The SuperFan Podcast – Episode 11 – Catherine Silverman

3000by3000 (1)This week on the podcast – Catherine Silverman (@catmsilverman) from InGoal Magazine and The Athletic!

Catherine shared her insights on the Arizona Coyotes season, how the roster has been constructed and the key drivers for their success thus far. I also got to learn about how the young core is progressing and the impact Taylor Hall has had as the club pushes for the top spot in the Pacific division. Catherine also gave her take on the Oilers goaltending situation and how best to deploy the Koskinen/Smith tandem.

Big thank you to Cat for her knowledge and insights!

Full segment below:

Podcast channels:

Music: Anitek. “Show me.” Anitek Instrumentals Vol. 4, 2010. Jamendo.com

Boosting the powerplay

coppernblue.com.full.54273With the Edmonton Oilers even-strength (5v5) performance being so poor this season – ranking 29th in the league in terms of goal-share with 44.69% (-19 goal differential) and often getting out-shot and out-chanced – it’s become even more critical that the Oilers generate as much offence as possible on the powerplay. It’s a game-state the Oilers should excel within given the high-end talent on the roster and the success the group – including McDavid, Draisaitl, Nugent-Hopkins and Klefbom – have had in previous seasons.

The good news is that the Oilers currently rank near the top of the league with the man-advantage, second only to the Tampa Bay Lightning, scoring 10.69 goals per hour – a metric that also captures how efficient teams are at scoring in a time-pressured situation. Since it’s a competitive results-driven league, it’s also important to dig behind the outputs and determine if the results are in fact sustainable and try to uncover any areas that might need attention from the coaching staff and management.

The first thing to note about the Oilers powerplay is that it’s allowed seven short-handed goals – the most in the league and the highest rate of goals against per hour with 1.92. That doesn’t drive down their overall results on the powerplay significantly as they would still rank third in the league if we factor in goals against. But not allowing those goals would have them at an even overall goal-differential today and likely a little higher up in the Pacific division standings. Worth noting that the Oilers don’t allow a lot of chances against, but when they have, their goaltending has been poor.

The second issue is that the Oilers are heavily reliant on their top powerplay unit, which isn’t surprising considering (a) their top end talent, (b) the amount of time the Oilers have typically trailed in games this season and (c) the lack of depth options on the roster. In fact, the Oilers for the second year in a row are near the top of league when it comes to the proportion of powerplay ice time the top unit is deployed. The league average proportion of ice time for top powerplay units is typically around 65% over the last three seasons. The Oilers currently deploy their top unit 81.0% – second in the league only behind the Washington Capitals.

Note: To determine the top powerplay units proportion of ice-time and their results, I took the forward with the most ice-time on each team and used them as a proxy for the first powerplay unit and divided their ice time with the team’s total powerplay time. The second powerplay units ice time and results were determined by subtracting the ice-team leaders TOI and results from the team totals. Data can be found in the Appendix.

It’s worth noting that a team like Tampa Bay who have historically been excellent with the man-advantage and are competing with the Oilers for the top powerplay have enough depth to ice two productive powerplay units this season. The first unit, using Kucherov as a proxy, is deployed for 66% of the total powerplay time – much closer to league average rates – and has generated 11.22 goals per hour (a total of 25 goals), well above the league average of 7.93 goals per hour. And while Kucherov and his group take a much needed break and to stay fresh for even-strength play, the second unit scores at a similar rate, generating 11.32 goals per hour; a total of 13 goals. That’s very impressive considering that second powerplay units on average generate 5.51 goals per hour. Both Lightning units are also above average relative to similar deployment groups when it comes to creating chances with the man-advantage, with the top powerplay unit generating 89.74 unblocked shot attempts per hour and the second unit generating 78.42 per hour. On average, top powerplay units generate 80.34 unblocked shot attempts per hour and second powerplay units generate 56.78 per hour. Put another way – the Lightning’s second powerplay unit generates more unblocked shot attempts per hour than half of the league’s top powerplay units.

Note: for special teams analysis, I include unblocked shot attempts (i.e., Fenwick) to assess a team’s ability to score or prevent goals against. Blocking shots and keeping pucks to the outside is a big part of killing penalties, so Fenwick gives us a sense of how well the skaters are doing their jobs and helping out their goaltenders.

The Oilers on the other hand don’t have enough depth to regularly deploy two powerplay units, as the bottom six forwards are predominantly penalty-killing specialists – a major focus for Oilers management this past off-season. In the limited minutes that McDavid hasn’t been on the ice for the powerplay, the Oilers have scored only one goal – a rate of 1.43 per hour – and generated 31.81 unblocked shot attempts per hour – both of which ranks last among all second powerplay units. What the Oilers roster is missing are depth players, individuals on the third and fourth lines at even-strength, that can play powerplay minutes and contribute, similar to the Lightning have available in Maroon, Killorn and Gourde and what the Oilers had in Letestu a few years ago.

As long as the Oilers second powerplay unit can’t even generate league-average rates of shots and chances, the pressure will remain on McDavid and the top group to continue producing. So far they’ve been excellent, generating 86.95 unblocked shot attempts per hour and 12.86 goals per hour – both of which rank highly among top powerplay units.

But since the Oilers management group ignored the results and findings from last year’s powerplay – one in which the top unit played close to 80% of the total time and the second unit was one of the worst in terms of goal-scoring – the pressure will remain on the likes of McDavid and Draisaitl to continue playing league-leading minutes and producing. Hopefully the Oilers recognize that their depth forward group isn’t just for penalty killing and begin finding cheap, reliable options for the second powerplay unit. It’s going to require critical analysis of their powerplay and looking past the results, something that’s historically been ignored by Oilers management when the overall results have been fine.

Data: Natural Stat Trick

Appendix A: NHL powerplay units, 2019/20 (As of January 5, 2020)

NHL PP First and Second Units - 20190106.jpg

Also posted at The Copper & Blue.

Chasing the game

22089843.jpg

The Oilers have been dreadful at even-strength (5v5) this season, now sitting 29th in the league with a goal-share of 43.75% – a -22 goal differential, scoring 77 goals and allowing 99. About 80% of the game is played in this state, so it’s critical that the Oilers start out-scoring their opponents more regularly if they intend on competing for a playoff spot.

One of the Oilers biggest issues is their inability to create offensive opportunities, as they currently rank 28th in the league when it comes to generating shot attempts, unblocked shot attempts (i.e., Fenwick, a proxy for scoring chances) and shots on goal. What’s worth noting is that defensively, they’ve been hovering around league average rates having done a decent job suppressing shot attempts and chances against. Unfortunately their goaltending has not been good, posting a team save percentage of 90.54% – good for 29th in the league.

So while we can expect that the Oilers rate of goals against to improve if the goaltending gets back to league average levels, we can’t have as much confidence in the offensive side of things as the team consistently spends more time defending and playing in their own zone.

What’s especially troubling is the Oilers lack of offence when they’re trailing in a game –  a game state when most teams start controlling the flow of play more, taking chances as teams with the lead get into more of a defensive shell. The Oilers have posted a goal-share of 44.44% when trailing in games, good for 29th in the league and only one of eight teams with a goal-share below 50%. They rank 21st in terms of goals per hour with 2.39 and 28th in terms of goals against per hour with 2.98. And their underling numbers aren’t that much better either.

Note that for this analysis I’ll be looking at unblocked shot attempts (i.e., Fenwick) as the Oilers appear to be less of a shot-quantity/volume-shooting team and more of a shot-quality team under Tippett.

Below is a breakdown of the Oilers share of unblocked shot attempts when they’re leading in a game, trailing in a game and tied with an opponent. I’ve also included the league’s top Fenwick For percentage and the worst to show where in the range the Oilers fall.

20200103-FF

As expected, we see that the Fenwick for percentage is highest when team’s are trailing, doing more and taking risks to tie up the game, compared to when team’s are leading and playing more defensively. The Oilers are just below average when the game is tied or when they’re leading. But when they’re trailing, they’ve only posted a 51.48% share, which ranks in the bottom quarter of the league.

What’s interesting is that while they do allow slightly more shot attempts against when trailing, it’s their offense that goes completely dry. Below is breakdown of their rate of unblocked shot attempts generated in different score states, along with the best and worst rates in the league.

20200103-FA60

20200103-FF60.png

Offensively, there’s not much of a difference for the Oilers whether they’re leading or trailing – it’s like they use the same tactics and playing style regardless of the score. On average, teams see an increase of 6.35 unblocked shot attempts per hour when trailing compared to when they’re leading – the Oilers are on the low end with an increase of only 5.10.

It should also come as no surprise that the Oilers do a lot better offensively with McDavid on the ice when trailing as the Oilers generate 45.07 unblocked shot attempts per hour with him on the ice and only 39.53  without him. In other words, much closer to league average rates with him and closer to the Detroit Red Wings without him. This is one of the drawbacks of focusing so much on improving the penalty kill this past off-season and bringing in defensive-minded forwards to fill-out the bottom six. Yes, McDavid, Nugent-Hopkins and Draisaitl are going to see an increase in their ice-time proportion when the team needs a goal, but there’s still a big portion of ice time when the bottom six needs to contribute offensively and at  least create chances.

For reference, below are the Oilers forwards, sorted by total even-strength ice time, and their proportion of the team’s ice time by the three different score-states.

20200103-Prop of ice time.png

Minor observation: Sam Gagner sees a spike in his ice time proportion when the team needs a goal. I thought that was strange until I noticed that he has the highest on-ice rate of unblocked shot attempts  when the Oilers are trailing in a game. Of the bottom-six forwards, he’s the one player who can chip in offensively. But unfortunately for him and others, the team can’t get a save. Below is a table of the forwards, sorted by the rate of unblocked shot attempts for per hour.

Player TOI FF/60 FA/60 GF/60 GA/60 SH% SV% PDO
Gagner 119.57 49.18 33.62 2.51 3.01 6.76 86.67 0.934
Kassian 222.83 45.77 41.74 4.04 3.50 12.30 89.08 1.014
McDavid 252.93 45.07 44.12 4.03 3.56 12.41 89.51 1.019
Nygard 91.03 43.50 38.23 3.95 0.00 11.32 100.00 1.113
Granlund 78.90 41.83 38.02 1.52 3.04 4.76 85.19 0.899
Draisaitl 247.67 40.94 47.48 2.91 3.63 10.08 90.26 1.003
Jurco 39.77 40.74 28.67 4.53 1.51 18.75 91.67 1.104
Haas 106.60 40.53 30.39 1.69 2.25 5.26 87.50 0.928
Neal 188.02 40.53 38.93 1.28 4.15 4.55 85.71 0.903
RNH 170.38 40.14 37.68 1.76 2.11 6.10 92.50 0.986
P. Russell 93.88 39.62 35.79 1.92 2.56 6.52 88.89 0.954
Khaira 134.50 38.81 35.24 0.89 3.57 2.94 86.67 0.896
Chiasson 149.10 38.23 38.63 1.61 2.01 5.63 92.06 0.977
Archibald 97.52 35.69 38.15 0.62 4.31 2.44 82.50 0.849
Sheahan 108.13 35.51 36.07 1.11 3.33 4.17 86.05 0.902

Another thing I found was that when the team needed a goal and McDavid was on the ice, the rate of shots against would go up. In fact, his on-ice Fenwick For% when the Oilers are trailing is currently at 50.53%, which among the Oilers forwards is only better than Chiasson (49.75%), Sheahan (49.61%), Archiabald (48.33%) and Draisaitl (46.30%).

The fact that Draisaitl ranks last made me wonder what the splits were like between him and McDavid. Here’s what their share of unblocked shot attempts are like with and without each other at even-strength when the Oilers are trailing. I also included the goal-rates and goal-share.

Combo TOI FF/60 FA/60 FF% GF/60 GA/60 GF%
Together 184.15 42.36 49.85 45.94 3.58 3.91 47.83
McDavid without Draisaitl 68.78 52.34 28.79 64.52 5.23 2.62 66.67
Draisaitl without McDavid 63.51 36.84 40.62 47.56 0.94 2.83 25.00
Neither 387.23 39.98 35.64 52.87 1.55 2.63 37.04

Small sample size, but it appears that part of the reason McDavid has only been league-average when the the team is trailing is because of Draisaitl. I don’t expect Draisaitl’s numbers away from McDavid to be great due to the lack of talent on the roster to play him with. But McDavid does see a significant jump in his Fenwick For% due to a big drop in the rate of shots against when he doesn’t have Draisaitl with him.

You’ll probably get a lot of resistance doing this, but one way to increase Draisaitl’s rest-time and potentially get more overall productivity from him is to reduce his minutes when the team is trailing. He’s not doing well in that game-state offensively, dragging down his linemates, and it appears that the low-event bottom six are at least getting a decent share of the shot attempts.

Summary

  • The Oilers aren’t very good when it comes to Fenwick For% at even-strength, especially when they’re trailing in games.
  • It’s their inability to create offence when trailing that’s especially troubling as they rank near the bottom of the league in terms of Goals For% and Fenwick For%.
  • Part of the problem is their bottom six who while do a good job suppressing shots when the team is trailing, have a lot of trouble creating chances. The team is in dire need of offensive talent, not just guys who can kill penalties.
  • The other problem is the goaltending, which ranks near the bottom of the league at even-strength and when the team is trailing.
  • Draisatl has been playing very poorly defensively as reflected by his on-ice share of shot attempts and the rate of unblocked shot attempts against when the team is trailing.

Data: Natural Stat Trick

Tracking the Pacific Division – As of December 31, 2019

gettyimages-1136590914-594x594

The Pacific Division is wide open at this point as not a single team has posted a points percentage that would rank in the top ten league-wide. Vegas currently ranks first in the division but is 13th in the league, Vancouver is 15th and Arizona is 16th.

What’s crazy is that only two of the eight teams have a goal-share above 50% at even-strength this season: Vancouver with 50.67% and Arizona with 51.35%. The remaining six teams are in the bottom ten league-wide.

Team Goal Differential Goals for% League rank
Arizona 4 51.35 13th
Vancouver 2 50.67 16th
Vegas -4 48.85 22nd
Anaheim -9 47.06 25th
Los Angeles -11 46.41 26th
Calgary -17 44.37 27th
Edmonton -22 43.6 28th
San Jose -29 41.81 30th

The Edmonton Oilers are only ahead of New Jersey, San Jose and Detroit when it comes to even-strength results – three teams destined for a draft lottery pick. What’s especially concerning is how poorly the Oilers rank in terms of shot-share metrics within the Pacific as they’re spending more and more time defending and playing in their own zone.

Below are the division results at even-strength (5v5) as of December 31, 2019, with teams sorted by point-percentage. For each of the shot-share metrics, I’ve applied a basic heat-map to show which teams are doing well compared to the division teams and which are struggling.  A description of each metric is at the end of this article.

Pacific Division - 20191231.JPG

The Oilers score/venue-adjusted Corsi For% at even-strength continues to slide as they currently rank 26th in the league and have posted a Corsi For% of 46.53% over their last twenty-five games – ahead of only Detroit and the New York Rangers. Vancouver and Arizona aren’t faring much better, posting a 47.0% Corsi For% and likely unable to keep up with Vegas whose shot-share numbers rank in the top ten league-wide. Los Angeles continues to do well in terms of shot-share, but don’t appear to have the talent to convert their opportunities into goals or enough consistent goaltending. San Jose, Calgary and Anaheim are all around the 49.0% mark in terms of Corsi For% over their last twenty five games.

Below is a snapshot of the last twenty-five games for each team at even-strength, sorted by points percentage. Note the Oilers goal-share of less than 40% and their Corsi For% of 46.53%. I’d like to believe the Oilers will bounce back even slightly, but they’ve been spending a lot of their time without the puck – and that’s with and without their top players on the ice.

Team PTS Point% CF% FF% xGF% GF% SH% SV% PDO
Vegas 29 0.580 53.49 53.48 54.54 52.88 8.12 91.74 0.999
Arizona 28 0.560 47.02 47.71 46.87 48.22 7.12 93.04 1.002
Los Angeles 27 0.540 55.09 54.82 55.19 48.95 6.01 92.51 0.985
Calgary 27 0.540 49.06 48.61 48.32 42.09 5.99 92.38 0.984
San Jose 26 0.520 49.58 49.01 48.19 45.54 8.96 89.22 0.982
Vancouver 25 0.500 47.05 47.36 45.73 44.49 7.09 92.36 0.994
Edmonton 24 0.480 46.53 47.69 47.14 39.96 8.05 89.34 0.974
Anaheim 19 0.380 48.84 48.73 47.77 41.95 7.03 90.98 0.980

From the looks of it, Vegas should be able to secure the top spot in the division, and I think should eventually move into the league’s top ten. The rest is anyone’s guess as every other team in the Pacific, except for Los Angeles, consistently gets out-shot and out-chanced.

ENIo2IuW4Ak_SjE.jpg

Quick glance at each team’s special team units, and the Oilers continue to do well and get production from their powerplay and penalty kill. Below is each team’s combined goal rates which adds together the rate of goals for and against on the powerplay and penalty kill.

Team Special Teams Combined Goal Rates
Edmonton 3.97
Vancouver 3.24
Vegas 2.83
San Jose 1.03
Arizona 0.21
Calgary -0.27
Anaheim -1.98
Los Angeles -3.98

Vegas has slipped slightly from last month when they posted a +4.09 goal rate, and San Jose has dropped from +3.26 to +1.03. Anaheim has improved quite a bit even though they’re still a negative team, jumping up from -4.17 at the end of November.

Data: Natural Stat Trick, Charting Hockey

Glossary:

  • Points-percentage (Point%) – The total points accumulated divided by the points that were available, including extra time.
  • Corsi For percentage (CF%) – The proportion of all the shot attempts the team generated and allowed that the team generated (i.e., Corsi For/(Corsi For + Corsi Against). This is used as a proxy for possession and can predict a team’s future share of goals (GF%).
  • Fenwick For percentage (FF%) – The proportion of all the unblocked shot attempts the team generated and allowed that the team generated (i.e., Fenwick For/(Fenwick For + Fenwick Against). This is used as a proxy for shot quality and considers shot blocking a repeatable skill. It can also predict a team’s future share of goals, slightly better than Corsi.
  • Shots For percentage (SF%) – The proportion of all the shots on goal that the team generated and allowed that the team generated (i.e., Shots For/(Shots For + Shots Against).
  • Expected Goals For percentage (xGF%) – This is a weighting placed on every unblocked shot based on the probability of the shot becoming a goal. This depends on the type of shot, location and uses historical shot and goals data to come up with the probability for each unblocked shot. This has been found to be a better predictor of future goals than Corsi and Fenwick.
  • Goals For percentage (GF%) – The proportion of all the goals that the team scored and allowed that the team generated (i.e., Goals For/(Goals For + Goals Against).
  • Shooting percentage (SH%) – The percentage of the team’s shots on goal that became goals (i.e., total goals divided by the total shots on goal).
  • Save percentage (SV%) – The percentage of the team’s shots on goal against that were saved (i.e., 1-(totals goals allowed divided by the total shots on goal against))
  • PDO – The sum of a team’s shooting percentage (SH%) and its save percentage (SV%). It’s based on the theory that most teams will ultimately regress toward a sum of 100, and is often viewed as a proxy for how lucky a team is. (Source)

Special teams tracking

EGsFNSeUEAAebcf.jpg

With the Oilers play at even-strength (5v5) being so poor, it’s become more and more imperative that the club gets as much out of their special teams as possible. The Oilers currently have a 43.90% goal-share at even-strength, a -20 goal differential – fourth worst in the league only ahead of New Jersey, San Jose and Detroit. What’s especially alarming is that their underlying shot-share metrics (47.30% Corsi For%, 48.17% Fenwick For%) are poor and continuing to decline, indicating that they’re spending more time defending and less time in the offensive zone at 5v5.

Now our expectations of the Oilers special teams should be higher considering the high-end offensive talent on the roster for the powerplay and the amount of work and resources management put towards fixing the penalty kill this past off-season. In my mind, the powerplay should be elite, while the penalty kill should at least be somewhere above league average rates (which would be a significant improvement from last season).

So far things have gone well for the most part as the Oilers rank second in the league scoring 10.69 goals per hour on the powerplay. And their penalty kill sits fourth in the league allowing  5.39 goals per hour on the penalty kill. To put things into context, the average rate of goals on special teams over the last three seasons is about 7.10 per hour.

Special Teams (2019/20) Rate League rank
Powerplay – Goals for/60 10.69 2nd
Powerplay – Fenwick for/60 78.12 8th
Powerplay – Shooting% 18.95% 1st
Penalty Kill – Goals against/60 5.39 4th
Penalty Kill – Fenwick against/60 74.11 21st
Penalty Kill – Save% 89.90% 3rd

Note: I do use rate stats (i.e., goals, shots per hour) for special teams as it factors in the time a team takes to achieve their objectives. Powerplay units are up against the clock when they’re deployed, so team’s that can get results faster should be considered more efficient. I also use unblocked shot attempts (i.e., Fenwick) to assess a team’s ability to score or prevent goals on special teams. Blocking shots and keeping pucks to the outside is a big part of killing penalties, so Fenwick gives us a sense of well the skaters are doing their jobs and helping out their goaltenders.

Powerplay

The Oilers powerplay is looking good for the most part, getting results and appearing to have the right tactics in place as reflected by their rate of unblocked shot attempts. Couple concerns however, the first of which is the team shooting percentage which could dip down if individual players get into slumps or if they run into a few hot goaltenders. The second concern is around their declining rate of shot attempts, which was at one point one of the best in the league. Below is a graph showing the Oilers rate of unblocked shot attempts per hour over rolling 25 game segments. They’re still around the league average rates, but it’s a little odd  that a team with that much elite talent has seen such a drop.

Oilers PP FF60.png

The other concern with the Oilers powerplay is that they’re allowing the highest rate of shorthanded goals against this season (2.08 per hour!), so far giving up seven. This one is on the goaltending as it’s not like the Oilers are allowing a ton of chances against on the powerplay – they allow the sixth lowest rate of shot attempts against, while the team save percentage is second worst in the league with 78.79% . This has to be a spot of bother for the coaching staff that is doing everything they can to improve their overall goal differential of -11.

Penalty kill

While the powerplay is good and has some underlying issues that need to be addressed, the penalty kill is good and is actually showing signs of progress. Over the full season, the Oilers have allowed 74.11 unblocked shot attempts per hour, slightly worse than league average and ranking 21st in the league. However, that rate has gradually been been declining over the course of the season, with the Oilers posting a rate of 67.71 unblocked shot attempts against per hour over their last twenty-five games – a rate that would have them in the top ten league-wide. Slight uptick happening recently, so it’s worth monitoring as the season progresses.

Oilers PK FA60

Something that I think the coaching staff might overlook is the rate of shot attempts against when certain players are on the ice. Below are the regular penalty killing forwards this season (minimum 30 minutes of ice time), sorted by their on-ice rates of unblocked shot attempts against (FA/60).

Player TOI TOI/GP FA/60 GA/60 On-Ice SV%
Ryan Nugent-Hopkins 58.75 1.68 57.19 5.11 87.50
Jujhar Khaira 63.58 1.72 59.45 2.83 93.33
Patrick Russell 33.67 1.12 67.72 12.48 76.67
Riley Sheahan 81.52 2.26 85.38 4.42 93.18
Josh Archibald 70.43 2.20 89.45 3.41 94.74
Markus Granlund 40.72 1.31 92.84 10.32 84.09
Leon Draisaitl 43.75 1.07 94.63 5.49 90.91

The team does a lot better keeping pucks to the outside when Nugent-Hopkins and/or Khaira are on the ice, with the actual rate of goals against being in a reasonable range. Where the team does see a spike is when the likes of Sheahan, Archibald or Granlund (who was demoted because of his poor penalty kill numbers) have been on the ice. What’s especially interesting to me is that Archibald’s on-ice rate of shots against isn’t that much better than Granlund’s. The only difference being that Granlund had a higher rate of goals against when he was on the ice  – which we know can often be goalie performance driven. Just compare Granlund’s on-ice save percentage (84.09) to Archibald’s (94.74). More on this in a second.

For those interested, here’s how the Oilers defencemen have done on the penalty kill so far this season, again sorted by their on-ice rate of unblocked shot attempts against. Interesting to me at least is that Nurse is not leading the team in ice time and looks more to be on the second penalty kill unit; the rate of shots against also increase when he’s on the ice. He’s also not a regular option on the powerplay, which should factor in to his next NHL contract.

Player TOI TOI/GP FA/60 GA/60 On-Ice SV%
Ethan Bear 73.58 1.79 74.20 5.71 88.89
Oscar Klefbom 118.40 2.89 79.05 6.59 88.79
Kris Russell 102.43 2.56 79.08 5.86 89.90
Darnell Nurse 81.78 1.99 81.43 5.14 90.91

The big question mark is the team save percentage (89.90%) on the penalty kill, which sits third in the league and can be expected to drop considering the talent-level between the pipes. Both goaltenders have been very good so far this season shorthanded. Among 41 goalies who have played at least 80 minutes on the penalty kill, Mikko Koskinen ranks ninth  with a save percentage of 0.894 (115 minutes), while Mike Smith ranks third with 0.917 (92 minutes). The issue is that both goalies are posting save percentages above their NHL career norms. Over the last three seasons prior to 2019/20, Mike Smith has played 152 games (752 minutes) and posted a save percentage of 0.865. Koskinen has only played 55 games (243 minutes) but posted a save percentage of 0.854 shorthanded. Worth noting that both of their historical penalty kill save percentages are slightly below the league average team save percentages from the last three seasons (0.867). I’d expect the Oilers to finish the year with a team save percentage closer to league averages.

One last note about the penalty kill. The Oilers do very little to squeeze out shorthanded goals while killing penalties. While they’ve allowed seven shorthanded goals, tied for most in the league, they’ve only scored once shorthanded themselves – the lowest rate of goals scoring in the league. They’re also generating the lowest rate of unblocked shot attempts in the league on the penalty kill, indicating that the coaching staff doesn’t really care to play more aggressively while shorthanded. Considering the the team is desperate for goals, it might be an area worth looking into.

Macaroni

To sum up how well the entire special teams are doing, I look to combine the rates of goals for and against on the powerplay and the penalty kill. Below is how the Oilers have done over the course of the season.

Oilers Macaroni - 20191231

To put things into context, teams with great special teams post a combined goal rate above +1.50 (Source). Last season, Tampa Bay was over 6.00 as their powerplay was ranked first and only allowed three shorthanded goals. And their penalty kill was first and also scored the fifth highest rate of shorthanded goals. Five other teams posted a combined goal rate above +1.50.

The Oilers got some outstanding contributions from the special teams early on, which helped bank those points but also masked their major issues at even-strength. Now we’re seeing the special teams come back down to earth a bit, but should still expect them to finish well considering the talent up front on the powerplay and the amount of resources put towards addressing the penalty kill this past off-season. With the Pacific division being wide open this season, it’s critical that the Oilers correct their underlying issues and adjust their roster and deployment tactics as necessary.

Data: Natural Stat Trick

The SuperFan Podcast – Episode 10 – Keith Anthony

3000by3000 (1)This week on the podcast I was joined by Keith Anthony (@keyantonyo) to discuss the Bakersfield Condors and the Oilers prospect pool. Keith has done a lot of work keeping tabs on the Condors using video analysis and his own tracking data – you can find his past work at his blog: Petro Praxis. Definitely worth a follow on Twitter if you want to know how the Oilers prospects are doing.

Keith shared his perspective on player development and the Oilers overall approach under Ken Holland. We also discussed specific players who’ll be critical to the Oilers long-term success including Caleb Jones, Kailer Yamamoto, Tyler Benson and William Lagesson.

Full segment below:

Podcast channels:

Music: Anitek. “Show me.” Anitek Instrumentals Vol. 4, 2010. Jamendo.com

Running operations

cut

Couple things occurred on Saturday that caught my attention and made me think about  how the Oilers choose to operate in year five of the McDavid era.

First twenty games

Dave Tippett made some comments to the local media about the Oilers current situation and how things compare to the first twenty games of the season.

The way I look at it, the first 20 games we were pretty solid coming out of the gate. We won some games. There were some games we didn’t play as well. But we had a lot of…there was a lot of purpose in what we were doing.

And the next 20 games, what happens is after the first month the league starts to get to know your team and you got two guys driving the offence, like driving it.  Fifty points and all of a sudden if you’re another team coming in here . You’re saying ‘Hey better take those guys [McDavid and Draisaitl] away’. So it gets harder. So when it gets harder for those guys to score then there’s a frustration that comes in and now you start trying to do more, trying to do more, trying to do more. And you’re chasing the game all the time. And that’s where Drai’s been chasing the game for a while. And he’s been chasing the game because he’s trying to win. And it’s gotten harder and harder for him because other team’s are really bearing down on those guys. – Oilers head coach Dave Tippett (Source: Edmonton Oilers)

Couple thoughts.

One, while the results were excellent for the Oilers in their first twenty games (12-6-2, 0.650 points percentage), there were some warning signs that indicated there might be trouble ahead. The biggest issue was their performance at even-strength (5v5), as the team posted only a +3 goal differential and a 52.00% goal-share, which was actually fourth best in their division. The Oilers even-strength deficiencies were clearly being masked by a very good powerplay and penalty kill, both of which continue to rank in the top five league-wide.

The Oilers other issue was their underlying shot-share metrics over those first twenty games, as the club posted a Corsi For% (i.e., a proxy for possession) of 48.62%, 23rd in the league and only ahead of Anaheim in the Pacific division. Their Fenwick For% (i.e., a proxy for scoring chances) wasn’t much better, ranking 20th in the league with 48.93% and again only ahead of Anaheim. Note that these numbers have been score and venue adjusted. What was especially troubling in those first twenty games is how poorly then ranked in terms of generating shot-attempts, unblocked shot attempts and shots on goal, as they ranked in the bottom six in the league – which is where they also sit today.

What the Oilers also should’ve identified in those first twenty games was the team’s sub-par underlying numbers when McDavid and Draisaitl were on the ice. The actual results were excellent, with the Oilers getting 70.59% of the goals when they were together, outscoring opponents 24-10. But they were also spending more time defending and playing in their own zone during those twenty games.

Here’s how the Oilers did when McDavid and Draisaitl played together in the first twenty games. McDavid spent 88% of his total ice time at even-strength playing with Draisaitl. Numbers have been score and venue adjusted.

TOI CF% FF% GF% On-Ice SH% On-Ice SV% PDO
300.72 47.84 47.22 70.59 15.27 94.21 1.095

While the goal-share was excellent, driven largely by an on-ice PDO of 109.5, the duo posted a Corsi For% of 47.84% and a Fenwick For% of 47.22. While we can expect a top-line’s shot-share to take a hit due to playing against the best competition, I would not expect the Oilers to actually do better without them on the ice. Without McDavid and Draisaitl, the Oilers posted a Corsi For% of 49.40% and a Fenwick For% of 50.03% in those first twenty games.

  • Related: Sliding – The SuperFan (2019, December 12

Seeing this early on, the team should not be surprised that after 41 games they rank 25th in terms of Corsi For% (47.30%) at even-strength and 23rd in terms of Fenwick For% (48.17%). And most importantly, they rank 28th when it comes to goal-share with 43.90%. As for Tippett’s comment, I would argue that the league already knew the Oilers weaknesses before the season started and exploited them even further as the roster didn’t change.

Granlund

Granlund

The second event on Saturday was the demotion of Markus Granlund, which on the surface isn’t a big deal as the player had been healthy-scratched often as of late, and is on a $1.3 million contract that expires this coming summer. What the move does indicate is that the Oilers still lack the ability to gather information, conduct analysis, and make decisions geared towards winning. Issues that really plagued the Oilers hockey operations during the previous manager’s regime and haven’t been addressed since by the owner.

When Granlund was signed this past summer, he was touted as a penalty kill specialist who could help the bottom-six forwards. However if the Oilers had conducted a simple analysis using publicly available data, they would have seen that his on-ice numbers on the penalty kill were actually poor and that his reputation was being bolstered by his previous team’s overall success. In case you missed it, below is what I put together shortly after the Oilers signed Granlund in July.

It really should be no surprise to the Edmonton Oilers that Granlund’s penalty killing numbers were so dreadful, as the Oilers allowed a rate of 10.32 goals against per hour when he was on the ice – the highest rate among the six most regular penalty killing forwards (those that have played at least 40 minutes). To put things into perspective, the Oilers as a team have allowed 5.39 goals against per hour this season, good for fifth best in the league. Over the previous three seasons, teams allow an average of 7.09 goals against per hour on the penalty kill.

Last season in Vancouver, which had one of the penalty kills in the league, Granlund posted an on-ice rate of 9.16 goals against per hour – one of the worst on the team and among forwards across the league who played a similar amount of minutes. The rate of unblocked shot attempts jumped up when Granlund was on the ice both this season in Edmonton and last season in Vancouver.

While the signing of Granlund was a low-risk move, Oilers management could have saved themselves some cap-space and potentially given those minutes to a more productive player had they just done a brief analysis using publicly available data. Instead, they’re paying an NHL salary to an AHL-bound asset and showing little overall progress in terms of building a championship contender.

If the Oilers are serious about competing, management really needs to ramp up their information gathering, analysis and overall decision-making. And that has to occur before and after every transaction they make to ensure they’re leveraging as much value as possible from their roster.

Data: Natural Stat Trick

Life line

mcdavid2

The Oilers even-strength (5v5) results have been very poor thus far, as the Oilers rank 27th in the league in terms of goal-share with 45.21% – scoring 66 goals and allowing 80. Thankfully their powerplay and penalty kill, which has posted a net goal differential of +12, has offset their poor even-strength results and helped them secure 42 points after 36 games.

While it remains to be seen if their special teams can continue bailing the team out (the powerplay doesn’t look like it’s slowing down anytime soon, while the penalty kill has begun to slip) the Oilers desperately need help, either in the form of on-ice personnel or possibly new coaching tactics to bring their even-strength results up to respectable levels.

The other factor that can drive results is of course McDavid who has the talent and track record now of being a true difference maker. Below is the Edmonton Oilers even-strength goal-share  since 2015/16 with McDavid on the ice, compared to their goal-share without him, along with the difference.

Season Goal-share with McDavid Goal-share without McDavid Difference
2015/16 50.70% 40.00% +10.70
2016/17 62.10% 48.90% +13.20
2017/18 57.04% 41.62% +15.42
2018/19 50.66% 40.74% +9.92

This season, it’s been no different. The Oilers have so far posted a goal-share of 55.07% with their captain on the ice, and a goal-share of 36.36% without him. Remarkably, their depth has posted results even worse than years past, posting a goal-differential of -21 without McDavid. And it doesn’t look like even McDavid can save them, as his  on-ice results have gradually been tapering off.

Below is the team’s cumulative goal-differential with McDavid on the ice (blue line), compared to the team’s cumulative goal-differential without him (red line).

20191217 - Goal differential.jpg

The first thing that jumps out is how quickly the results without McDavid have declined, and the amount of work now required to dig out of this mess. A big reason why the Oilers have struggled is their inability to control the flow of play and generate scoring chances. Over the last 15 games, which is where we’ve seen the sharpest decline, the Oilers have posted a 46.81% Corsi For% (a proxy for possession) and a 48.14% Fenwick For% (a proxy for possession) without McDavid on the ice. It was obvious in the summer and it’s been made very clear now that the moves the team made to shore up the depth aren’t good enough and are now costing the team wins in the standings.

The second thing that jumps out is McDavid’s goal-differential, which has stagnated around the +10-mark for quite sometime now and currently sits at +7. While that isn’t bad, keep in mind he finished the 2016/17 season with a +30 on-ice goal-differential at even-strength, and followed that up with a +20 on-ice goal-differential in 2017/18; goal-shares were above 55% in both of those seasons. Last season appeared to be an anomaly for him as he only finished with a +2 on-ice goal-differential, which can probably be attributed to being overplayed and the coaching tactics that season. He wasn’t able to make up for the goals scored against when he was on the bench, and unfortunately it looks like the same thing will happen this season.

One reason for that is the team’s inability to control play even with McDavid on the ice, as the Oilers often get out-chanced even with him on the ice. What’s especially troubling is the fact that the Oilers on-ice share of scoring chances (i.e., Fenwick) have been gradually declining, making me wonder if the injury McDavid sustained at the end of last season hasn’t healed or if he’s feeling fatigued. Note that the Oilers on-ice shot-share metrics with McDavid on the ice over the course of the 2019/20 season are well below his career norms (i.e. seasons prior to 2019/20).

McDavid on-ice (5v5) Career prior to 2019/20 2019/20
Corsi For% 51.28 47.21
Fenwick For% 51.81 47.05
xGoals For% 53.05 47.89

Below is the Oilers share of scoring chances with McDavid on the ice, over rolling ten game segments. It’s mind-blowing to see the best player in the world in his prime post such poor on-ice shot-share numbers, spending more and more of his ice-time without the puck.

McDavid FF.jpg

While McDavid has the ability to carry a team and make up for his teammates mistakes, it doesn’t appear he alone will be able to secure the Oilers a playoff spot this season. The Oilers desperately need skill and talent to remain competitive, contributing offensively and to potentially take some of the workload off of McDavid who clearly hasn’t been himself. The other frustrating part in all of this is that the lack of talent and effective coaching tactics to control the flow of play is really holding back McDavid’s offensive potential, as he could be scoring more often if the Oilers had the puck more frequently.

Data: Natural Stat Trick

Line combinations that might help

Haas-1040x572.jpg

In an attempt to find balance against a good Dallas team tonight, the Oilers have shuffled their line combinations with McDavid, Draisaitl and Nugent-Hopkins each centering their own lines.

On the surface, these line combinations are fine. Ideally, the Oilers have three good lines that can chip in offensively and give opposing teams problems with match-ups. The problem is that the Oilers don’t have enough talent on the wings to be a three-line team, but Tippett seems to expect the top players to carry the load and make the best of their situation.

We’re looking for some balance, just some stability in our line up. I look at those three guys, three of our top players. And we need those players not only to play well themselves, but to drag some people along with them. That’s what good players do, they make other players around them better and we’re hoping that’s the scenario tonight. Source: Edmonton Oilers

It’ll be interesting to see how long the line combinations stick, especially if the Oilers fall behind on the scoreboard early on, something that’s very likely to happen.

While the line combinations aren’t terrible, I don’t think they’ll address the Oilers current issue preventing goals at even-strength (5v5) or their inability to out-shoot and out-chance opponents. Over their last ten games, the Oilers have been outscored 13-30, a goal-share of 30.23%  – second worst only to the Red Wings. Their even-strength issues, I should note, is not a recent trend as they’re now fourth worst in the league over the entire season in terms of goal-share, ahead of only New Jersey, San Jose and Detroit.

And when it comes to shot-based metrics, the Oilers have seen their numbers decline over the course of the season (as noted in my last post). Taking the full season into account, the Oilers rank in the bottom third of the league when it comes to Corsi For%, Fenwick For% and Expected Goals For%.

20191213 - Shot share at 5v5

Over the last ten games, the Oilers have posted the following:

  • Corsi For% – 45.41% (29th)
  • Fenwick For% – 46.02% (29th)
  • Expected Goals For% – 43.50% (30th)

The somewhat good news is that there may be a way to address these underlying issues and help improve the goal-share leveraging the existing roster of players. These potential  line combinations won’t carry the Oilers to a championship, but can at least help stop the bleeding. Especially when it comes to their rate of expected goals against, which ranks 23rd in the league with 2.34, and is getting worse. And it’s based on line combinations that have been tried before this season and posted good shot-share percentages, but maybe didn’t get the results desired by the coaching staff

Below are the line combinations, including each trios time spent together at even-strength this season, along with their shot-share metrics (score adjusted) and actual results as captured by their goals-for percentage. More details can be provided in the appendix. Starting off with the top six.

5v5 TOI CF% FF% xGF% GF%
Draisaitl-McDavid-Kassian 399.25 47.66 46.51 47.93 57.40
Neal-RNH-Chiasson 103.47 59.38 58.41 60.79 50.00

The top line is one where the underlying numbers have been poor all season for the trio, but the results have been very good. Ideally their possession metrics and expected goal-share is even higher, giving us some assurances that the results are sustainable. For now, we can assume their results will likely slip, but there’s that McDavid factor to take into account.

The second combination, featuring Nugent-Hopkins, Neal and Chiasson had results that were a little surprising. While the goal-share has broken even (4 GF, 4 GA), their possession numbers as captured by their Corsi For% of 59.38% has been excellent. What’s interesting is that it was their on-ice save percentage that pulled down their goal-share, as their expected rate of goals-for (2.38) was very close to their actual rate of goals (2.32). The only caveat with this line is that they did see a higher share of offensive zone face-offs, which might be fine considering the next two line combinations spent more time starting in their own zone.

5v5 TOI CF% FF% xGF% GF%
Khaira-Sheahan-Archibald 90.08 54.19 52.22 48.26 26.23
Granlund-Haas-Russell 64.75 59.36 62.25 64.94 50.00

These results were surprising to me. In just under 100 minutes at even-strength, the combination of Khaira, Sheahan and Archibald posted a Corsi For percentage of 54%. Unfortunately, they weren’t able to convert their time with the puck into actual scoring chances as measured by expected goals, but it was still impressive considering how often Sheahan starts in his own zone. I wouldn’t trust this line to be a shut down line, but they should be able to match-up well against the other team’s bottom six and prevent bad things from happening.

The last line is the one I’m really curious about in a larger role. The trio has only played an hour together at even-strength, but they’ve been quite effective supporting the puck and just being in good spots at both ends of the rink. I assumed their Corsi For% and expected goal-share would have been fine, maybe around break even. Just didn’t think it would be above 55%.

What I especially like about the bottom six here is that the team has two players who have played center on each line that can split centermen duties depending on what side and what end of the ice they’re on. It was interesting to hear Tippett address faceoffs in his media availability when asked about Haas and Sheahan on a line together tonight.

Those two will flip back and forth depending on where the draw is, left side or right side. So they’ll both get a little bit of a look that way. Both are smart players, they can adapt to the other.

I’ve actually done a bunch of that stuff before way back to some world championships with a lot of players where centermen if you got a right side guy and a left side guy on a line it really gives you an advantage to your percentages in your own end especially. That’s the thinking behind there. Source: Edmonton Oilers

Going to back to an earlier point about chances for and against, over the last ten games the Oilers have posted an expected goals for percentage of 43.50% (30th in the league), generating 1.98 expected goals per hour, and allowing 2.57 expected goals per hour. Keep in mind, league average rates over the last three seasons is 2.25 for both metrics so the Oilers are very far from reasonable rates.

Below are how the proposed line combinations have done in their time together this season at even-strength when it comes to the rate of expected goals.

5v5 xGoals For/60 xGoals Against/60 xGF%
Draisaitl-McDavid-Kassian 2.72 2.96 47.93
Neal-RNH-Chiasson 2.43 1.57 60.79
Khaira-Sheahan-Archibald 1.87 2.00 48.26
Granlund-Haas-Russell 2.82 1.52 64.94

All four line combinations have posted an expected goals-for share above their teams recent share. Three of the four line combinations have posted an expected goals-for per hour rate above what the Oilers have posted recently and closer to league average rates. And three of the four lines post an expected goals against rate below what the Oilers have recently posted and league average rates.

Again – these potential line combinations won’t win championships, but they have results this season that indicate that they can help the Oilers improve their chances of outscoring opponents today. The team is still in need of talent up front and could use more speed and finishing skill. Those issues aren’t likely to be addressed anytime soon, and it doesn’t appear that the young prospects will get a look either. But for now, the Oilers can use their present roster to help improve their share of shots and chances and bring their underlying numbers back to respectable levels. They’ll need any edge they can find to hold on to a playoff spot.

Data: Natural Stat Trick

Appendix 1: Line combinations details

5v5 TOI CF% FF% xGF% GF% On-Ice SH% On-Ice SV% PDO Off. Zone FO%
Draisaitl-McDavid-Kassian 399.25 47.66 46.51 47.93 57.40 13.03 91.97 1.050 50.38
Neal-RNH-Chiasson 103.47 59.38 58.41 60.79 50.00 6.95 90.23 0.972 63.93
Khaira-Sheahan-Archibald 90.08 54.19 52.22 48.26 26.23 4.67 85.30 0.900 33.33
Granlund-Haas-Russell 64.75 59.36 62.25 64.94 50.00 6.69 90.72 0.974 43.33

 

Sliding

22316220.jpg

The Oilers are sliding, and they shouldn’t be surprised.

The good news is that the Oilers currently rank second in the Pacific with 40 points, and fifth in the Western Conference. The problem is their overall results aren’t good enough, especially at even-strength, and they’ve been trending downwards for a while.

20191213 - Goal differential

While the powerplay and penalty kill continues to thrive, it’s the even-strength (5v5) play that is dragging down the Oilers overall results. Below is the cumulative total of goals-scored, which now sits at -12 – a goal-share of 45.77%, good for 26th in the league.

20191213 - Goal differential at 5v5.png

The underlying shot-share numbers have been poor, and appear to be getting worse. While they did show some signs of life only a few weeks ago, their overall play has been declining ever since. I wrote last week that I think it had to do with the team potentially trying to generate more offence by focusing less on defensive play, but that’d be something only the coach could confirm.

Point % CF% FF% xGF% GF% SH% SV% PDO
0.588 47.88
(26th)
48.43
(22nd)
48.90
(22nd)
45.77
(26th)
8.54 90.68 0.992

Below are the Oilers Corsi For%, Fenwick For% and xGoals For% over rolling 10-game segments this season. A glossary describing the metrics can be found below.

20191213 - Shot share at 5v5

The last ten games have been extremely poor, with the Oilers posting shot-share metrics well below league average levels. The expected goal-share has slid down to 45.0%, which again might be due to the team taking more risks in an attempt to generate offence. The Oilers were doing something right early on, but it appears they’ve adjusted their tactics in an attempt to generate offence.

The other concerning issue is the team’s performance at even-strength both with and without McDavid this season. In year’s past, we would see all the shot-share numbers be above at least 51.0% with their best player on the ice. This year, it’s been a different story, as the Oilers even with McDavid have posted shot-share numbers below 50.0%.

20191213 - WOWY 97.png

The good news is that the overall goal-share has been outstanding with McDavid. But it does make you wonder how much better his on-ice goal-share and point totals would be if the Oilers were spending less time in their own zone and more time with the puck. The lack of skill on the roster, combined with a blueline that has only a few puck-movers, is what I think is driving the shot-share numbers downward.

The last ten game have been especially concerning, with the whole team including McDavid struggling to generate offence and spending more time defending. Below we see that the shot-share metrics are closer to the 45.0% percent range (!), with the goal-share being below the break-even mark even with McDavid.

20191213 - WOWY 97-10

The Oilers have posted a -1 goal differential with McDavid, and a -12 goal differential without him over the last ten games. Not even the special teams can bail them out as the outstanding powerplay has scored ten goals (and allowed one) over the last ten, while the penalty kill has actually struggled allowing six goals.

Can’t say their current results are too surprising as the team lacks skill and depth, and their shot-share metrics have been below average for most of the season. The question again comes back to the Oilers management and how they view the overall results and where the team has been trending.

With the team having accumulated 40 points, does management think the results are real and sustainable? Or do they recognize that they’re lacking a significant amount of skill and depth in all positions, and start to add pieces, either from their own development system or through the trade market. Thinking heading into the season that this was going to be a development year, and based on their actual results and underlying trends, I would expect the team to hold off on making any drastic changes and start to give some of the younger prospects a chance to play in Edmonton and further their development.

It would have been nice if the team took a positive step and posted underlying numbers that demonstrated sustainability. But the reality is that they’re not good enough to compete for a championship and should be using this season to properly evaluate the prospects that they do have.

Data: Natural Stat Trick

Also, I made an appearance on CBC Radio Active this week. Will post the link when it’s available: CBC Radio Active (2019, December 11)

Glossary

  • Corsi For percentage (CF%) – The proportion of all the shot attempts the team generated and allowed that the team generated (i.e., Corsi For/(Corsi For + Corsi Against). This is used as a proxy for possession and can predict a team’s future share of goals (GF%).
  • Fenwick For percentage (FF%) – The proportion of all the unblocked shot attempts the team generated and allowed that the team generated (i.e., Fenwick For/(Fenwick For + Fenwick Against). This is used as a proxy for shot quality and considers shot blocking a repeatable skill. It can also predict a team’s future share of goals, slightly better than Corsi.
  • Expected Goals For percentage (xGF%) – This is a weighting placed on every unblocked shot based on the probability of the shot becoming a goal. This depends on the type of shot, location and uses historical shot and goals data to come up with the probability for each unblocked shot.
  • Goals For percentage (GF%) – The proportion of all the goals that the team scored and allowed that the team generated (i.e., Goals For/(Goals For + Goals Against).
  • Shooting percentage (SH%) – The percentage of the team’s shots on goal that became goals (i.e., total goals divided by the total shots on goal).
  • Save percentage (SV%) – The percentage of the team’s shots on goal against that were saved (i.e., 1-(totals goals allowed divided by the total shots on goal against))
  • PDO – The sum of a team’s shooting percentage (SH%) and its save percentage (SV%). It’s based on the theory that most teams will ultimately regress toward a sum of 100, and is often viewed as a proxy for how lucky a team is. (Source)