TheStatican
Registered User
- Mar 14, 2012
- 1,714
- 1,452
A discussion in the Lemieux as the GOAT thread got me thinking about taking a another look at the Gretzky vs Lemieux debate. I found myself disagreeing with a number of members but my argument is not about who is the Goat but rather focuses on Peak Performance. Fair warning, this post will probably set records for the use of statistics, hey I don't call myself "TheStatican" for nothing. Simply put I'm a fan of statistics because they don't lie, the only thing in question is how much relevance one chooses to place on them.
Peak, longevity & consistency, are factors that all matter when it comes to the Goat debate. Lemieux had an extremely high peak but due to injuries he did not consistently play at that level year in year out. Gretzky meanwhile maintained his peak at least 5 years straight with very little deviation. Gretzky is also clearly ahead of Lemieux in overall longevity and every one else in that department, save for a player like Howe. Mario would need to have more than just a little edge in peak performance to overcome the differences in longevity & consistency. But anyway you cut it the difference in peak between them is too small to change any of the other factors and so I agree that there is no argument for Lemieux as the hockey Goat.
In terms of peak performance there are those who also feel that matter has long since been settled as well. Certainly if you just take the raw numbers at face value this would be correct. Not much point in reading beyond this point as you already have your answer then. Some say it's not just the records but the fact he reached his peak level repeatedly in many years. Well, that's precisely what makes Gretzky the GOAT. No one else in the history of the game had his combined level of consistency and peak performance. But that alone does not mean that Gretzky's absolutely peak was the highest of all time, that is confusing longevity of peak, for peak performance. I would agree I'd be more likely to view Lemieux's peak performance as suspect if he only did it once. But he did it twice, that to me proves it was no "fluke season" the first time.
If you ask very casual hockey fans they're likely to tell you that the raw totals are what matters most and they do so because the NHL continues to reinforce this belief. Hockey is the only major North American sports that focuses almost exclusively on raw counting stats when it comes to awarding the best offensive performers in a season. The Art Ross and the Rocket are simply awarded to the players who score the "most points" and "most goals" respectively in the league during a particular season. These are awards which serve to dramatically increase the value of a players legacy and yet winning those awards does not necessarily mean the player in question was actually the best point producer or goal scorer in the league during that season. The problem lies in the concepts of performance and value; people commonly confuse the two. Performance is exclusive of value and value is simply what happens when a level of performance is sustained over a certain period of time. But people like it when things are kept simple and so the NHL continues to keep it simple for them.
Let's consider Alexander Mogilny's and Teemu Selänne's 1992-93 season, that year they both scored 76 goals. I consider Mogilny as having the superior goal scoring season of the two. This is because he scored his 76 goals in 77 games while Selänne did so over the course of 84 games. Selanne had 73 goals in his best 77 game stretch that season, impressive no doubt but Mogilny bested that by scoring 73 goals in a 68 game stretch. On a per-game basis Mogilny's scoring rate was notably higher; 0.987 vs 0.905 but the NHL instead essentially sees no difference between the two and recognizes both as the winners of the goal scoring title without any attention given to their games played. I believe this to be a faulty and incorrect assessment. Furthermore, neither Mogilny nor Selänne were even the best goal scorer in the league that season. That honor belonged to Mario Lemieux who scored 69 goals in 60 games(and really even less than that), easily besting both of them by scoring at blistering pace of 1.150 goals per game.
So how can the true measure of a players performance be determined if not by raw totals? The 'eye test' is one way but unfortunately it can easily be influenced by personal biases and it naturally fades with time. While it can be useful in supporting an argument it's far from a definitive fact. Statistics are definitive, however while we can make use of statistics as an absolute we must first understand which statistics hold the most value. The most obvious distinction between statistics are those that measure a quantity of production and those that measure a rate of production. A counting stat is one that measures a player's total production without reference to his number of opportunities. Common counting stats include goals, points and shots. While the most common counting stats are raw numbers some sites provide sabermetric stats that are also expressed the same way, such as hockey reference's point share. In contrast to counting stats, rate stats try to show a rate of productivity by dividing the number of successes or failures by the number of opportunities. Most rate statistics will include a term like "average" or "percentage" in their name as in indication that they are the result of a division process. Some examples of traditional rate stats including shooting percentage and goals against average. It's important to note that many counting stats are given with an implicit number of opportunities, like a season or career. This kind of implicit number of opportunities may or may not balance out. While two 1st line centers will be seen as playing in very similar conditions to facilitate the collection of goals and assists, the same is not true of a center and a defensemen so a comparison of their counting stats is not valid which brings me to Bobby Orr... There are a fair number of people who believe that it is neither Gretzky nor Lemieux but rather Orr who had the highest peak in NHL history. Unfortunately it's difficult to quantify his value in comparison to the other two using the statistics we have available to us because most of the data collected only considers offensive contributions. We would have to make adjustments and estimates for how much value to place on his defensive contributions and there is no general consensus on what those adjustments should be. Gretzky and Lemieux are more easily comparable since not only did they play the same position they also achieved their peaks within a fairly close proximity of each other(hence why the title to says "Highest Offensive Peak" instead of just highest peak). Orr may or may not have had the highest "Peak" but as incredible as he was, he absolutely did not have a higher offensive peak than Lemieux or Gretzky, but I digress.
Now I'm not saying that NHL should suddenly change the way they determine who is the Art Ross or Rocket winner, but clearly these awards and the raw totals that are associated with them both not tell the full story, for that we require more advanced stats. Thankfully the hockey community has begun to evolve and even a good portion of the fan base now understands that raw totals alone do not necessarily convey the value that they appear to at first glance. That they are not necessarily the best parameter to use when measuring one players performance against another and the teams themselves have long since move beyond this metric, commonly employing much more sophisticated analytics to make conclusions about player performance and comparisons between them. Even baseball has moved on and places more value in analytics and rate stats than raw totals. While they still likewise make use of many counting stats for the most part averages and not raw totals are used to inform us who the most productive players in the league are in a given season. A player having 100 RBI's and 30 home runs in a season used to be one of the main metrics by which a player was seen as having a great year. We now understand that a player can reach those totals and actually be provide negative to their team. Heck, even a 40 home run season can have negative value. It would have been blasphemous to even suggest such a thing only a couple decades ago. Even batting average, a rate statistic, is now understood to be a flawed assessment of a players performance and other more advanced metrics like OBP(on base percentage) are seen as better tools. I'll end my point with this - While I don't disagree that the measure of a player's value can be determined by their raw totals, the true measure of a players performance should not be measured solely by them. Per-game production is far more relevant in that regard. Of course it does come with an important caveat - in order to be considered valid this per-game production must be sustained over a long enough period of time. I will touch upon that matter in the next post below but I think it's time to provide some actual data.
forums.hfboards.com
The majority seems to feel that Lemieux was not at his peak during this season an assessment I agree with. Though by some measures that season actually compares favorably to both his and Gretzky's best. Others view his 1987-88 season as a peak year as well but much like his 1989-90 & 1991-92 seasons, both of which are within his generally recognized peak period, Lemieux's offensive production in that season was clearly a step below his very best. It's clear that he had yet another step in him which he showed the following year when he posted superior numbers and on an adjusted basis 87-88 season is even less impressive. Again, this is not about longevity of peak, therefore there is no point in comparing seasons which are a step behind the others.
Gretzky meanwhile, well to be frank he made things difficult... The magnificent bastard had such a long and consistent peak that it's almost impossible to pin down exactly when he was no longer performing at his absolute best. Though it is much simpler if your just looking at the raw numbers. For you his best season is going to be one of these three; 81-82(most goals), 83-84(highest ppg) or 84-85(most points) and his 'peak period' will be that five year stretch. But in actuality his peak point production was incredibly consistent over a 6 and a half year stretch. His goal scoring did vary a bit more though. While the Art Ross may value goals and assist equally on a 1 to 1 basis and a good number of assists are more indicative of skill than the goals that resulted from them, generally speaking goals are acknowledged as having a higher value than assists in point totals. This has been frequently been discussed on the boards already and has clearly been established as the general consensus;
Are GOALS worth more than ASSISTS?
Which is why to me once Gretzky's goal scoring declined by a certain amount I could no longer consider him as being in his peak. Prior to this analysis I saw Gretzky as having 5 peak seasons; 1981 through to 1986. However others have argued that his peak extended beyond this with various time frames ranging from 1981-87, 1981-88, 1981-89, 1981-91 or even as long as 1981-93. That debate rages on and will probably continue ad infinitum without resolution. Personally I see his peak as beginning somewhere between a little after the halfway mark of the 1980-81 season and ending about one-third of the way into the 87-88 season, essentially 6 full seasons plus parts of two others which puts him just a little shy of 7 full seasons. But obviously since the first and last of those seasons are only partial seasons(80-81 & 87-88) they can not be counted as peak seasons in and of themselves. Hence I arrived to the peak period of 1981-87 for Gretzky. I'm sure some will argue that his 87-88 season should also be considered a peak season as on a per game point basis it is on par with 86-87. But as I noted above, his goal scoring rate over the course of that season had just fallen far too much to be comparable to either his previous best seasons or Lemieux's best. A season of 40-109-149 is still quite impressive but is not in the same realm as a season of 69-91-160. Hence that season and all of Gretzky's beyond it are not included in this evaluation. Now that we finally got that settled its time to show some cold hard data.
There are three parts to my analysis;
1)Scoring levels
2) Supporting cast and
3) Stat padding/play time
Playing conditions have changed radically over the course of NHL history. In 2018-19 the average team scored 247 goals while 20 years earlier only 2 out of 27 teams exceeded that total; Toronto with 268 and NJ by a single goal with 248. The most commonly use statistics in hockey; raw totals, make no attempt to adjust for these kinds of situational factors. Kucherov won the scoring race in that 2018-19 season with 128 points while 20 years prior it was Jagr who led the way with 127. So who had the better season? A casual hockey fan with no understanding of the history of the game would probably say it was Kucherov and how could you blame him? On face value it certainly appears that way when you take a quick look at any website that lists the Art Ross winners and their point totals. Also when Kucherov tallied his 128th point media pundits were quick to point out it was the "best" single season total since Lemieux in 95-96. Yes it is absolutely correct to say it was the highest total since Lemieux '96. It's the usage of certain terms that is problematic, rather than saying it was the "highest" total since... they instead commonly labeled it as being the "best" total since... In a sense discarding Jagr's 98-99 season as being less noteworthy. But ANY knowledgeable hockey fan knows that is absolutely not the case. Jagr had a few more goals and a smidgeon higher ppg; 1.57 vs 1.56 but we all know that's not why Jagr's season is understood as being the greater of the two.
Here's another coincidence; the goals per game difference between that 2018-19 and 1998-99 season was 0.38 (3.01 vs 2.63) This is exactly the same as the goals per game difference between two of the seasons that are most commonly recognized as Gretzky and Lemieux's greatest; 1981-82 and 1992-93; 0.38 (4.01 vs 3.63) But just as how we recognize there is a difference between what Kucherov did and what Jagr did despite practically identical totals, we also realize that the difference scoring-wise between 1981-82 and 1992-93 is far less dramatic than the difference between 1998-99 and 2018-19 despite an identical raw difference in scoring. So what does this tell us? It tells us that using raw statistics alone is a poor metric of evaluation and that most inherently understand that.
As we all know there has never been an era in NHL history where scoring was higher than it was during Wayne's peak. To be sure Lemieux's peak also happened in higher scoring seasons, but it is well known that Lemieux produced at a higher PPG relative to league scoring rates:
Adjusted data like that has been known for some time and was the original reason why I rated Lemieux as having the higher peak. However I've since recognized that adjusting point totals based just on overall scoring averages are flawed since they do not take into consideration increased scoring by offensive players in seasons where more powerplay goals are scored. My analysis of the two hardly revolves around comparing simplistic adjusted point totals. I wouldn't waste everyone's time using such a flawed metric. I agree that further adjustments are needed to account for the differences in powerplay scoring levels throughout each season.
Those in the Gretzky for peak camp note that Lemieux benefited from increased powerplay opportunities and certainly there is no disputing that he did in fact score a greater percentage of his goals and points on the powerplay. But there's two things people who say that are failing to recognize. One - it's not as if this was done by choice, Lemieux didn't purposefully choose to became some sort of super power-play leach. By the time he had reached his peak ALL superstar players were accumulating a progressively share of their points on the powerplay. Two - It was more difficult to score at even strength during his peak than it was during Gretzky's in the early to mid 80's. Even strength scoring rates had dropped by double the amount compared to overall scoring rates:
Not only that but it was also easier to score on the powerplay during Gretzky's peak, the drop in powerplay percentage from the early 80's attests to that - scoring per 60 declined in all situations. The one thing that countered these trends was the increase in powerplay opportunities. But that increase was not enough to change the over trend of declining goal totals.
Chart of seasonal scoring averages per situation;
Gretzky's advantage in even strength scoring is often brought up in comparisons between him and Lemieux to support the argument that he was the better of the two at their best since he scored more of these 'harder to get' points. But in an environment where it's between 18 to 42% easier to score even strength points exactly who do you think is going to get more points if all else was even between the two?
Here's what happens to their Even Strength totals when you normalize even strength scoring alone to the same level(adjusted to 81-82's 2.98 ES goals per game):
When adjusted to the same even strength scoring levels Lemieux 1992-93 season is the best, narrowly beating out three of Gretzky's seasons. Lemieux's 92-93 also has the highest even strength goals per game rate besting Gretzky's record 81-82 seasons where he had 68 goals. Lemieux was on pace for 66 ES goals in a season where ES scoring rates were nearly 20% lower. That doesn't even consider Gretzky's 4 on 4 scoring which would lower his numbers even further. While that's technically even strength it's also a situation where scoring levels are much higher than 5 on 5 which is what most people think about when you say "even strength". The 81-82 season where Gretzky set the record for ES goals and points isn't actually his best season ES season when you make adjustments for the fact that ES scoring was at all-time record levels that year(post-war). He had 3 seasons which were actually better than it to it.
Before anyone get's all up in arms here - Gretzky's 6 year stretch of consistent near record performance is still unmatchable, Lemieux did not touch that and THAT is why Gretzky is the Goat. But on a per-game basis the numbers show that Lemieux's 1992-93 season is in fact THE BEST ES scoring pace anyone in the history of the game has achieved when you actually take into consideration league-wide scoring rates. Now the adjusted totals will change based on which season you choose to adjust them to, however the difference between them always stay the same. For example when adjusted to 92-93's 20% lower scoring rate of 2.439 EV goals per game Gretzky's 81-82 season comes out to 56G 65A 120PT in 80 = 0.70GPG 1.50PPG vs Lemieux's actual totals of 47G 49A in 60 = 0.78GPG 1.60PPG.
Moving on to Power Play totals, once again adjusted to the highest scoring year for powerplays 1988-89 season @1.058 PP gpg;
Once again Lemieux has the best powerplay scoring season though this time it's 95-96 . His powerplay performance that season was otherworldly, nobody not even himself comes to within 80% of his powerplay production in 95-96. Lemieux's 1992-93 powerplay performance is well back of his 95-96 season, but it's still ahead of all of Gretzky's peak seasons. Interestingly enough Conner McDavid just had a season that rivaled Lemieux's 95-96 season when it came to powerplay production and likewise he did not lead the league in EV points. I'm curious to see if those who derided Lemieux for "getting so many powerplay points" will do the same with McDavid as well...
The one situation where Gretzky does come out on top is in Short Handed scoring, the highest scoring average for which also occurs in 88-89 @0.155 SH gpg:
I'll say this - What the numbers show is was that while his advantage declines considerably Gretzky should still be considered best even strength scorer of all time seeing as how he still has 6 of the 7 highest adjust ES totals of all time. However, Lemieux DID in fact match slightly best him - or in the least you can say he matched him, at ES scoring in their peaks when you add context to the comparison. On the powerplay Lemieux advantage likewise declines but he was still a superior powerplay scorer, Gretzky never came close to matching Lemieux's production even when his figures are adjust for powerplay scoring levels. Now some might say it's not that powerplay scoring was higher in Lemieux's best years but that his team also got way more powerplays than usual. But that only applies to his 80's seasons, the Penguins were at or around the league average in opportunities in the 90's. And Gretzky overall totals would not have ballooned massively if he played in those high powerplay scoring seasons because his gains would have been countered by the losses in ES scoring and 4v4 scoring, where his team was already scoring at levels approaching a powerplay.
In any case these charts only show their adjusted situational totals in isolation. When we adjust for varying scoring levels on ES, PP & SH together during each season we get varying figures. For example if you adjust Gretzky's seasons to 1988-89 it will favor Lemieux's numbers and conversely if you adjusted Lemieux's seasons to 1981-82 it will favor Gretzky's numbers. In order to have a proper comparison we essentially have to adjust all their peak seasons to every other season in consideration.
The calculation of these totals is rather straightforward. For example when I adjust Gretzky's 81-82 numbers to 82-83 scoring levels I take his ES, PP & SH totals, divide them by the average per game totals for 81-82 and then multiple them by the per game numbers for 82-83
ES totals; 68-79-147 times 2.980 then divided by 2.862 = 65-76-141
PP totals; 18-39-57 times 0.917 then divided by 0.889 = 17-38-55
SH totals; 6-2-8 times 0.115 then divided by 0.114 = 6-2-8
Added all the new totals together and you get = 89-116-204
Sometimes this leads to point totals that are off by one as above(89+116 should equal 205) but that's just due to rounding. To one decimal place his converted numbers add up as such; 88.7-115.7-204.4
The seasonal scoring averages for ES, PP & SH for each year is listed at the top(Hockey reference list of overall & powerplay scoring levels). The numbers for the season being converted are in BLACK, the top season is highlight in GREEN and those in YELLOW are within 90% of the top season. calculations were all done on an excel spreadsheet
In ALL instance Lemieux's 1992-93 season has the highest goal and point per game scoring levels when you account for differing Even Strength, Power Play and Short Handed scoring levels.
Also the claim that Lemieux 88-89 season stats and 95-96 were inflated by an excusive amount of powerplay opportunities doesn't hold true. When adjusted to ES/PP/SH scoring levels in any year of the 80's they both always rank in the top 5 in goals and points. That's because the loss of points from a reduction in the powerplay opportunities is countered by the increase in even strength scoring levels per min AND a reduction in SH playing time, a much lower scoring situation that Lemieux played much time in, some seem to forget this.
So what's the baseline season? I believe the fairest method to use would be an average of seasons rather than a single season. Since the end of the original six era in 1967-68 to last season 2021-22 these are the league scoring averages;
When adjusted to the scoring averages over the previous 55 seasons here's what the adjusted totals come out to:
Once again Lemieux's 1992-93 season prevails on a per game basis.
*content recently consolidated for easier reading
Peak, longevity & consistency, are factors that all matter when it comes to the Goat debate. Lemieux had an extremely high peak but due to injuries he did not consistently play at that level year in year out. Gretzky meanwhile maintained his peak at least 5 years straight with very little deviation. Gretzky is also clearly ahead of Lemieux in overall longevity and every one else in that department, save for a player like Howe. Mario would need to have more than just a little edge in peak performance to overcome the differences in longevity & consistency. But anyway you cut it the difference in peak between them is too small to change any of the other factors and so I agree that there is no argument for Lemieux as the hockey Goat.
In terms of peak performance there are those who also feel that matter has long since been settled as well. Certainly if you just take the raw numbers at face value this would be correct. Not much point in reading beyond this point as you already have your answer then. Some say it's not just the records but the fact he reached his peak level repeatedly in many years. Well, that's precisely what makes Gretzky the GOAT. No one else in the history of the game had his combined level of consistency and peak performance. But that alone does not mean that Gretzky's absolutely peak was the highest of all time, that is confusing longevity of peak, for peak performance. I would agree I'd be more likely to view Lemieux's peak performance as suspect if he only did it once. But he did it twice, that to me proves it was no "fluke season" the first time.
If you ask very casual hockey fans they're likely to tell you that the raw totals are what matters most and they do so because the NHL continues to reinforce this belief. Hockey is the only major North American sports that focuses almost exclusively on raw counting stats when it comes to awarding the best offensive performers in a season. The Art Ross and the Rocket are simply awarded to the players who score the "most points" and "most goals" respectively in the league during a particular season. These are awards which serve to dramatically increase the value of a players legacy and yet winning those awards does not necessarily mean the player in question was actually the best point producer or goal scorer in the league during that season. The problem lies in the concepts of performance and value; people commonly confuse the two. Performance is exclusive of value and value is simply what happens when a level of performance is sustained over a certain period of time. But people like it when things are kept simple and so the NHL continues to keep it simple for them.
Let's consider Alexander Mogilny's and Teemu Selänne's 1992-93 season, that year they both scored 76 goals. I consider Mogilny as having the superior goal scoring season of the two. This is because he scored his 76 goals in 77 games while Selänne did so over the course of 84 games. Selanne had 73 goals in his best 77 game stretch that season, impressive no doubt but Mogilny bested that by scoring 73 goals in a 68 game stretch. On a per-game basis Mogilny's scoring rate was notably higher; 0.987 vs 0.905 but the NHL instead essentially sees no difference between the two and recognizes both as the winners of the goal scoring title without any attention given to their games played. I believe this to be a faulty and incorrect assessment. Furthermore, neither Mogilny nor Selänne were even the best goal scorer in the league that season. That honor belonged to Mario Lemieux who scored 69 goals in 60 games(and really even less than that), easily besting both of them by scoring at blistering pace of 1.150 goals per game.
So how can the true measure of a players performance be determined if not by raw totals? The 'eye test' is one way but unfortunately it can easily be influenced by personal biases and it naturally fades with time. While it can be useful in supporting an argument it's far from a definitive fact. Statistics are definitive, however while we can make use of statistics as an absolute we must first understand which statistics hold the most value. The most obvious distinction between statistics are those that measure a quantity of production and those that measure a rate of production. A counting stat is one that measures a player's total production without reference to his number of opportunities. Common counting stats include goals, points and shots. While the most common counting stats are raw numbers some sites provide sabermetric stats that are also expressed the same way, such as hockey reference's point share. In contrast to counting stats, rate stats try to show a rate of productivity by dividing the number of successes or failures by the number of opportunities. Most rate statistics will include a term like "average" or "percentage" in their name as in indication that they are the result of a division process. Some examples of traditional rate stats including shooting percentage and goals against average. It's important to note that many counting stats are given with an implicit number of opportunities, like a season or career. This kind of implicit number of opportunities may or may not balance out. While two 1st line centers will be seen as playing in very similar conditions to facilitate the collection of goals and assists, the same is not true of a center and a defensemen so a comparison of their counting stats is not valid which brings me to Bobby Orr... There are a fair number of people who believe that it is neither Gretzky nor Lemieux but rather Orr who had the highest peak in NHL history. Unfortunately it's difficult to quantify his value in comparison to the other two using the statistics we have available to us because most of the data collected only considers offensive contributions. We would have to make adjustments and estimates for how much value to place on his defensive contributions and there is no general consensus on what those adjustments should be. Gretzky and Lemieux are more easily comparable since not only did they play the same position they also achieved their peaks within a fairly close proximity of each other(hence why the title to says "Highest Offensive Peak" instead of just highest peak). Orr may or may not have had the highest "Peak" but as incredible as he was, he absolutely did not have a higher offensive peak than Lemieux or Gretzky, but I digress.
Now I'm not saying that NHL should suddenly change the way they determine who is the Art Ross or Rocket winner, but clearly these awards and the raw totals that are associated with them both not tell the full story, for that we require more advanced stats. Thankfully the hockey community has begun to evolve and even a good portion of the fan base now understands that raw totals alone do not necessarily convey the value that they appear to at first glance. That they are not necessarily the best parameter to use when measuring one players performance against another and the teams themselves have long since move beyond this metric, commonly employing much more sophisticated analytics to make conclusions about player performance and comparisons between them. Even baseball has moved on and places more value in analytics and rate stats than raw totals. While they still likewise make use of many counting stats for the most part averages and not raw totals are used to inform us who the most productive players in the league are in a given season. A player having 100 RBI's and 30 home runs in a season used to be one of the main metrics by which a player was seen as having a great year. We now understand that a player can reach those totals and actually be provide negative to their team. Heck, even a 40 home run season can have negative value. It would have been blasphemous to even suggest such a thing only a couple decades ago. Even batting average, a rate statistic, is now understood to be a flawed assessment of a players performance and other more advanced metrics like OBP(on base percentage) are seen as better tools. I'll end my point with this - While I don't disagree that the measure of a player's value can be determined by their raw totals, the true measure of a players performance should not be measured solely by them. Per-game production is far more relevant in that regard. Of course it does come with an important caveat - in order to be considered valid this per-game production must be sustained over a long enough period of time. I will touch upon that matter in the next post below but I think it's time to provide some actual data.
Peak Seasons
The first thing we must do before making a comparison between the two is to identify their peak seasons. For Lemieux it's quite easy, his peak 'period' is generally seen as stretching from either 1987-93 or 1988-93 and within those stretches he very clearly has two defining seasons; 1988-89 and 1992-93. Some feel that Lemieux's 1995-96 can also be considered a peak season and I made a thread discussing that exact matter just to see what the general consensuses on that is;
Was Lemieux still in his peak in 1995-96?
Beliveau-Geoffrion were yet to be factor during Howe 53 season, Richard was over 30 by that point. So for Howe peak competition during Howe peak season he could have a point, if we include Ricahrd to Beliveau prime, Howe 95 points and 49 goals look great still obviously...
The majority seems to feel that Lemieux was not at his peak during this season an assessment I agree with. Though by some measures that season actually compares favorably to both his and Gretzky's best. Others view his 1987-88 season as a peak year as well but much like his 1989-90 & 1991-92 seasons, both of which are within his generally recognized peak period, Lemieux's offensive production in that season was clearly a step below his very best. It's clear that he had yet another step in him which he showed the following year when he posted superior numbers and on an adjusted basis 87-88 season is even less impressive. Again, this is not about longevity of peak, therefore there is no point in comparing seasons which are a step behind the others.
Gretzky meanwhile, well to be frank he made things difficult... The magnificent bastard had such a long and consistent peak that it's almost impossible to pin down exactly when he was no longer performing at his absolute best. Though it is much simpler if your just looking at the raw numbers. For you his best season is going to be one of these three; 81-82(most goals), 83-84(highest ppg) or 84-85(most points) and his 'peak period' will be that five year stretch. But in actuality his peak point production was incredibly consistent over a 6 and a half year stretch. His goal scoring did vary a bit more though. While the Art Ross may value goals and assist equally on a 1 to 1 basis and a good number of assists are more indicative of skill than the goals that resulted from them, generally speaking goals are acknowledged as having a higher value than assists in point totals. This has been frequently been discussed on the boards already and has clearly been established as the general consensus;
Are GOALS worth more than ASSISTS?
Which is why to me once Gretzky's goal scoring declined by a certain amount I could no longer consider him as being in his peak. Prior to this analysis I saw Gretzky as having 5 peak seasons; 1981 through to 1986. However others have argued that his peak extended beyond this with various time frames ranging from 1981-87, 1981-88, 1981-89, 1981-91 or even as long as 1981-93. That debate rages on and will probably continue ad infinitum without resolution. Personally I see his peak as beginning somewhere between a little after the halfway mark of the 1980-81 season and ending about one-third of the way into the 87-88 season, essentially 6 full seasons plus parts of two others which puts him just a little shy of 7 full seasons. But obviously since the first and last of those seasons are only partial seasons(80-81 & 87-88) they can not be counted as peak seasons in and of themselves. Hence I arrived to the peak period of 1981-87 for Gretzky. I'm sure some will argue that his 87-88 season should also be considered a peak season as on a per game point basis it is on par with 86-87. But as I noted above, his goal scoring rate over the course of that season had just fallen far too much to be comparable to either his previous best seasons or Lemieux's best. A season of 40-109-149 is still quite impressive but is not in the same realm as a season of 69-91-160. Hence that season and all of Gretzky's beyond it are not included in this evaluation. Now that we finally got that settled its time to show some cold hard data.
There are three parts to my analysis;
1)Scoring levels
2) Supporting cast and
3) Stat padding/play time
Part 1 - Scoring Levels
Playing conditions have changed radically over the course of NHL history. In 2018-19 the average team scored 247 goals while 20 years earlier only 2 out of 27 teams exceeded that total; Toronto with 268 and NJ by a single goal with 248. The most commonly use statistics in hockey; raw totals, make no attempt to adjust for these kinds of situational factors. Kucherov won the scoring race in that 2018-19 season with 128 points while 20 years prior it was Jagr who led the way with 127. So who had the better season? A casual hockey fan with no understanding of the history of the game would probably say it was Kucherov and how could you blame him? On face value it certainly appears that way when you take a quick look at any website that lists the Art Ross winners and their point totals. Also when Kucherov tallied his 128th point media pundits were quick to point out it was the "best" single season total since Lemieux in 95-96. Yes it is absolutely correct to say it was the highest total since Lemieux '96. It's the usage of certain terms that is problematic, rather than saying it was the "highest" total since... they instead commonly labeled it as being the "best" total since... In a sense discarding Jagr's 98-99 season as being less noteworthy. But ANY knowledgeable hockey fan knows that is absolutely not the case. Jagr had a few more goals and a smidgeon higher ppg; 1.57 vs 1.56 but we all know that's not why Jagr's season is understood as being the greater of the two.
Here's another coincidence; the goals per game difference between that 2018-19 and 1998-99 season was 0.38 (3.01 vs 2.63) This is exactly the same as the goals per game difference between two of the seasons that are most commonly recognized as Gretzky and Lemieux's greatest; 1981-82 and 1992-93; 0.38 (4.01 vs 3.63) But just as how we recognize there is a difference between what Kucherov did and what Jagr did despite practically identical totals, we also realize that the difference scoring-wise between 1981-82 and 1992-93 is far less dramatic than the difference between 1998-99 and 2018-19 despite an identical raw difference in scoring. So what does this tell us? It tells us that using raw statistics alone is a poor metric of evaluation and that most inherently understand that.
As we all know there has never been an era in NHL history where scoring was higher than it was during Wayne's peak. To be sure Lemieux's peak also happened in higher scoring seasons, but it is well known that Lemieux produced at a higher PPG relative to league scoring rates:
Adjusted data like that has been known for some time and was the original reason why I rated Lemieux as having the higher peak. However I've since recognized that adjusting point totals based just on overall scoring averages are flawed since they do not take into consideration increased scoring by offensive players in seasons where more powerplay goals are scored. My analysis of the two hardly revolves around comparing simplistic adjusted point totals. I wouldn't waste everyone's time using such a flawed metric. I agree that further adjustments are needed to account for the differences in powerplay scoring levels throughout each season.
Those in the Gretzky for peak camp note that Lemieux benefited from increased powerplay opportunities and certainly there is no disputing that he did in fact score a greater percentage of his goals and points on the powerplay. But there's two things people who say that are failing to recognize. One - it's not as if this was done by choice, Lemieux didn't purposefully choose to became some sort of super power-play leach. By the time he had reached his peak ALL superstar players were accumulating a progressively share of their points on the powerplay. Two - It was more difficult to score at even strength during his peak than it was during Gretzky's in the early to mid 80's. Even strength scoring rates had dropped by double the amount compared to overall scoring rates:
Overall scoring in 81-82 was 7.2% higher than it was 88-89, 10.7% higher than 92-93 and 27.6% higher than 95-96.
But even-strength scoring in 81-82 was 17.8% higher than it was 88-89, 22.2% higher than 92-93 and 42.1% higher than 95-96.
Not only that but it was also easier to score on the powerplay during Gretzky's peak, the drop in powerplay percentage from the early 80's attests to that - scoring per 60 declined in all situations. The one thing that countered these trends was the increase in powerplay opportunities. But that increase was not enough to change the over trend of declining goal totals.
Chart of seasonal scoring averages per situation;
Season | Overall | EV | PP | SH | PPO | PP% | SHG % | SV% | % off peak | % off EV peak |
4.01 | 2.98 | 0.92 | 0.12 | 4.00 | 22.89 | 2.88% | 0.873 | 100.0 | 100.0 | |
3.86 | 2.86 | 0.89 | 0.11 | 3.87 | 22.94 | 2.95% | 0.875 | 96.4 | 96.0 | |
3.94 | 2.88 | 0.92 | 0.14 | 4.20 | 21.98 | 3.40% | 0.873 | 98.4 | 96.5 | |
3.89 | 2.86 | 0.89 | 0.14 | 4.01 | 22.20 | 3.44% | 0.875 | 96.9 | 95.9 | |
3.97 | 2.81 | 1.02 | 0.14 | 4.62 | 22.10 | 3.03% | 0.874 | 99.0 | 94.2 | |
3.67 | 2.64 | 0.90 | 0.13 | 5.04 | 20.99 | 2.60% | 0.879 | 91.5 | 88.5 | |
3.74 | 2.53 | 1.06 | 0.15 | 5.04 | 20.99 | 3.07% | 0.879 | 93.3 | 84.9 | |
3.63 | 2.44 | 1.03 | 0.15 | 5.28 | 19.57 | 2.93% | 0.885 | 90.4 | 81.8 | |
3.14 | 2.10 | 0.90 | 0.14 | 5.04 | 17.93 | 2.82% | 0.898 | 78.3 | 70.5 |
Gretzky's advantage in even strength scoring is often brought up in comparisons between him and Lemieux to support the argument that he was the better of the two at their best since he scored more of these 'harder to get' points. But in an environment where it's between 18 to 42% easier to score even strength points exactly who do you think is going to get more points if all else was even between the two?
Here's what happens to their Even Strength totals when you normalize even strength scoring alone to the same level(adjusted to 81-82's 2.98 ES goals per game):
| Season | Gm | ES PT | League GPG | | Adj PT | GPG | PPG | ES Pt's in 80 Gm | G% off top season | P% off top season |
Gretzky | 81-82 | 80 | 147 | 2.98 | | 147 | 0.85 | 1.84 | 147 | 88.8% | 94.0% |
Gretzky | 82-83 | 80 | 132 | 2.86 | | 137 | 0.61 | 1.72 | 137 | 63.9% | 87.9% |
Gretzky | 83-84 | 74 | 135 | 2.88 | | 140 | 0.77 | 1.89 | 151 | 80.4% | 96.7% |
Gretzky | 84-85 | 80 | 146 | 2.86 | | 152 | 0.70 | 1.90 | 152 | 73.5% | 97.3% |
Gretzky | 85-86 | 80 | 143 | 2.81 | | 152 | 0.50 | 1.90 | 152 | 52.7% | 97.1% |
Gretzky | 86-87 | 79 | 124 | 2.64 | | 140 | 0.60 | 1.77 | 142 | 62.8% | 90.8% |
| | | | | | | | | | | |
Lemieux | 188-8 | 76 | 102 | 2.53 | | 120 | 0.64 | 1.58 | 126.5 | 66.4% | 80.9% |
Lemieux | 192-9 | 60 | 96 | 2.44 | | 117 | 0.96 | 1.955 | 156 | 100.0% | 100.0% |
Lemieux | 95-96 | 70 | 73 | 2.10 | | 104 | 0.61 | 1.48 | 119 | 63.6% | 75.8% |
When adjusted to the same even strength scoring levels Lemieux 1992-93 season is the best, narrowly beating out three of Gretzky's seasons. Lemieux's 92-93 also has the highest even strength goals per game rate besting Gretzky's record 81-82 seasons where he had 68 goals. Lemieux was on pace for 66 ES goals in a season where ES scoring rates were nearly 20% lower. That doesn't even consider Gretzky's 4 on 4 scoring which would lower his numbers even further. While that's technically even strength it's also a situation where scoring levels are much higher than 5 on 5 which is what most people think about when you say "even strength". The 81-82 season where Gretzky set the record for ES goals and points isn't actually his best season ES season when you make adjustments for the fact that ES scoring was at all-time record levels that year(post-war). He had 3 seasons which were actually better than it to it.
Before anyone get's all up in arms here - Gretzky's 6 year stretch of consistent near record performance is still unmatchable, Lemieux did not touch that and THAT is why Gretzky is the Goat. But on a per-game basis the numbers show that Lemieux's 1992-93 season is in fact THE BEST ES scoring pace anyone in the history of the game has achieved when you actually take into consideration league-wide scoring rates. Now the adjusted totals will change based on which season you choose to adjust them to, however the difference between them always stay the same. For example when adjusted to 92-93's 20% lower scoring rate of 2.439 EV goals per game Gretzky's 81-82 season comes out to 56G 65A 120PT in 80 = 0.70GPG 1.50PPG vs Lemieux's actual totals of 47G 49A in 60 = 0.78GPG 1.60PPG.
Moving on to Power Play totals, once again adjusted to the highest scoring year for powerplays 1988-89 season @1.058 PP gpg;
Adjusted to 1.058 PP GPG | Season | Gm | PP PT | League GPG | | Adj PP PT | GPG | PPG | PP Pt's 80 Gm Proj | G% vs top season | P% vs top season |
Gretzky | 81-82 | 80 | 57 | 0.917 | | 66 | 0.26 | 0.82 | 65.8 | 50.1% | 62.2% |
Gretzky | 82-83 | 80 | 54 | 0.889 | | 64 | 0.27 | 0.80 | 64.3 | 51.7% | 60.8% |
Gretzky | 83-84 | 74 | 47 | 0.924 | | 54 | 0.31 | 0.73 | 58.2 | 59.7% | 55.1% |
Gretzky | 84-85 | 80 | 44 | 0.891 | | 52 | 0.12 | 0.65 | 52.2 | 22.9% | 49.4% |
Gretzky | 85-86 | 80 | 54 | 1.021 | | 56 | 0.14 | 0.70 | 56.0 | 27.5% | 52.9% |
Gretzky | 86-87 | 79 | 46 | 0.903 | | 54 | 0.19 | 0.68 | 54.6 | 37.2% | 51.6% |
| | | | | | | | | | | |
Lemieux | 88-89 | 76 | 79 | 1.058 | | 79 | 0.41 | 1.04 | 83.2 | 78.7% | 78.7% |
Lemieux | 92-93 | 60 | 55 | 1.032 | | 40 | 0.27 | 0.94 | 75.2 | 52.7% | 71.1% |
Lemieux | 95-96 | 70 | 79 | 0.904 | | 92 | 0.52 | 1.32 | 105.7 | 100.0% | 100.0% |
Once again Lemieux has the best powerplay scoring season though this time it's 95-96 . His powerplay performance that season was otherworldly, nobody not even himself comes to within 80% of his powerplay production in 95-96. Lemieux's 1992-93 powerplay performance is well back of his 95-96 season, but it's still ahead of all of Gretzky's peak seasons. Interestingly enough Conner McDavid just had a season that rivaled Lemieux's 95-96 season when it came to powerplay production and likewise he did not lead the league in EV points. I'm curious to see if those who derided Lemieux for "getting so many powerplay points" will do the same with McDavid as well...
The one situation where Gretzky does come out on top is in Short Handed scoring, the highest scoring average for which also occurs in 88-89 @0.155 SH gpg:
Season | Gm | SH PT | League GPG | Adj SH PT | GPG | PPG | G% vs top season | P% vs top season | |
Gretzky | 81-82 | 80 | 8 | 0.115 | 11 | 0.10 | 0.10 | 58% | 30% |
Gretzky | 82-83 | 80 | 10 | 0.114 | 14 | 0.10 | 0.13 | 58% | 37% |
Gretzky | 83-84 | 74 | 23 | 0.143 | 25 | 0.176 | 0.337 | 100.0% | 100.0% |
| Gretzky | 84-85 | 80 | 18 | 0.138 | 20 | 0.15 | 0.23 | 88% | 67% |
| Gretzky | 85-86 | 80 | 18 | 0.140 | 20 | 0.04 | 0.23 | 24% | 67% |
Gretzky | 86-87 | 79 | 13 | 0.131 | 15 | 0.10 | 0.16 | 60% | 49% |
Lemieux | 88-89 | 76 | 18 | 0.155 | 18 | 0.17 | 0.24 | 97.3% | 70% |
Lemieux | 92-93 | 60 | 9 | 0.155 | 9 | 0.10 | 0.15 | 57% | 45% |
Lemieux | 95-96 | 70 | 9 | 0.142 | 10 | 0.12 | 0.14 | 71% | 42% |
I'll say this - What the numbers show is was that while his advantage declines considerably Gretzky should still be considered best even strength scorer of all time seeing as how he still has 6 of the 7 highest adjust ES totals of all time. However, Lemieux DID in fact match slightly best him - or in the least you can say he matched him, at ES scoring in their peaks when you add context to the comparison. On the powerplay Lemieux advantage likewise declines but he was still a superior powerplay scorer, Gretzky never came close to matching Lemieux's production even when his figures are adjust for powerplay scoring levels. Now some might say it's not that powerplay scoring was higher in Lemieux's best years but that his team also got way more powerplays than usual. But that only applies to his 80's seasons, the Penguins were at or around the league average in opportunities in the 90's. And Gretzky overall totals would not have ballooned massively if he played in those high powerplay scoring seasons because his gains would have been countered by the losses in ES scoring and 4v4 scoring, where his team was already scoring at levels approaching a powerplay.
In any case these charts only show their adjusted situational totals in isolation. When we adjust for varying scoring levels on ES, PP & SH together during each season we get varying figures. For example if you adjust Gretzky's seasons to 1988-89 it will favor Lemieux's numbers and conversely if you adjusted Lemieux's seasons to 1981-82 it will favor Gretzky's numbers. In order to have a proper comparison we essentially have to adjust all their peak seasons to every other season in consideration.
The calculation of these totals is rather straightforward. For example when I adjust Gretzky's 81-82 numbers to 82-83 scoring levels I take his ES, PP & SH totals, divide them by the average per game totals for 81-82 and then multiple them by the per game numbers for 82-83
ES totals; 68-79-147 times 2.980 then divided by 2.862 = 65-76-141
PP totals; 18-39-57 times 0.917 then divided by 0.889 = 17-38-55
SH totals; 6-2-8 times 0.115 then divided by 0.114 = 6-2-8
Added all the new totals together and you get = 89-116-204
Sometimes this leads to point totals that are off by one as above(89+116 should equal 205) but that's just due to rounding. To one decimal place his converted numbers add up as such; 88.7-115.7-204.4
The seasonal scoring averages for ES, PP & SH for each year is listed at the top(Hockey reference list of overall & powerplay scoring levels). The numbers for the season being converted are in BLACK, the top season is highlight in GREEN and those in YELLOW are within 90% of the top season. calculations were all done on an excel spreadsheet
1981-82 = EV 2.980 | PP 0.917 | SH 0.115
1982-83 = EV 2.862 | PP 0.889 | SH 0.114
1983-84 = EV 2.877 | PP 0.924 | SH 0.143
1984-85 = EV 2.858 | PP 0.891 | SH 0.138
1985-86 = EV 2.807 | PP 1.021 | SH 0.140
1986-87 = EV 2.636 | PP 0.903 | SH 0.131
1988-89 = EV 2.529 | PP 1.058 | SH 0.155
1992-93 = EV 2.439 | PP 1.032 | SH 0.155
1995-96 = EV 2.097 | PP 0.904 | SH 0.142
In ALL instance Lemieux's 1992-93 season has the highest goal and point per game scoring levels when you account for differing Even Strength, Power Play and Short Handed scoring levels.
Also the claim that Lemieux 88-89 season stats and 95-96 were inflated by an excusive amount of powerplay opportunities doesn't hold true. When adjusted to ES/PP/SH scoring levels in any year of the 80's they both always rank in the top 5 in goals and points. That's because the loss of points from a reduction in the powerplay opportunities is countered by the increase in even strength scoring levels per min AND a reduction in SH playing time, a much lower scoring situation that Lemieux played much time in, some seem to forget this.
So what's the baseline season? I believe the fairest method to use would be an average of seasons rather than a single season. Since the end of the original six era in 1967-68 to last season 2021-22 these are the league scoring averages;
GP | G | EV | PP | SH | PPO | PP% | |
| Totals | 51848 | 317056 | 229330 | 77082 | 10644 | ||
| Per Game | 3.058 | 2.212 | 0.743 | 0.103 | 3.89 | 18.79 |
When adjusted to the scoring averages over the previous 55 seasons here's what the adjusted totals come out to:
Once again Lemieux's 1992-93 season prevails on a per game basis.
*content recently consolidated for easier reading
Last edited:
Just imagine what Lemieux could have done with Jari Kurri on his side while adding Messier & Anderson level-players onto the power play. With those upgrades I'm almost certain Lemieux easily sets the record for goals and points in '89. The problem of course is there is no means by which to accurately extrapolate how much of a difference this could have made in his numbers. It's simply something that can never be accurately quantified. But there's is no disputing that the lack of quality teammates hindered his production by some measure and considering this season already typically ranks as the 5th highest(sometimes 4th) when ES/PP/SH-adjusted there is certainly an argument to be had that it may have in fact been the best offensive season of all time.
