After integrating career and pace value into PPS as the key cogs in the formula, the results looked promising. It generated a befitting hierarchy of NHL players, their performances effectively contextualized. Unlike many “best of” lists, this one didn’t lean nostalgic or overrepresent skaters from high-scoring times. Instead, the list better balanced short and long careers, and extracted true performance relative to era. But stopping at career and pace as the main ingredients fell short of capturing the full picture.
The first sign that the existing formula remained incomplete was clear. Among goaltenders, Roberto Luongo rated higher than Dominik Hasek. While greatly admired off the ice, Luongo remains a criminally underrated player; in fact, he’s someone I hoped these ratings would give proper credit to in retirement. But, to be blunt, he was not Dominik Hasek. While Luongo played over 300 more games than Hasek, I couldn’t justify the order after digging into their yearly results. Hasek’s best was simply too much better than every other goalie, and the system did not give him sufficient credit. However, this was not the result that ultimately forced further changes.
Rather, it was that Bobby Orr did not have the highest rating among defencemen at this stage. Orr famously played only 8.6 AdYrs, earning the last scoring title by a defenceman in his final full season. His physical demise remains among the cruelest ‘what ifs’ in sports history. Yet, Orr’s career value remarkably ranks 12th among blueliners. By pace value, Orr is in another galaxy – over five(!) more point shares per year than any other defenceman. Robbed the ability to accumulate value long-term, Orr benefits from the statistical gift of never experiencing a decline in efficiency. Yet it was still not enough to articulate the quality of his career. In both the cases of Hasek and Orr, each reached a level of play that was not being adequately captured. Enter peak as the third major factor in PPS.
How will we measure peak value?
The terms ‘peak’ or ‘prime’ are often used, but not explicitly defined. It could be a player’s single best season, or best stretch of consecutive seasons, or handpicking any exceptional seasons at any point in a career. To integrate it into PPS, peak value needed to be labelled. In his Historical Baseball Abstract, Bill James used two elements of peak performance in his ratings – win shares in a player’s five best consecutive seasons, and win shares from a player’s three best seasons at any point. Jay Jaffe’s JAWS methodology uses a player’s seven best seasons at any time in their career.
Neither approach translated well to hockey; superstar players frequently missed too many games at their best to define peak by in-year output. Orr, for instance, missed at least nine games in all but five of his 12 seasons. Mario Lemieux only played consecutive seasons of 70+ games twice. Eric Lindros played 90% of the NHL schedule only twice – once as league MVP in a lockout-shortened year and not again until his age-29 season. Sidney Crosby’s best three-year run included only 99 games total. It became clear that some iteration of pace (i.e. a per-game basis) needed to be considered when defining a player’s peak value.
After much trial and error, a consecutive seven-year stretch of seasons appeared to best represent a player’s peak. Three or five seasons tends to overstate players with very short, outstanding careers. Given we are not penalizing for missed games within a player’s peak, using seven seasons forces a player to maintain their value for a solid stretch of time, regardless of injury or interruption. The logic for using consecutive seasons is that primes in hockey, with few exceptions, tends to be an identifiable band of years in a player’s 20s, perhaps extending into their early 30s. Think Terry Sawchuk’s earth-scorching debut, featuring a 221-114-78 record and five straight seasons with a GAA under 2.00. Or Guy Lafleur’s 836-point burst, inclusive of three scoring titles within his age-23 and age-29 seasons. Or Brett Hull’s explosive breakout of 411 goals in 510 games from 1989-90 through 1995-96. This is the slice of a career that we want to emphasize within PPS through peak value.
Thus, peak value is defined as the best pace value (i.e., AdPS per 82 games for skaters, or per 60 games for goalies) in a seven-season window. At a minimum, this stretch must include at least 3 AdYrs (i.e. 246 AdGP for skaters, 180 AdGP for skaters). We’ll again illustrate the calculation through Jeremy Roenick, as we seek his highest seven-season run of pace value over the course of his 20-season career.
JEREMY ROENICK, PPS PACE VALUE BY SEVEN-SEASON STRETCH
AdPS = Adjusted Point Shares; AdGP = Adjusted Games Played; Pace = AdPS ÷ AdGP x 82
Roenick’s aging curve is strikingly linear, a typical trend for an elite player. Comfortably exceeding the minimum 246 AdGP for every seven-year stretch of his career, all such septennials are eligible to be Roenick’s peak. Roenick tops out in seven-year pace value across his third through ninth seasons (1990-96 with Chicago through 1996-97 with Phoenix). His peak value, therefore, is 9.8 (i.e., 62.2 AdPS ÷ 521 AdGP x 82). Updating Roenick’s PPS score through three factors, it totals 25.6, as follows:
- Career Value = 8.4
- Pace Value = 7.4
- Peak Value = 9.8
PPS Factor #3: Peak ValueBest pace value over 7 consecutive seasons
If career is 7+ seasons: Minimum 3 adjusted seasons within the 7-season peak;
If career is <7 seasons: Minimum 3 adjusted seasons in career
Some guidelines are necessary to explain qualifying for peak value:
- A player need not get into 7 seasons to qualify – they just need to play at least 3 AdYrs in 7 seasons or less. If we did not set a minimum on AdYrs, active players could unfairly climb the PPS chart much too quickly. This threshold requires an up-and-coming player to amass the equivalent of three full seasons to earn peak value points. To get a sense of time served, each of Alex DeBrincat and Charlie McAvoy qualified for peak value during 2020-21. Similarly, a complete career that lasted less than seven seasons can qualify, so long as they play 3+ AdYrs (examples: pioneering star Newsy Lalonde, WWII veteran Johnny Mowers, and tragic plane crash victim Bill Barilko).
- A qualifying seven-season stretch must contain at least 3 AdYrs. We want to scrap “peaks” that are one or two decent partial seasons spread over seven scarcely played years. Example: Philipp Grubauer got into some NHL action each year from ages 21 through 27, playing just 139 AdGP. With seasons of 1, 2, and 17 AdGP baked into the seven-year window, this won’t qualify as an eligible peak for Grubauer.
- The lost 2004-05 lockout year, or any season where a player did not dress in a single game, does not count as a season. For example, Joe Thornton’s peak comprises his fourth through 10th NHL seasons. This will cover eight years but only seven seasons (i.e., 2000-01 to 2003-04, and 2005-06 to 2007-08).
- The same pro-ration for goaltenders used in pace value applies to peak value. Given peak is right sized to fit a single season, a goalie’s peak will also be measured across 60 games as we did with pace.
Peak value is therefore inherently higher than pace value, given it comprises the best cross-section of seasons by pace. A player at the peak of their powers is essential to the system, as an elite career is most commonly associated with the player’s best years, not their ascent or descent toward the pack. With the criteria now in place, peak value has its definition, and PPS has its three foundational factors.
While career value measures an accumulation of value, and pace value measures the efficiency of career value, peak value best captures the imagination. It represents the stretch of time where a player was at their absolute best in terms of contributing value to their team – the “peak of their powers.” This factor rightfully boosts Hasek and Orr upward in PPS, the missing piece of the formula to encapsulate the zenith of their dominance. Let’s check in on the top qualifying player by era and position in line with the definition and minimums established.
TOP PLAYER BY ERA AND POSITION, PPS PEAK VALUE
|1||Howie Morenz||151||Eddie Shore||132||Clint Benedict||147|
|2||Gordie Howe||129||Red Kelly||124||Bill Durnan||144|
|3||Wayne Gretzky||188||Bobby Orr||204||Ken Dryden||135|
|4||Jaromir Jagr||147||Al MacInnis||127||Dominik Hasek||148|
Stats through 2021-22; Peak Value = Best Pace Value over 7 consecutive seasons (for Skaters, expressed over 82 GP; for Goaltenders, expressed over 60 GP)
The players identified represent hockey played at its finest in each segment of the NHL’s history. The list depicts famously dominant performances – the highest per game values across any seven-season run by era. Perhaps only Era 4 stars Jagr and MacInnis warrant further discussion. Many would expect Crosby or Ovechkin atop the Era 4 forward list; however, Jagr’s peak remarkably featured a direct role in seven consecutive Art Ross Trophies (five by Jagr, two by Lemieux). Any dismissal of Jagr’s individual success as a product of Lemieux is unfounded; four of Jagr’s five scoring titles came with Lemieux out of the league entirely. Even if McDavid should pass Jagr’s peak (he is at 12.6 through six seasons), it will likely go under Era 5, the pair debuting 25 years apart.
MacInnis, meanwhile, is perhaps the most surprising inclusion on the list. Despite the eighth-best peak among all defencemen, he is often overshadowed by a pair of contemporaries, Coffey and Bourque. Given five of MacInnis’ seven peak seasons fall under Era 3, for those curious, Chris Pronger’s peak value (11.6) leads defencemen with seven-year peaks exclusively in Era 4.
Pivoting to the all-time best peaks regardless of era, we’ll list the top 10 per position. We have to keep in mind that timeline adjustments are not directly reflected in peak value, so the further back in history, the greater the ability to dominate. While career values lean toward long careers, and pace values toward short careers, the top peaks have no such predispositions. Thinking of the sport played at its best is pure exhilaration. Here is a celebration of the 10 most outstanding seven-season peaks in NHL history.
TOP 10 PLAYERS BY POSITION, PPS PEAK VALUE
|1||Wayne Gretzky||3||188||Bobby Orr||3||204||Dominik Hasek||4||148|
|2||Mario Lemieux||3||183||Denis Potvin||3||140||Clint Benedict||1||147|
|3||Howie Morenz||1||151||Paul Coffey||3||136||Bill Durnan||2||144|
|4||Phil Esposito||3||147||Ray Bourque||3||134||Ken Dryden||3||135|
|5||Jaromir Jagr||4||147||Eddie Shore||1||132||George Hainsworth||1||133|
|6||Guy Lafleur||3||138||Larry Robinson||3||127||Tony Esposito||3||132|
|7||Eric Lindros||4||138||Al MacInnis||4||127||Roberto Luongo||4||130|
|8||Nels Stewart||1||135||Brad Park||3||127||Terry Sawchuk||2||130|
|9||Sidney Crosby*||4||135||Red Kelly||2||124||Glenn Resch||3||129|
|10||Gordie Howe||2||129||Guy Lapointe||3||121||Johnny Mowers||1||128|
|*||Sidney Crosby*||4||135||Erik Karlsson*||4||106||Craig Anderson*||4||116|
Stats through 2021-22
There are only four names among the 30 outside the Hall of Fame, two of which remain ineligible (Crosby and Luongo are not eligible). This leaves the last two goaltenders on the list, a pair of surprises among the game’s most celebrated brick walls. Glenn ‘Chico’ Resch did not exceed 25 appearances until his age-27 season, arriving via the non-traditional path of NCAA at the time. Evenly splitting the crease with Billy Smith for six-plus seasons, Resch quietly dominated with annual save percentages near the top of the league every year. Dealt following the Islanders’ first of four straight Cups, Resch backstopped mediocre teams for the second half of his underrated career. Johnny Mowers, meanwhile, just qualifies for a peak value score. Cruelly losing his job during three absent years amidst WWII service, Mowers’ short, Vezina-winning career is largely lost to history.
The crew of ten forwards needs no introduction. Notably, Lemieux’s seven-year peak falls short of Gretzky’s (0.5 AdPS per season). The list of top 10 blueliners offers little shock value, aside from the absurdity of Orr’s 20.4. Astoundingly, the gap between Orr and next-best Denis Potvin (6.4 AdPS per season) is the same chasm separating Potvin and the 132nd best peak value by a defenceman (Jason Woolley). The defencemen list leans heavily toward Era 3 as opposed to Era 4, both demonstrating how actively elite blueliners participated in offence in these days, and further reinforcing the need to include timeline adjustments. Lastly, the goalie leaderboard is topped by The Dominator himself, inclusive of two or three goaltenders from each eras.
By closing our plunge into peak performance, PPS has its base. Career, pace and peak value will collectively anchor the scores. The fourth and fifth factors – timeline adjustments and bonus values – remain critical to the validity of the PPS methodology, their inclusion adding necessary refinement. A player can jump meaningfully up or down based on these final two factors despite the relatively low statistical weight we will assign.
PPS is rounding into fighting shape, a few steps away from bringing all the elements together to roll out a comprehensive NHL player evaluation system. We’ll complete the core scoring for Jarome Iginla below, adding his peak of 10.4, upping his core total to 27.8. Iginla’s peak is the 47th-best among forwards through 2020-21, covering his age-24, Richard and Ross-winning breakout through his age-31 seasons, inclusive of three Hart finalist nods.