better when they aren’t hitting as much, since by definition you don’t have possession of the puck if you’re the one doing the hitting. And as far as I’m aware, high hitting rates aren’t particularly associated with strong defense.
  • The definition of a hit is fairly murky, which leads to more subjectivity among the scorers at each NHL arena.

The timing could not have been more perfect because just as I was thinking about this, I was working through my copy of Stat Shot (highly recommended, by the way!) and came across their chapter on hits. I really appreciated their process, as they talked through hits and various ways to make it more useful as a statistic, and I decided to replicate their work for the most recent 17-18 season. Since I’m a hockey analytics beginner, it was a useful exercise!

I’m going to walk through this process at a fairly high level, but to get the most details, you should definitely pick up a copy of Stat Shot. To see the data visualization, explore below or click here. (If it isn't showing up at the bottom of this page, refresh!)

  1. As is customary, we’re looking at even-strength situations only. Power plays and penalty kills are different beasts, so we’re eliminating differences in play in those specialized situations.
  2. The first step is to tackle the possession issue. Some players may hit a lot, but that isn’t as valuable if they have low possession numbers. We can adjust the hit total with Fenwick (also known as unblocked shot attempts), one of the most common proxies for puck possession.
  3. We also need to address the potential sample size problem by taking into account time on ice. The cutoff used here was 800 even-strength minutes: anyone with fewer minutes had their hits adjusted by taking a weighted average of their hits, relative to their time on ice, and the average amount of hits for their position.
  4. In order to (partially) eliminate the subjectivity problem in actually recording the statistic, we can look at the number of hits registered for each team at home versus on the road for the past few years. Using average possession time, we can determine how the actual hits recorded at home correspond to the expected values each year, and then use a weighted average to calculate an overall “bias factor” for each team that’s applied to their player’s hits.
  5. Lastly, we convert the adjusted hits to a per-60 rate.

I was also personally curious as to what the scatterplot looked like between this adjusted hitting rate and points per 60 (which was also adjusted to take into account TOI using the same procedure as above). In the visualization, I added dotted lines to represent the top 20 percent for both adjusted hitting rate and points per 60. It’s fascinating to see who shows up in or near that quadrant (e.g., Evander Kane, Dustin Brown, David Backes).

You can see all of these numbers in the visualization. The first tab shows the scatterplot (with the ability to find a particular player and choose how you want the data to be color-coded), and the second tab shows the summary per team. (I did restrict the games played to 30, just for simplicity’s sake. And due to my time constraints, the underlying data aren’t as accurate as they could be when it comes to players who spent time on multiple teams. Players were assigned to the team for which they played the majority of their games.)

Meghan Hallhockey