Advertisement

If the Toronto Maple Leafs end up searching for a rental option on the right side of their defence, they may find that all roads lead to Ottawa — or possibly to the desert.

This is a continuation of my series on player trade evaluations, ranging from the Leafs ‘second-tier’ wingers here and here to the most talked-about top-four options, Matt Dumba and Josh Manson.

My previous reviews focused on the numbers, using tools like my SKATR interactive visualization, which resides here at MLHS and the heap maps of hockeyviz.com (behind a paywall). SKATR’s glossary and guide can be found at the link if you have any questions.

The Search

Miner Background

Searching for a player who meets several criteria among hundreds of defensemen can take a very long time, or you can jot down the names that come up in trade discussions and hockey media. I decided to do a little of both and see what I could find out.

The main tool in my search was Miner, a part of the SKATR package. It allowed me to zero in quickly so I could spend more time investigating qualified players based on their five-on-five statistics. I’m not sure what tools teams use; I suspect a team like Toronto keeps up-to-date player files based on lots of scouting and analytics research.

An example of SKATR Miner starting with a large subset of RD defensemen across 24 statistics.

Criteria

While the search starts with a long list, it’s possible to narrow that down fairly quickly using some of the filters in Miner. Here is a list of the main criteria I used in my search:

  • Cap hit under $5 million
  • Contract expiry 2020-22
  • Age
  • Handedness
  • Relative to teammate expected goal suppression (defense)
  • Relative to teammate expected goal share (two-way play-driving)
  • Ice time and percentage of ice time versus elite forwards
  • General availability given team playoff chances, etc.

However, once I found some candidates, I also forced players onto the list because Toronto fans and insiders were throwing their names around — players like Travis Hamonic, Sami Vatanen, Alec Martinez, and Colin Miller. I wanted to make sure that I at least looked at a few of those names, just in case I was missing something.

This list of eight defensemen in the chart below isn’t perfect — it isn’t meant to be.  It’s a hybrid between players receiving some rumour buzz and players who met my personal criteria.  For this exercise, it’s also important to note I used Miner’s 2019-20 player data. I could have selected 2018-20 combined totals as a means to stabilize some of the performance data, but with the historical heat maps as a second resource, I chose to highlight their current performance with Miner.

Your mileage may vary. That’s why SKATR is interactive. That’s also why they have GMs and scouts instead of a supercomputer that rings a door chime when the player is found.

‘Drill-Down’ Chart

This is simply a customized, simpler version of Miner.

The numbers in each square are percentiles from SKATR; meaning, they are the player’s five-on-five 2019-20 percentiles, ranking from 0 (worst) to 100 (best) among all NHL defensemen with at least 100 minutes of ice time. The percentiles are calculated after adjusting for score and venue and converting raw counts to per-60 rates.  For this ‘drill-down,’ I have selected a sample of 12 of the 24 SKATR statistics. They are arranged across the top in three sections: Box Score, Context, and Relative to Teammates.

The columns I focused on the most are “Percentage of TOI vs Elite” forwards (competition faced), “Expected Goal Share” relative to teammates (i.e. play-driving, aka Rel TM xGF%), and “Expected Goals Against” relative to teammates (i.e. defense, expected goal suppression, aka Rel TM xGA60).

The SKATR guide has a section on Relative to Teammate statistics. It’s worth a read if you aren’t clear on them.

SKATR ‘Drill-Down’ is a custom version of Miner not available online.

Key SKATR/MINER Takeaways

  • The Senators’ Dylan DeMelo has had a standout season when it comes to driving play (Rel TM xGF%, 100th percentile) and suppressing expected goals against (92nd percentile). Brenden Dillon in San Jose, though left-handed, ranks next.
  • The Coyotes’ Jason Demers has faced elite forwards the most often as a shutdown defender (89th percentile) while still ranking high on defense (Rel TM xGA60 65th percentile). Adam Larsson and Sami Vatanen have played shutdown roles and deserve more investigation.
  • Travis Hamonic in Calgary has been able to drive on-ice offense (Expected Goals For relative to teammates or Rel TM xGF%), but Hamonic’s defensive contribution has been weak this season; relative to teammate expected goals against per 60 is down at the fourth percentile.

We can cull down the list, but let’s do a little more exploring first by looking at their heat map results.

Heat Maps

Background

Isolated heat maps use advanced math and statistics to estimate the rink map threats associated with a player after accounting for teammates, competition, and other factors.

One thing they also do is blend some of the prior season into the current season’s numbers to help stabilize the estimates (a Bayesian method for more technical readers). Therefore, the 2019-20 map is not strictly this season.

[For those interested in the unique math involved, here’s a link. Those who understand regression and math will see that the Gaussian functions (basically, the coloured areas in the rink maps that are derived from shot location probability densities) are actually used as variables in the regression model.]

Heat Map Threats Summary

The table below gives the 2019-20 offensive and defensive zone threat impacts as percentages above or below league average. Negative percentages in the Defensive Threat column represent improvements in shot threat suppression.

For reference, here are the Leafs right-side defender threat percentages:

  • Justin Holl: +2.4% on offense, + 9.4% on defense
  • Tyson Barrie: -2.9% on offense, + 15.6% on defense
  • Cody Ceci: -10.5% on offense, +4.4% on defense

Although there is other evidence that results have improved under Keefe, I can’t separate out the Babcock and Keefe results here. Still, you can see why Kyle Dubas would be placing a premium on a right-handed defender with term.

Before reviewing the results in the table, let me show an example of where these threat numbers came from. Here is what the maps look like for Edmonton’s Adam Larsson.  His defensive zone heat map is at the bottom. Larsson is estimated to have reduced the threat on his own net by a sizeable -8.7% in 2019-20.

Red areas represent higher shot threat locations, while bluer areas represent reduced threats. Right-handed Larsson has his greatest impact in front of and to the right of his goaltender.

Key Takeways from the Heat Map Threats

  • Martinez, Hamonic, Vatanen, and Miller have actually allowed a higher threat level threat against their net when they are on the ice at five-on-five. These may be trade options for some teams, but the Leafs are ideally seeking more evidence of defensive support on the right side. I have seen reports that Vegas has expressed interest in Sami Vatanen. I would pass on him, I think.
  • Of the four who have shown a defensive impact, 29-year-old left-hander Brenden Dillon and 27-year-old Adam Larsson have the most impressive threat reductions (-10.6% and -8.7%), followed by Jason Demers in Arizona (-5.5%) and the Senators’ Dylan DeMelo (-1.1%).

Narrowing Down the Targets

For the purposes of this exercise, let’s eliminate the four who haven’t shown an ability to suppress shot threats in 2019-20 and zero in on those remaining.

Brenden Dillon – LD – SJ – $3.27M – Expiry 2020


Reports suggest that Kyle Dubas much prefers a trade that helps the right side of his defense — ideally with some term. That would disqualify left-hander Brenden Dillon, at least for this article on right defense options.

However, Dillon’s 2019-20 performance, at least defensively, has made hockey people stand up and take notice. The Boston Bruins are reportedly interested in him. Dillon is also a pending UFA and appears to be in line for a raise in his next contract. San Jose may choose to get some trade deadline return for Dillon since they are out of the playoff race with only 48 points in 52 games.

Here is Dillon’s SKATR Solo profile for the combined 2018-20 seasons and his heat map history. He has a cap hit of $3.27M and has size. If the Leafs or another team need a left-handed rental, this one would be high on my list given his two-way play driving and relative to teammate defense.


It looks like Dillon has had two good defensive seasons in a row.

More on Adam Larsson and Jason Demers

Moving down the list, Adam Larsson of Edmonton and Jason Demers of Arizona have played tough shutdown minutes, have played reasonably well defensively, and are relied on higher in the lineup. Right now, both of their teams are fighting for a playoff spot.

Playoff Probabilities – Jan 29, 2020

MoneyPuck

When I highlighted Matt Dumba and Josh Manson in my last articles, we all knew that their teams were further off the pace. I have trouble imagining either of these two teams being interested in a trade right now because the hole on the right side would hurt their chances significantly.

Another factor that has to be given some weight is the SKATR results outside of defense. Larsson (seventh percentile) and Demers (12th percentile) are ranked very low on Game Score this season — a metric that although it is weighted towards offense, still captures other factors.

Both also have negative offensive threat estimates in the heat map summary. Adam Larsson, in particular, nets out with relative to teammate expected goal share and shot share rankings that are sub-par in 2019-20.

If I look at their SKATR profiles for the combined 2018-29 and 2019-20 seasons, I get a more stable view of what they have been doing over the past one-and-a-half years.

There are a lot of similarities. As shutdown defensemen who face a heavy load of top competition, their general results are not that impressive — especially the key relative to teammate expected goal metrics, which are below average.

All things considered, I would not rate these two as the most likely top trade options at this point in time, although Jason Demers could certainly deserve a look if the Coyotes fall a little further out of the playoff race.

Just in case that happens, let’s take a moment and look at Demers a little further. (I don’t see Edmonton falling out of the race anytime soon.)

Jason Demers – RD – ARI – $4.5M – Expiry 2021


Below are his heat maps for the past four seasons. You can see that he has contributed defensively every year — that’s notable given the level of opponents he typically faces in a shutdown role.

At the same time, his negative impact on offense is apparent. This is something we often see in defensive defenders. The question is: How much would his defensive approach to the game hurt the offense of his team? That, in a nutshell, is why a player like Demers is usually best partnered with a more offensive partner — for example, Morgan Rielly.

On the penalty kill, Demers leads with 2:22 minutes per game, but his +.57 xGA60 is not as strong as the -2.16 and -1.57 results of Jakob Chychrun and Niklas Hjalmarsson.

Demers’ cap hit is another issue that would need to be addressed. At $4.5 million, he has an identical cap hit to Cody Ceci. Unless there is salary retention, the Leafs would only have room for one of the two when Morgan Rielly returns; Ceci would either need to be sent to Arizona as part of the deal or be traded elsewhere. However, even if Cody was traded out, Demers’ cap hit would be barely manageable next year given their other cap pressures.

Let’s keep the veteran Demers on the back burner for now but keep an eye out should the Coyotes falter. He has an eight-team No-Trade list, and the Leafs are probably not one of them. You may recall that they missed out on a chance to acquire him a few years ago when he was playing for Florida.

That leaves one player unaccounted for: Dylan DeMelo of the Ottawa Senators.

Dylan DeMelo – RD – OTT – $900K – Expiry 2020



Dylan DeMelo‘s name has been brought up in the Toronto sports media, and rightly so: He is a 26-year-old right-handed defenseman, has a cheap $900K cap hit, and he has put up some good play-driving numbers. DeMelo is also a pending UFA, so the Leafs’ preference for term would not be met. That said, he is an intriguing and inexpensive option on a tanking budget-conscious team that has traded before to gain a Toronto second-tier winger (Connor Brown).

When San Jose traded for Erik Karlsson, the sixth-round pick from the 2011 draft was considered a minor piece. Perhaps Pierre Dorion did his homework on the 6’1″, 191-pound rearguard from London, Ontario. In any event, DeMelo is playing well on one of the league’s weakest teams.

SKATR

DeMelo is a 20-point-a-year defenseman. Below his SKATR profile for the past two seasons. Keep in mind that this is a 5-on-5 report and the bars are percentile rankings among NHL defensemen based on rate statistics — e.g goals per 60 minutes — not just the number of goals. The 2019-20 profile on the left will have the same percentiles as those reported in the ‘Drill-Down’ chart shown earlier.

While his offense is mixed, his very high relative to teammate numbers this season point to a player who is driving two-way play and suppressing goals and shots in his own end. This is a significant improvement over last year. It’s possible that the weak team around him has contributed to these results, although there are small adjustments for team strength in the calculations.

DeMelo Heat Maps

DeMelo has also started from his own end more than most defensemen (OZ start ratio 31st percentile) and faced elite forwards at an average level (52nd percentile). Over half of his minutes have been spent with Mark Borowiecki as his partner.

Below we can see DeMelo’s isolated heat maps for the last five years. They describe a player who had a bounce up in offensive zone threat this year (+8.1%) while simultaneously improving his defensive zone results (-1.1%). Whether that projects into the future is uncertain given that his 2018-19 results were almost the mirror version of the previous year.

Overall, his two-way improvement this season is driven by increased shot threats in the offensive zone. The defensive zone result is less enthusiastic than the relative to teammate expected goals against numbers mentioned earlier — perhaps in part due to the partial blending/smoothing of 2019-20 and 2018-19 shot threats.

Lastly, DeMelo’s penalty kill numbers are only so-so. He averages 2:30 minutes per game and has allowed more expected goals against than most of his teammates (xGA60 Rel = +.94).

Dylan DeMelo Clips

Now let’s take quick a look at DeMelo in action. There’s a lot to like in the way he has played defense for the Senators this season.

As Justin Bourne pointed out, this is a smart play by DeMelo to shift over quickly to the left boards from his right side and nullify a rush.

Here is a perfect stretch pass from the corner:

And here is some excellent defensive positioning and use of his stick to disrupt the Montreal forward:

Trade Options

There are so many defensemen available in theory and buckets of possibilities are bandied about in trade discussions. Assessing defensemen is difficult in 2020 despite all the advances in hockey analytics. While I can’t possibly cover or comment on them all, I hope this writeup gives an inkling of some of the candidates out there.

NHL teams like the Leafs undoubtedly go through a much more rigorous process, with video analyses, scouting reports, cap experts, up-to-date player files, and private tracking data.

Having said all that, I have a feeling that defensemen such as Brenden Dillon, Adam Larsson, Jason Demers, and Dylan DeMelo are on Kyle Dubas’ radar.

The Dylan DeMelo Option

DeMelo seems to be the most attainable, both because of his low cap hit ($900K) and the fact that Ottawa is clearly out of the race. His numbers also suggest that he might be a better two-way play-driving option with an offensive element to add to his resumé. That could turn out to be a better fit for Keefe’s system than some of the more one-dimensional defensive candidates.

The biggest issue with DeMelo is that he is a pure rental and he doesn’t fit the Leafs reported desire to sign a right-side player with term. At the end of the day, though, teams don’t always find the perfect solution (like Jake Muzzin).

The Senators have prospects in the system they may want to bring up next year, so it’s an open question as to whether a budget-constrained Pierre Dorion gives DeMelo an offer or chooses the rental route. The Ottawa Sun published a recent article on Dylan DeMelo’s status for those interested.

Personally, I’m thinking a young dynamic prospect like Dmytro Timashov together with a third- or fourth-round pick could be enough to acquire DeMelo. That’s not cheap, but as Leafs fans know, prices are higher at the trade deadline for quality rentals. The Leafs would also need to be convinced that DeMelo would improve their right side playoff roster over Holl/Barrie/Ceci (and possibly Liljegren) before they would go the rental route.

The Jason Demers Option

If veteran defender Jason Demers is in play, my guess is we’re probably looking at a solid prospect and a second-round pick in lieu of a first. Maybe Jeremy Bracco? Egor Korshkov? I’m not sure.

Almost by definition, the cost to acquire Jason Demers would be between the Jake Muzzin and Tomas Plekanec deals. Odds are that John Chayka and the Coyotes won’t make a move unless they are well out of the playoff picture. After all, they invested in Taylor Hall this season.

The Brenden Dillon LD Option

Brenden Dillon is a very attractive left-handed defender with good size and good results. Boston appears to be one of the teams going after him. If the Leafs have an interest in bolstering the left side, I would be after him, too. He would probably do a good job filling in on the top four in Rielly’s absence.

That would mean Dermott and Sandin would be pushed down the lineup and we could see a top four that looks like: Muzzin – Holl / Dillon – Barrie.  What would it take to acquire Dillon as a rental? It would probably be a similar scenario to what I described with DeMelo.

The other option for the Leafs may be to simply review the landscape and opt to forego a trade in the absence of Morgan Rielly. They could just wait until the off-season, when prices aren’t as high, the cap situation is clearer, free agents are free, GMs have a drink together, and teams are attempting to move puzzle pieces around at the draft. The return of Jake Muzzin and the recent play of Tyson Barrie and Rasmus Sandin have made that seem more feasible than it was even three weeks ago.

Kyle Dubas has a lot of moving parts and constraints to deal with right now. Making a trade that helps the club won’t be easy, but it’s certainly possible, as I hope this series has shown.

The following data sources were instrumental to this article and in its data viz:
Corsica, HockeyViz, MoneyPuck, Natural Stat Trick, PuckIQ, and PuckPedia. I encourage your support to these providers, including Patreon where available.