Hockey Analytics: Divisional Drive-By

Gus Katsaros
·6 min read


This content is not available due to your privacy preferences.
Update your settings here to see it.


The NHL schedule has taken some rough hits on the nose during the first quarter of the COVID filled NHL 2020-21 season, but most teams have now passed the 14 game first quarter mark. Despite some stragglers, and as games start to ramp up again, enough of a sample size exists to have an initial view of each of the divisions. Restricting the views to divisions is much different this season, due to COVID restrictions.

This season doesn’t fully translate to the true strength/weaknesses of teams overall, with competition limited to inter-divisional play, but there are distinct favorites emerging to shape up the playoff picture fairly early.

Challenges will exist during the playoffs that didn’t in previous season, except for the 1994-95 season where teams only played within their conference and teams from opposing conferences met only once – in the Stanley Cup Finals.

This season is still different, however, in that during those seasons, teams at least had a wider field of opposition – and got special looks at each, even if the schedule didn’t seem balanced they played outside of the division.

With teams restricting play to within the division, come playoff time, the first two rounds will be familiar but when teams cross over to other divisions, we should see more unfamiliarity with team play.

Data for this article was sourced at NST and doesn’t include games played on Tuesday Feb 16.

DIVISIONS

In normal times, I divvy up the season with a 20, 21, 21, 20 split to represent quarters. In a 56-game season – assuming everyone does end up playing 56 games – 14 games marks the end of the first quarter. As stated before, not every team has played the same number of games, but when looking at total division games, they’re all within an acceptable range. COVID – and snow – will further delay games.

NHL Games Played 2020-21
NHL Games Played 2020-21

The disparity in games/minutes played requires a reliance on rate stats instead of straight counting numbers. Rate stats are normalized over 60 minutes – or any value as long as it makes sense, which 60, does as a full game – and despite the small sample size, there’s at least some normalcy in evaluating performance over a larger population.

The North has played the most minutes, affected less by the delays in scheduling and in a slightly safer environment in Canada than the wider open United States.

The chart below outlines rate stats (per -60) for even strength and power play situations. Included in the metrics are shots (SF), shot attempts (CF), goals for (GF), expected goals (xGF), and scoring chances (SCF & HD – high danger).

There are some minor discrepancies, with just straight visuals, like the North division producing the most shot attempts per 60, at both 5v5 and 5v4 – and gets to high danger areas the most at both strengths. The West runs the shot attempts at 5v4.

NHL High level shot metrics 2020-21
NHL High level shot metrics 2020-21

Shot attempts lead into scoring chances and location plays a part in whether there’s a goal scored. Getting closer to the net increases the propensity for scoring goals, but as I’ve alluded to in the past, shot assists, and the study of passing and the effects on scoring have indicated the value of pre-shot movement to influence the shot as well as location. This is a great resource for a more in depth dive into pre-shot expected goals models.

When viewing heat maps, keep in consideration that a ‘shot’ may have been made possible only due to the series of events prior to that shot being fired.

NHL Scoring Chances Per 60 2020-21
NHL Scoring Chances Per 60 2020-21

The chart above provides a clue as to the effectiveness of special teams in the West. The West division leads others in low, medium and high danger scoring chances per 60 minutes. The North seems to have abandoned the concept of shooting from the points and low danger areas, lagging other divisions by a fair margin. This is evident at both 5v5 and 5v4, with more medium and high danger chances replacing the low danger threats.

Scoring chances give way to goals scored and with expected goals models, we can assess the expected value of shots to measure against actual (observed) goals to calculate a differential that applies a narrative to overall performance.

Editor’s Note: Drafting is only half the battle! Get an edge on your competition all season long with rankings, projections, trade evaluator, lineup adviser and much more with our NHL Season Tools. Click here to learn more!

By division, focusing scoring chance generation on 5v4 and 5v5, we see the West leading the way in goals per 60 minutes, outpacing their expected goals (xGF). In the chart above we saw a mediocre metric for scoring chances coming from the East, but when switching to goals for and expected goals, the East has had the best power play goals rates, outpacing their expected goals by 11 goals.

NHL Expected Goals 2020-21
NHL Expected Goals 2020-21

In the West, they are scoring at the third best pace for goals, while generating the highest expected goals per 60 – while taking more perimeter shots and relying on rebounds and net front presence to augment scoring.

So, to recap, in the West, they’ve had general goal scoring success, have generated the highest expected goals, and fire most scoring chances from low danger scoring areas.

Below we see each game situation shooting and save percentage by division and recall the amount of low danger chances being generated at 5v4 by teams in the West, the result is lagging other divisions in shooting percentage.

Once scoring chances have been considered, we measure effectiveness by using shooting and save percentage. Overall shooting percentage at 5v4 is in reverse order for save percentage at 4v5. The West has had the least amount of power play luck and had timely goaltending while shorthanded.

Of note –in the North they are generating less low danger scoring chances, but keeping with the evolution of hockey, they’re presenting an impressive 10.4% shooting percentage while shorthanded. Successful teams will try to find ways to score goals in every situation. It’s why the Maple Leafs have given Mitch Marner penalty killing duty in 2019-20 and even started giving Auston Matthews some PK time – even if it’s specifically designed for face off management. Win back the puck, start from an offensive position and force the power play to play defense – that is, try to retrieve the puck to set up the power play. This idea is becoming much more mainstream and we should become more aware of this as we observe going forward.

Teams that utilize a ‘power-kill’ mentality can capitalize offensively when down a man and it’s encouraging to see more scoring chance generation even in what is supposed to be a defensive situation.

NHL Shooting and Save Percentages 2020-21
NHL Shooting and Save Percentages 2020-21

What the second quarter holds is still up for grabs. Some teams games delays are being rescheduled and as the disparity in the division becomes more evident, we will see a much clearer playoff picture emerge. In the meantime, we can periodically check in on divisional play to measure results across a greater spectrum of sample sizes using rate stats.

On to the second quarter.