Scoring levels of the "pack" over the past 30 years in relation to League GPG

[daver]

I think the concept of the "pack" of elite offensive players is generally accepted, and how much a player separated themselves from the "pack" is generally accepted as the best way to compare players from other seasons/eras. The challenge is defining what constitutes the "pack" over league history with different league sizes.

Starting in '93/94, the league continued it's drop in overall scoring after a crazy 92/93 season. I measured the "pack" as being the average point total of the #10 scorer, the #25 scorer and the #50 scorer and compared that to the league GPG. I excluded seasons shortened by lockouts and Covid. As the # of teams increased, so does the # of the player used in the calculation. E.g. 2023/24 I used the #12 scorer, #30 scorer, and the #60 scorer.

Here is how the first seasons looked:

93/94 - League GPG: 3.24, scoring level of the "pack - 86 points (avg. scoring of Top 9 - 110)

95/96 - League GPG: 3.14, scoring level of the "pack - 87 points (avg. scoring of Top 9 - 116)

96/97 - League GPG: 2.92, scoring level of the "pack - 75 points (avg. scoring of Top 9 - 99)

97/98 - League GPG: 2.64, scoring level of the "pack - 67 points (avg. scoring of Top 9 - 89)

98/99 - League GPG: 2.63, scoring level of the "pack - 67 points (avg. scoring of Top 9 - 100)

99/00 - League GPG: 2.75, scoring level of the "pack - 69 points (avg. scoring of Top 10 - 86)

..

03/04 - League GPG: 2.57, scoring level of the "pack" - 65 points (avg. scoring of Top 10 - 75)


Comment- There is a clear, and not surprising, connection between league scoring levels and the scoring level of the pack. Generally speaking ES scoring decline was the main reason for the league drop. PP scoring was pretty consistent. There seems to be a somewhat disproportionate effect on the scoring levels of the pack vs. the overall drop in league GPG. I.e. the pack dropped a bit more %-wise. From '96/97 to '03/04 (the DPE), the GPG averages 2.66 and the "pack" averages 69 points


IMMEDIATE POST LOCKOUT


05/06 - League GPG: 3.08, scoring level of the "pack - 79 points (avg. scoring of Top 10 - 105)

06/07 - League GPG: 2.95, scoring level of the "pack - 79 points (avg. scoring of Top 9 - 104)

--

09/10 - League GPG: 2.84, scoring level of the "pack - 71 points (avg. scoring of Top 9 - 96)


Comment- League scoring, and the "pack" scoring take significant jumps as PP opportunities take a significant jump. As PP opportunities decrease so does league scoring and the "pack". It seems clear that ES scoring has not really been affected by league changes to open things up.


DPE 2.0

10/11 - League GPG: 2.79, scoring level of the "pack - 67 points (avg. scoring of Top 10 - 89)

11/12 - League GPG: 2.73, scoring level of the "pack - 68 points (avg. scoring of Top 10 - 87)

--

16/17 - League GPG: 2.77, scoring level of the "pack - 65 points (avg. scoring of Top 10 - 84)


Comment - From '10/11 to 16/17, the GPG averages 2.75 and the "pack" averages 66 points. There appears to be a clear difference in the "pack" scoring from the DPE that can be attributed to the difference in PP opportunities (an average of 4.35 in the DPE vs. 3.21 in the DPE 2.0).

There is also a clearer disproportionate effect on the scoring levels of the "pack" vs. the overall drop in league GPG in comparison to the DPE. In the five seasons that league GPG was above 2.90, the average GPG was 3.07 and the pack scoring was 81. During DPE 2.0, the "pack" scoring decreased 10% more than expected (expected was 73 points).


POST DPE 2.0

17/18 - League GPG: 2.97, scoring level of the "pack - 74 points (avg. scoring of Top 11 - 95)

18/19 - League GPG: 3.01, scoring level of the "pack - 79 points (avg. scoring of Top 11 - 104)

--

3.11 - League GPG: 3.11, scoring level of the "pack - 80 points (avg. scoring of Top 11 - 111)


Comment - From '17/18 to 23/24, the GPG averages 3.08 and the "pack" averages 79 points. The "pack" has returned to scoring levels very similar to those seen in the five other seasons where the league GPG was above 2.90. The difference this time is an increase in ES scoring as PP opportunities have not wavered too much since 14/15.


OVERALL COMMENT

Scoring for the "pack" was the most difficult during the DPE 2.0. Accordingly, it was the most difficult period for superstar talent to separate themselves from the pack. Leading scorers in the DPE 2.0 and the DPE are handicapped in terms of relative dominance vs. the leading scorers in higher scoring seasons. Both eras had multiple seasons of "meh" level of leading scorers where under 100 points was winning Rosses and noone was getting close to 50 goals despite many of the same players, still in their primes, doing just that in seasons before and after the DPE eras.

Click to read more...

 

[MadLuke]

Scoring for the "pack" was the most difficult during the DPE 2.0. Accordingly, it was the most difficult period for superstar talent to separate themselves from the pack.

Not sure that automatically follow ? For the pack to separate themselve from second liner would sound maybe more obvious than for superstar to separate themselve from the pack, historically Howe-Jagr-Crosby separation from the pack is in some way pretty much almost exactly the same, prime non peak McDavid/Gretzky/Lemieux has well.

It is a bit hard to know what we are looking at (or the added information if you will) here

86/3.24 = 26.54
69/2.63 = 26.23

Without calculating I would not have known mentally how close those 2 situation were and now I am not sure what it mean.

 

[DitchMarner]

I would say the statistical information and conclusions presented in this thread allude to the question I posted in a thread I made on the Numbers board a little while ago, that being: how much does the lack of differentiation between different games states matter when adjusting point totals?

Of course, I was referring specifically to HockeyReferences adjusted stats, but the question can be asked about any adjustment method in general.

No Distinction Between Different Games States In HockeyReference's Adjusted Formula - How Much Does This Matter?

I like HR's adjusted formula for valuing goals and points. Points and goals are worth more (ie more valuable) when scoring across the League is lower. This is simply a basic mathematical fact. There is really no subjectivity involved. A 50 goal season in a league where 8.02 goals are scored on...

forums.hfboards.com

In general, League-wide scoring scales fairly intuitively with scoring at the highest levels. If scoring across the League drops, you generally will see lower point totals from the top scorers than you will in seasons where scoring across the League is very high. This trend confirms that common adjustment methods aren't total voodoo and do have some indicative value.

The Power Play Bias Effect (ie top stars in certain eras collectively receiving more PP time than the top stars in other eras and thus having a good opportunity to score a greater percentage of the League's goals compared to other eras) is something that I'm not sure any adjustment method accounts for.

 

[DitchMarner]

Not sure that automatically follow ? For the pack to separate themselves from second liner would sound maybe more obvious than for superstar to separate themselves from the pack, historically Howe-Jagr-Crosby separation from the pack is in some way pretty much almost exactly the same, prime non peak McDavid/Gretzky/Lemieux has well.

It is a bit hard to know what we are looking at (or the added information if you will) here

86/3.24 = 26.54
69/2.63 = 26.23

Without calculating I would not have known mentally how close those 2 situation were and now I am not sure what it mean.

Well, it's one thing for the star and superstar level players to separate themselves from the pack and another for the very best scorers to do it. If a talent is truly transcendent, shouldn't he be able to dominate his peers in scoring to a large degree regardless of how hard it is for run-of-the-mill stars and superstars to do that?

For example: if Connor McDavid is able to have a 153 point season and finish 40 points ahead of the third and fourth leading scorers in the League (who both played all 82 games) - that being equal in difference in scoring between the third and fourth scorers and the 48th leading scorer (73 points) - does that suggest that it is easy/easier for top scorers to separate themselves from the pack in this era or just that it's possible for him to do it in a way we haven't seen recently?

He also won a scoring Title in 56 games in a season where it would have taken the third leading scorer 82 games to outscore him.

If players from other eras were unable to achieve this type of dominance, is it because of some sort of scoring environment disadvantage that we have to try to unearth or simply because they were not as good as he is offensively?

 

[daver]

For example: if Connor McDavid is able to have a 153 point season and finish 40 points ahead of the third and fourth leading scorers in the League (who both played all 82 games) - that being equal in difference in scoring between the third and fourth scorers and the 48th leading scorer (73 points) - does that suggest that it is easy/easier for top scorers to separate themselves from the pack in this era or just that it's possible for him to do it in a way we haven't seen recently?

Except Kucherov (and to a slightly lesser degree, MacKinnon), did just that this past season. Kucherov was only slightly behind McDavid's level of domination over the #10/#25 and #50 scorer.

 

[daver]

The Power Play Bias Effect (ie top stars in certain eras collectively receiving more PP time than the top stars in other eras and thus having a good opportunity to score a greater percentage of the League's goals compared to other eras) is something that I'm not sure any adjustment method accounts for.

There is no adjusting, just comparing relative dominance of respective peer groups. These peer groups played in the same conditions as you noted. It should not matter that Player A from Era C got more PP time (and more PP points) than Player B from Era D; what matters is where they stood among the other players who played in the same league conditions.

As for the OP, the amount of PP time clearly affected the scoring level of the pack; to a slightly lesser degree overall league scoring.

 

[VanIslander]

Noise.

 

[daver]

I added in the average Point total of the Top 9/10/11 point scorers in the OP.

In the five seasons between 1993/94 and 2006/7 when league GPG was above 2.90: the average GPG was: 3.07, the pack scoring was 81, and the avg. Top 9/10 scoring was 107

During DPE, the average GPG was 2.66, the pack scoring was 69, and the avg. Top 9/10 scoring was 89.

During DPE 2.0, the average GPG was 2.75, the pack scoring was 66, and the avg. Top 10 scoring was 85.

From '17/18 to 23/24, the GPG averages 3.08 and the "pack" averages 79 points and the avg. Top 11 scoring was 105

COMMENT

In the ten seasons between 1993/94 and 2023/24 when league GPG was above 2.90: the average GPG was: 3.06, the pack scoring was 80, and the avg. Top 9/10 scoring was 106.

In the thirteen seasons between 1993/94 and 2023/24 when league GPG was below 2.80: the average GPG was 2.70 (88% lower), the pack scoring was 68, (85% lower) and the avg. Top 9/10 scoring was 87 (82% lower).

It seems like when the league tightened up, it brought more parity among forwards, among Top 6 forwards, and among star/superstar talent. This makes sense as tighter games means tighter checking, less risky plays, more even TOI etc..

 

[WarriorofTime]

I think the concept of the "pack" of elite offensive players is generally accepted

I have no idea what this means and why it would be "generally accepted"

 

[daver]

I added in the average Point total of the Top 9/10/11 point scorers in the OP.

In the five seasons between 1993/94 and 2006/7 when league GPG was above 2.90: the average GPG was: 3.07, the pack scoring was 81, and the avg. Top 9/10 scoring was 107

During DPE, the average GPG was 2.66, the pack scoring was 69, and the avg. Top 9/10 scoring was 89.

During DPE 2.0, the average GPG was 2.75, the pack scoring was 66, and the avg. Top 10 scoring was 85.

From '17/18 to 23/24, the GPG averages 3.08 and the "pack" averages 79 points and the avg. Top 11 scoring was 105

COMMENT

In the ten seasons between 1993/94 and 2023/24 when league GPG was above 2.90: the average GPG was: 3.06, the pack scoring was 80, and the avg. Top 9/10 scoring was 106.

In the thirteen seasons between 1993/94 and 2023/24 when league GPG was below 2.80: the average GPG was 2.70 (88% lower), the pack scoring was 68, (85% lower) and the avg. Top 9/10 scoring was 87 (82% lower).

It seems like when the league tightened up, it brought more parity among forwards, among Top 6 forwards, and among star/superstar talent. This makes sense as tighter games means tighter checking, less risky plays, more even TOI etc..

Added in the #100 scorer (#126 in 2023/24) into the mix to expand the data.

In the ten seasons between 1993/94 and 2023/24 when league GPG was above 2.95: the average GPG was: 3.06, the average points for the #100/#126 scorer was 52

In the thirteen seasons between 1993/94 and 2023/24 when league GPG was below 2.80: the average GPG was 2.70 (88% lower), the average points for the #100/#126 scorer was 46 (88% lower)

This seems to be more evidence that when ES scoring tightened up, it was the elite scorers who were affected the most.

It is also worth noting that during the DPE, scoring by the #200 player (#246 in 23/24) averaged 30. Since then, it has not risen above 34, (it averaged 31 during DPE 2.0) most notably in the last five years when scoring by the #115 player has risen by a greater % (along with scoring by everyone else above).

So it seems easier for the elite offensive talent to separate themselves in recent years as ES scoring has shown a significant increase. This goes back to the other thread of wondering how to view Kucherov's and MacKinnon's 23/24 seasons vs. some of the stronger DPE and DPE 2.0 seasons (Jagr, Crosby, Kane).

Was it possible for superstar talent during the DPE, and seemingly moreso in the DPE 2.0, to have "Best of their era"/"Among the best Non-Big 4" level seasons?

 

[daver]

Was it possible for superstar talent during the DPE, and seemingly moreso in the DPE 2.0, to have "Best of their era"/"Among the best Non-Big 4" level seasons?

There are some interesting questions to ask?

Was peak Jagr closer to Mario given we saw a glimpse of prime Mario as the DPE was setting in 96/97 when scoring took a drop? Or was Mario half checked out that season. And what about a 35 year old Mario in 00/01 when he matched Jagr's PPG when they played together? Was he just as good as peak Jagr or were his numbers inflated due to the PP and playing with a peak Jagr?

Should Crosby and Kane DPE 2.0 dominant Ross wins get more attention? Kane's Art Ross win was not unexpected as he was knocking on the door (probably longer than one would think) on being an Art Ross threat since 2010, and made some noise after 15/16. But did he have a truly dominant Art Ross win in him? We could be asking the same thing about a 30 year old Kucherov and MacKinnon this year.

Crosby's 13/14 was not quite at the level of his previous three partial seasons also during DPE 2.0, which adds more questions than answers, but begs the question of how he would have fared if he played at his peak in the recent seasons with the major uptick in ES scoring.

It will be interesting to see if the trend towards higher scoring, and higher ES scoring, holds and if other star forwards start putting up numbers that throw some doubt on generally accepted narratives:

Before this season:

Peak Kucherov and peak Kane are very close

Peak MacKinnon was not particularly close to peak Malkin and peak Crosby

Peak McDavid is not "gettable" like Wayne/Mario