#2. Assist Environment

As straightforward as the concept may be to a hockey fan today, assists have an interesting origin story. Despite the fact they are less direct and easier to earn, both primary and secondary assists are awarded the same respect as goals on a scoresheet – one point. However, the most overlooked aspect of assists is that they are inextricably linked to the frequency of goals. We’ll examine that history as part of our exploration of Hockey Reference’s second adjustment.

When it comes to awarding assists in the present day, the NHL loosely defines one in its legislation under Rule 57: “When a player scores a goal, an ‘assist’ shall be credited to the player or players taking part in the play immediately preceding the goal, but no more than two assists can be given on any goal. Each ‘assist’ shall count one point in the player’s record. Only one point can be credited to any one player on a goal.”

It’s been accepted in modern times to mean that assists are distributed to the last two teammates of the goal scorer that made any contact with the puck. Whether artfully or incidentally, whether after a series of passes or directly, whether they are still on the ice or have left it, so long as the other team has not controlled the puck in between, an assist is earned. A convoluted concept, if we’re being honest. Aided with slow motion video replay and multiple camera angles, the awarding of assists is an easier process compared to previous generations. Historically, some combination of the referee’s memory and on-ice position, combined with the honesty of the players, were the only factors available to dish out assists.

Before we account for the second adjustment, a brief history of the NHL’s distribution of assists is in order. Unlike goal scoring, when it comes to assists, there are no wild annual fluctuations across the league’s recent timeline. In the NHL’s last 64 seasons, assists per goal in each year are exclusively between 1.61 and 1.75 – a reasonable range of approximately 8%. It is the formative years of the league where issuance of assists is the most volatile.  In the league’s first four decades, its priorities included establishing itself as a viable business entity, navigating two World Wars, creating legitimate rules that generated excitement, and overseeing the health and viability of its fleeting franchises.  Unsurprisingly, accurately defining and tracking assists garnered less buzz.

The chart below encapsulates the NHL’s distribution of assists per goal since its inaugural season.  The dotted line is fixed at the 1.67 assist per goal mark (i.e. 10 assists on six goals per game) that will serve as our established baseline for assist adjustments.

NHL AVERAGE ASSISTS PER GOAL, 1917-18 TO 2020-21

The trendline above highlights that the NHL’s pioneering players’ assist totals were severely suppressed.  In its debut season, the league awarded only 145 assists on 342 goals – a stingy average of 0.42 assists per goal.  The NHL is roughly four times more generous in awarding assists on a goal in 2020-21. The league’s most efficient set-up man in 1917-18, Reg Noble of the Toronto Arenas, earned a measly 10 assists in 20 games played. Assists were scarce.

There is little evidence to suggest that in the NHL’s first quarter century that assists were well monitored, uniformly assigned, or that there was significant intrigue in the statistic.  It remains unclear exactly how assists were dished out. Was a direct pass required? Did the sequence require uninterrupted passing plays for multiple assists? Did shots generating rebounds count as assists if resulting in a subsequent goal? Did a player need to visibly contribute to a goal via a subjective level of impact? Further complicating matters, of course, exists the partiality of scorers from city to city, given limited communication amongst the league’s recordkeepers. Though documentation of official assist rules in the league’s first decade are unavailable, secondary assists were not awarded until the late 1920s; this spike is clear in the chart above. Even primary assists were scarce in those days – it would take until the 15th season of NHL play in 1931-32 until league-wide assist totals surpassed goal totals. 

A key development in the history of assist assignment occurs via the donation of the Art Ross Trophy in 1947.  The annual award, assigned to the league’s leading point scorer, brings new attention to league scoring totals. With goals typically straight forward to assign, assists inevitably become a more scrutinized piece of the point scoring puzzle. Though a direct link cannot be proven, assists continue to rise after the introduction of the trophy. In the season before the Ross is first awarded, assists per goal measures 1.30; within five seasons, the rate is 1.50, and within 10 seasons, the rate is 1.61.

The assist needle has moved very little since; in fact, each of the last six seasons has either been a rate of 1.68 or 1.69 assists per goal. All data suggests assists have become predictable in their allotment.  Video replay closely examines every possible touch. There is little reason to expect fluctuation from the current assist level, barring significant changes in the definition.

When it comes to assist totals, however, they cannot be analyzed in isolation.  As we’ve learned, the number of assists per goal the NHL assigns historically is critical to the exercise, but adjustment #1 (goal environment) is also extremely important here.  We can adjust assists up or down based on NHL patterns through history, but assists are only earned when goals are scored – they are not mutually exclusive tallies. So, our lessons on scoring environment apply to adjusting assists. 

With that in mind, let’s look at the historical plot of assists per game – this chart factors in the generosity of assist assignment and the goal environment in each season.  With our 10 assists per game baseline highlighted below, we obtain a more accurate sense of the true volume of assists distributed by year.

NHL AVERAGE ASSISTS PER GAME, 1917-18 TO 2020-21

Consider the 1928-29 season, now infamous for being the rock bottom of NHL offence – 2.92 total goals per game. A skater in that time period not only had to earn assists when the NHL awarded only 0.60 of them per goal, but with such a suppressed goal environment, the impact on assists is significantly compounded.  That season, there were only 1.75 total assists per game – not per goal, but rather per game. In 1985-86, Wayne Gretzky himself earned more assists per game (2.04) than both teams combined in an average NHL game in 1928-29. For further perspective, Hall of Famer and celebrated catalyst Frank Boucher paced the league in assists with 16 in a 44-game season. Simply adjusting assists for how they were awarded at the time is therefore insufficient on its own – the goal environment of the day greatly impacts its assist totals.

With a review of the NHL’s history in dishing out “apples” to its players, combined with the interconnectedness of goal scoring trends on assists, how does this affect our adjustments?  Hockey Reference establishes our adjusted baseline at 1.67 assists per goal.  The figure is both logical and keeps things as simple as possible.  Over 100-plus seasons, roughly 300,000 goals and 500,000 assists, the NHL has awarded an average of approximately 1.68 assists per goal.  But by using 1.67 as its baseline, combined with the six goal per game baseline established for goal environment in the previous chapter, we have a nice and even figure of 10 to play with (i.e., 6 goals per game x 1.67 assists per goal = 10 assists per game). In its simplest terms, the adjustment to assists places every season in history into a setting where there are exactly six goals and 10 assists per game. This closely mirrors the present-day scoring environment – 2020-21 featured 5.96 goals per game and 10.00 assists per game, and 2020-21 featured 5.80 goals per game and 9.81 assists per game.

---

Let’s demonstrate through the Hockey Reference example of Gordie Howe‘s 1952-53 season again, this time focusing on the legend’s 46 assists that year. To calculate the adjustment, we divide 10 by the league average assists per game, excluding Howe himself.

The 1952-53 season featured 1,513 assists over 210 games played; removing Howe’s own impact from that year, we get:

• (1,513 – 46) ÷ 210 = 6.99 assists per game

To arrive at our assist assignment adjustment, we get:

• 10 ÷ 6.99 = 1.43

Howe notched 46 assists, which using the adjustment, calculates:

• 46 x 1.43 = 66 assists

This adjustment equates Howe’s 46 assists in 1952-53 to 66 assists in a neutral environment for earning assists.  We’ve neatly factored in the goal environment and the assist environments in one fell swoop.

Adjustment #2: Assist Environment                        

Adjusts a player’s assist total to an assist environment of 10 assists per game

Howe’s output of 49 goals, 46, assists, and 95 points blossoms to 65, 66, and 131, respectively.  The adjustments launch upward faster than an errant elbow from Howe himself.  Again, Howe operated in a 70-game NHL schedule in that era, so his 66 adjusted assists are across only 70 games at this point.  But we must consider that we are only halfway through the four adjustments – Mr. Hockey will be impacted further by tweaks for schedule length and roster size.

Via the second adjustment, we’ve learned that assists were awarded casually and scarcely in the NHL’s earliest days, ultimately leveling off in recent decades at around five assists for every three goals scored. Like goal scoring, adequately adjusting for the chasm in assists produces drastic changes to the historical record. This wraps the two more complex of the four Hockey Reference adjustments, with season length and roster size our final two tweaks in building adjusted point totals.

Data from Hockey-Reference.com.