Inspired by @IanGraph‘s recent realization that NHL.com publishes data shots that hit the post and crossbar, this post will investigate the leafs who most frequently break our hearts by hitting the post or crossbar. The data includes the last 3.5 seasons, and includes those who have played more than 20 games for the leafs over those seasons. You can find Ian’s post on which leafs are exceeding expectations here.

As a disclaimer, this is more of a fun “post” (lol!) in that hitting the post is almost certainly not a repeatable “skill” so we’re more just observing some interesting totals. Enjoy!

First we look at aggregated totals over the past 3 seasons – and Kadri leads the bunch again with 21 post or crossbar hits over that period. For context, that’s nearly as many goals as Jake Gardiner has had (23) over that period – so Kadri would certainly win the “How did he miss that?” award for the leafs. Let’s face it, we all saw flashes of Vasilevskiy coming across to rob Kadri, or a devastating *ping* on his backhand goal last night. We also see some of the leafs snipers rise to the top of the chart as well in Matthews, Nylander, and Rielly, likely due to higher shot volumes from these players.

On a per shot basis – we see more of the 4th line type in Frederik Gauthier  rise to the top. For every 100 shots Frederik Gauthier takes (theoretically. He’s only taken 41 shots over the years according to our data), he hits the post 7 times. The rest fall into the 3-4 range per 100 shots – finally concrete evidence that hitting the post is not a repeatable skill!

Next we’re going to shift to misses – either wide or high – over these seasons. Again, Kadri tops the list in aggregating totals- followed by a few other leafs regulars in Rielly and Gardiner.

On a per shot basis, which again is more useful given the changes to the leafs roster, we’re seeing a lot more 4th liners and 3rd pairing players in our list. Par Lindholm appears particularly bad at shooting – he tops the list and is the only forward in the top 10 besides Leivo (who I recall frequently blasting wristers over the net for a nice breakout for the other team).

When we look at forwards only, we get a better idea of just how bad Lindholm is at shooting – he’s way more likely to miss the net when he ratchets up a wrist shot – and is probably better off passing to his linemate in Gauthier who will almost certainly hit the post (but will probably dump it in the corner).

In a season where a lot has gone right, leafs fans have found no shortage of annoyances between backup goaltender play, Nylander’s slow start, or the right defensive side’s difficulty driving play. Today we will briefly discuss the leafs recent Powerplay struggles, and why they probably didn’t need to switch things up as they did against the Devils on Thursday.

The leafs sent out the following units in that game – however things did get a little jumbled at the end – when Matthews stayed out on the ice with the top unit and generated a few grade-A chances.

PP1: Tavares, Marner, Kadri, Rielly and Kasperi Kapanen

PP2: Matthews, Patrick Marleau, William Nylander, Andreas Johnsson and Jake Gardiner

To start, both units looked like they had never played together (probably because they never played together) with a few failed zone entries against a top New Jersey penalty kill. Things looked pretty bad until Marner finally was set up on the half wall and made some dangerous passes across the crease. While I don’t think these units would be particularly bad (and I do like seeing Kapanen on one of the units finally), the change may have been unnecessary – as the following charts will attempt to depict.

The charts are broken into 10 game chunks of the leafs season to observe changes over time. Looking at the raw totals – we see where the narrative that the leafs Powerplay is slowing down could be emerging:

However, on a per minute basis – which makes more sense given the leafs have also seen less PP time as the season has wore on – things look slightly different. We see much more fluctuation between a very hot start, a cool off period, a return to form, and another drop off.

What might be causing this? A few things: we may see the leafs struggling to generate more scoring chances, which would definitely be a reason to make a change. Or it could be a more boring answer (but one that I think is more likely), and the change is driven by variance in shooting percentage. This was certainly a point brought up by Tavares in a post-game interview after the Predators lost who brought up that the PP has to be better, but posts and and near-misses have also played a role.

The chart below supports this theory, with the leafs goals scored per minute closely mirroring the leafs shooting percentage on the Powerplay.

This means that the leafs could be generating less scoring chances on the Powerplay, and because of this we are seeing their shooting percentage drop. That would be a very good reason for Babcock to make a tactical change to generate some better looks, but it may not be true.

Looking at these two charts, we see the level of offense generated by the Powerplay over our 10 game chunks. They have continued to generate roughly 2 shots towards the net for every Powerplay minute, and have actually generated more High Danger chances per minute than they were early on in the season when they were on fire from a production standpoint.

It’s hard to evaluate a Powerplay without analyzing game footage, and at times the leafs Powerplay has seemed to lack some creativity, but it seems mostly a product of a fluctuating shooting percentage based on the high-level numbers. We will see if  Babcock sticks to his guns on the PP changes, since they may have been unnecessary in the first place.

Phil Kessel’s 2014 season continues to stand as the best production we’ve seen at 5v5 over the course of an 82 game season, but the new core is hot on his heels. This may come as a surprise to some, but the leafs stars haven’t fully clicked for a full season until this year. Mitch Marner’s slow start in 2017 seems like a distant memory, Matthews injury stints have made it difficult to amass full season accolades, and Tavares is in his first season as a leaf.

This year has been different, at least for Tavares and Marner, whose prorated even-strength numbers will surpass Phil if they continue at the current pace. This obviously doesn’t account for powerplay time intentionally – which will further inflate totals across the board (particularly Matthews). Perhaps we’ll take a look at powerplay totals in another post! We actually see a lot of 2018/2019 season in the chart below – a sign that things have changed in leaf land.

Data is from Corsica Hockey – chart is the top 10 5v5 points going back to 2008.

The leafs were downed by the Minnesota Wild on January 3rd despite their late game push. While the leafs created tons of scoring chances in the game, many of which were thwarted by Devan Dubnyk, the Wild once again made the leafs defense luck suspect on multiple occasions – a recurring theme this season. The blemishes look worse without Andersen between the pipes – so many suspect the leafs will be active at the trade deadline to beef up their defense, especially on the right side.

One of the rumored additions is Alex Pietrangelo of the St. Louis Blues, a two time second team allstar, who would certainly bolster the right side. There are certainly questions about his value in a trade – he is 29 years old and his contract will be up at the end of the 2020 season. The leafs will likely not be able to afford his next contract, one in which he will likely be overvalued heading into his 30s.

The following charts illustrate where Pietrangelo would fit on the leafs D corp – aggregating on-ice expected goals from the past 3 seasons. It’s important to view these charts with context – Connor Carrick often finds himself at the top given his sheltered role on the 3rd pairing during his leafs career.

Pietrangelo ranks 3rd behind Carrick and Gardiner – but the upgrade is obvious – the leafs top 2 righthand D in Hainsey and Zaitsev rank at the bottom.

Looking at expected goals against, Pietrangelo would be the best leafs D in the top pairings (removing Carrick and Polak). Gardiner and Reilly’s high event play-style has them at the bottom of the rankings.

Finally, looking at both xGF and xGA – Pietrangelo’s more stable game results in less goals both for and against his team while on the ice. His overall effect is certainly positive using these metrics, and would be a nice addition to a team struggling to cover the opposition in it’s own end.

Source: Evolving Hockey

The Leafs had a hard time at home against the Islanders last night, evidenced by the 4-0 scoreline and normalized expected goal totals in the chart below:

John Tavares tallied 2 goals tonight against the Blue Jackets – bringing his total to 25 on the season and putting him on pace for 56 goals on the year in what would be a career high. The following graphs display his 5v5 statistics, and just how much he’s produced, relative to the rest of his career. Charts do not include data from tonight’s (12/29) game and the Points and Goals charts prorated each season to 82 games.

 

 

The leafs entered the Christmas break on a 4 – game win streak against mostly below average teams, scraping out a win against the Red Wings on Sunday night. The recent streak is a nice characterization of the leafs half-season to date – supported by great results but leaving us wanting more. The leafs have left some questions about their ability to drive play with sub-par possession statistics , and ability to contend against Boston and Tampa when they aren’t supported by dynamite goaltending through Freddie Anderson.

These narratives seem to indicate us leafs fans are a truly miserable bunch! The team sits second in the league behind Tampa and has a 97% chance of making the playoffs according to Moneypuck.com – what would have been a dream only a few years ago has become an annual reality. To offer some positivity going into the holiday season – this post will focus on the leafs contract bargains and some players who have surprised based on their price tag.

The Leafs Bargain Bin

To pick out the leafs 2019 bargains-to-date, we’re going to focus specifically on point production metrics, and how it relates to a player’s cap hit. Obviously there are plenty of other metrics to look at – especially with bargain players who may put up less points and do more of the “little things – but this analysis will focus on measuring how much players have contributed on the score sheet.

Figure 1 above depicts the leafs Points per 60 minutes (P/60) plotted against their cap hit in millions. While I want to assume that Trevor Moore will score 9 points per 60 minutes for the rest of the season, players with less than 5 GP are excluded (but I liked his game when he did play). Taking a look at the players above 3 points per 60 minutes – the results tend to support the narrative so far this season. Marner and Matthews, both still on their ELC deals, have produced at a very high rate relative to their cap hit (and have been forced to join the Nutcracker to pay the bills). Tavares has been as advertised – coming at a high price but producing at a high rate.

The second tier – between 2 and 3 points per 60 minutes – features some of the leafs bargains we will take a deeper look at. Kasperi Kapanen, Andreas Johnsson, and Tyler Ennis have all produced at a high level for under $1 million. Morgan Reilly also finds himself in that tier – his hot start and ability to produce elite offense from the backend has been well documented.

Figure 2 shows players’ Points per 60 minutes divided by their cap hit to illustrate who is giving the leafs the best production at the lowest cost. Figure 3 shows pure production without accounting for TOI – Matthews drops down due to his injury and Marner looks even more impressive – but the bargains are pretty consistent. Again, barring the two superstars, Ennis, Johnnsson, and Kapanen stand out as great value.

Tyler Ennis – The refurbished on-sale laptop

Ennis’s career looked like it may have been over following difficulties with concussions in 2015/16 and troubles finding his footing with Minnesota the next year. The leafs saw a savvy veteran player who could offer some upside if he hit his stride. He was doing just that when he unfortunately fractured his ankle against the Rangers last Saturday – but we’ll cover here just how good he has been.

For context, Ennis put up 22 points in 73 games with the Wild the season prior – not great production but someone you can plug on your fourth line. However, on a per minute basis – he’s starts to look more like the player he was in Buffalo when he put up 20 goals (Figure 4). We’ve seen this in flashes during the course of the season – his goal against the Blue Jackets comes to mind – and the leafs should be happy he’s surpassing his recent totals.

In comparison to the rest of the league, we see a similar story. On raw point totals – Ennis ranks 224th at 5v5, which is bottom level 3rd line material – certainly fine for a player making $650K. Again switching to P/60, his 1.98 is good enough for 115th – which is more like 2nd line production. It’s not necessarily the case that Ennis deserves more minutes – we saw him struggle to produce on the Matthews line – but his ability to put up points in the role he was given shouldn’t be looked over and the leafs will count on signings like this one when the cap situation gets tighter down the road.

Kasperi Kapanen – a new house

Kasperi Kapanen is a new house – you expect his value to rise but you don’t know for sure what you’ve got until you spend some time with him. The leafs didn’t pick Kapanen out of a bargain bin, at least not from a contract perspective, but he makes the Kessel return look pretty strong. Rather the leafs signed him 3 years ago when he was still in the AHL and a huge uncertainty. This year he has surpassed the expectations of leafs fans by showcasing his elite speed but also his playmaking ability and work in the corners.

His production is at first line levels in both raw and P/60 totals – he ranks 46th in 5v5 points and 62nd in P/60 – and he’s proved resilient enough to stay on the top line while Nylander figures out his game. He played strong without Matthews as well – giving Mike Babcock a slew of options to rotate the lineup when he sees fit.

Andreas Johnsson – A broken TV

Leafs fans had high hopes for Johnsson coming off a strong playoff series against Boston last year – and he disappointed out of the gate. Just like that TV you bought that you immediately had to return – Johnsson spent a handful of games out of the lineup and took awhile to get his legs and his game going and we’re seeing the results start to surface.

Now, Johnsson is producing at the border of 1st and 2nd line rates – he ranks 118th in raw Points, and jumps to 68th in P/60 – just behind Kapanen.

A lot of leafs fans were hoping to sign Johnsson for longer than the 2-year contract he ended up with. After the early hiccup, it looks like he was oversold – but now that he’s hitting his form and the points are following – it will be another interesting contract for the leafs to handle at the end of 2019.

A topic of growing interest in the NHL is the recent influx of players with high-end talent. While hockey is known for being a team game, where individual players tend to have a more marginal impact on results than other sports, team models that emphasize retaining a core of high-end talent have had success and is becoming the blueprint for cup contention. Namely we’re thinking of the Penguins and Blackhawks, who’ve together have won 6 of the last 10 Stanley cups.

This isn’t necessarily anything new – teams have always needed high end talent to win a cup and they always will need it too. However, these teams have shifted to a model that places more importance on players at the top the skills distribution at the cost of depth. The Blackhawks were forced to let good players go to keep their core intact, while the Penguins paid handsomely for a star-studded forward group at the sacrifice of a deep d-core.

How exactly this plays out for the leafs moving forward will be interesting to see. The team is still looking to sign Nylander this year, and have negotiations with Matthews and Marner coming up next year. It’s highly possible they will need to let Gardiner walk, and as the team progresses, they may lose out on depth players that are good, not great, and thus not a part of the core. Edmonton is finding themselves in a similar situation with McDavid, Draisaitl, and Nugent-Hopkins but have a few more bad contracts that could crunch them.

So this begs the question, have teams responded to the success of this new model by paying more for superstar players? We’ve seen McDavid be rewarded with a 12.5M contract – how has this shaken out with the rest of the NHL?

The Gini Coefficient

To determine whether or not inequality is growing in the NHL, we are going to use the Gini coefficient, which provides an index we can use to measure inequality by looking at income distributions.

The Gini coefficient is a way of comparing how distribution of income in a sample compares with a similar sample in which everyone earned exactly the same amount. Inequality on the Gini scale is measured between 0, where everybody is equal, and 1, where all the country’s income is earned by a single person. In our case, we will compare income distribution between NHL teams to see which are more equal, and which are more top heavy.

More details on the Gini coefficient can be found in the Methodology section at the end of this story.

An Example

Using this calculation, let’s take a quick look at some of the more equal and unequal teams in the NHL to see if the index is properly capturing inequality. Looking at the 2017-2018 season, the last full season, we can see the most unequal teams based on the Gini coefficient are Chicago, Pittsburgh, Dallas, Washington, and Nashville:

This certainly passes the eye test that the Gini is accurately finding teams that are top heavy – our Chicago and Pittsburgh examples are both included, along with other teams with superstar talent – Dallas, Washington, and Nashville. Let’s take a quick peak at the top of their distributions.

These teams all have high salaries at the top of their distribution, and tend to have more players on ELCs and cheap contracts at the bottom. Another way we could look at this is the % of the CAP the top players are taking up – but we’ll leave this methodology for a separate blog post and focus on the Gini for now. We also notice that these teams span decent to great, not surprising considering their top end skill – but supports our working hypothesis that teams with a higher Gini are more successful.

The bottom of the distribution is a little more interesting – we find the Islanders, Ducks, Red Wings, Leafs, and Devils at the bottom:

As expected for the Gini bottom feeders – we’re seeing contracts at the top of the distribution that are lower than the more unequal teams – topping out around $6M. Two other interesting points – what seemingly makes the NYI the most equal is their lack of ELC’s coupled with the smaller top end contracts. Around 2/3rds of the team makes over $2M – while teams like Toronto and Detroit are relying more on entry – level deals.

Do Unequal Teams Make The Playoffs More Often?

The table below shows the data for the 2018 season – and I’ll now take the time to take a descriptive look at how inequality may impact results. Using playoff qualification as a sort of crass measurement – we see there isn’t a clear relation between Gini and qualifying for the playoffs. There is a dynamic evident between teams who are further into their development and have had to pay out large contracts to their core. Gini doesn’t seem to impact success with making the playoffs, perhaps because having young players on their ELC is beneficial, as is having top end talent resulting in a higher Gini. There are a lot of factors here we are passing over – how much cap space teams are using to name just one – and the analysis would benefit from a more statistical approach which we will save for the next post.

Back to the Question!

Finally, we will look to answer the question we set out at the beginning – has the NHL become more unequal in recent years? Compiling data from 2014 to 2019, the average Gini coefficient in the NHL increased steadily from 2014 to 2018, before dropping back down to 2014 levels this year.

This result is not too surprising considering the variety of factors that could be impacting results. One, as good teams become great and pay out contracts for their top players – we also expect rebuilding teams to filter to the bottom of the list as they rebuild and shed those large contracts. And second, our hypothesis was that high-end skill would claim more money – but it’s also possible that the bottom of the distribution has become more skilled, offsetting any salary gains at the top of the distribution.

Next Steps

It’s difficult not to caveat such a high-level analysis, but in our next post we will look to apply some statistical methods that may give us a better answer in our inequality influences results.

Methodology

Salary cap data was pulled from Cap Friendly, and Corsica Hockey was used for accurate team names for 2014 – 2019.

Gini coefficient explanation: The Gini coefficient is computed using a Lorenz curve, where population is plotted on the x axis and cumulative share of income is plotted on the y axis. The “Lorenz curve” represents the plotted points of the actual distribution in question, and the perfect distribution line represents a society where everyone earned the same amount.

Finally, we calculate the Gini coefficient by taking the area underneath the actual income distribution and the line of perfect income equality. In the diagram pictured, this is A/(A+B)

The coefficient doesn’t capture very explicitly changes in the top 10% – which has become the focus of much inequality research in the past 10 years – an area we may be able to further explore after this analysis

Hi everyone, and welcome to the Mackinaw Stats hockey analytics blog. I plan on using this blog to conduct analysis on hockey data, with a focus on data from the National Hockey League. As the title suggests, the stories I plan on writing will sometimes focus on the Toronto Maple Leafs, but will always entail answering a question about hockey that I will look to answer with data.

The title of the blog is dedicated to the beloved Joe Bowen, the voice of the Toronto Maple Leafs, who made famous (to leafs fans anyway) the phrase “Holy Mackinaw!”. Bowen will be inducted into the media wing of the Hockey Hall of Fame tomorrow, and was honored at the game last night against the New Jersey Devils. As his broadcast partner Jim Ralph recounted last night, “No one has bled blue more than Joe.”