On “Talent Dilution” – re-assessing one of history’s greatest foregone conclusions – PART 6 (Hartman’s Equation, and Common Sense)

PREVIOUS: Part 5 (Intuition, and what is “talent”?)

In the last section, I referred at some length to former Kansas State basketball coach Jack Hartman.  Or, because my basketball knowledge is admittedly limited, I used a couple of quotes from Bill James to refer to Jack Hartman.

Hartman’s Equation

If you skipped over the last section, which is a dumb thing to do but I don’t control your life, Jack Hartman referred to talent as “being where you’re supposed to be and doing what you’re supposed to do”.

With this in mind, I propose paring it down to a simple equation.

Being where you’re supposed to be + doing what you’re supposed to do = talent, or

On-field/court/diamond/ice intelligence + physical skill = talent, or

IQ + skill = talent

I’ll refer to this from here on as Hartman’s Equation, if for no other reason than it sounds good and it makes sense.

The Old-Timers

One does not need to look very far to find quotes from old-timers who swear that the game was never played at a higher or better level than when they were active.  It applies to all sports, it applies to all time periods, and the complaints are usually the same: opponents are too nice to each other, modern players are paid to well to truly care, kids have everything handed to them and aren’t well-rounded, blah blah blah.

I’ll refrain from quoting any one of them here, mostly because it’s pointless and because it’s entirely possible that someone I quote may have changed their opinion in the years since.  I don’t believe that opinions are static, and I don’t want to refer to a retired player as an outdated caricature of something that he may not even believe any more.

What I do remember is a quote from a player about 15 years ago, who was talking about how amazing it was to be playing in the NHL since it meant that he was one of the 700 best players in the world.  This is not an uncommon sentiment; players from the so-called Original Six era talk about how they had to be one of the 150 best players in the world, and goalies talk about how they had to be one of the six best, and so on.  It stands to reason that to be one of the 150 best requires a great deal of talent, and being one of the 700 best…still talented, but probably not as much.

Let’s look at any Hall of Fame and think of the best players to ever suit up.  Babe Ruth and Honus Wagner come to mind in baseball, Michael Jordan and Kareem Abdul-Jabbar in basketball, Wayne Gretzky and Gordie Howe and Bobby Orr, and so on.  If those aren’t the best players in their respective sports, the number of challengers to the top spot can be counted on one hand.

But now look further down the list.  Baseball has Ruth and Wagner and Musial and Mathewson and Schmidt and Mantle…but it also has Jesse Haines and Ray Schalk and Harold Baines and George Kelly and Chick Hafey.  Hockey has Gretzky and Orr and Howe and Lemieux and Plante and Harvey and Bourque…but it also has Shorty Green and Buddy O’Connor and several others along the same lines.  Even among Hall of Famers, there is a massive talent disparity – how much more so is the disparity in a league at any given time when the best players are Hall of Fame-caliber and the worst ones don’t last more than a season or two?

The Kids Have Their Say

Let us presume for a moment that the quotes from old-timers are in fact accurate, and that the best xxx number of players in the world are in the NHL.  If this were the case, then talent dilution exists and we should all be very concerned.  If the NHL had two teams, then it’s fairly likely that we’d be seeing the best 46 players in the world (based on the current active roster limit of 23 players). If it were four teams, then the best 92 players.  If six teams, the best 138.  And therefore, each time that a team is added, the 23 next-best players are added in on equal footing as the best in the world.

It seems plausible to suggest that the best player in the NHL is the best player in the world, but how far down does this go?

We can see the hole in the argument when looking at rookies over time.

In 1981-82, an 18-year-old rookie named Dale Hawerchuk stormed onto the scene with the Winnipeg Jets.  The previous year saw one of the worst teams in NHL history skate to a 9-57-14 record – one of only two teams in modern history to have fewer than ten wins in a full season (the 1974-75 Capitals, in their first year of existence, still hold the record with eight).  For their remarkable futility, Winnipeg received the first overall pick in the 1981 Entry Draft and selected Hawerchuk.

In his first season as a professional player, Hawerchuk scored 45 goals and added 58 assists, and Winnipeg surged to a 33-33-14 record and a playoff spot.

A decade later, another Jets rookie erupted onto the scene: a 22-year-old from Finland named Teemu Selanne.  In his first year in the NHL, Selanne scored 76 goals to tie for the league lead and was named first-team All-Star as the top right wing.  The top left wing that year was Luc Robitaille, who a few years earlier had also burst onto the scene as a rookie, scoring 45 goals to lead the Kings to a 16-point improvement over the previous season and into the playoffs.  The center on that All-Star line was Mario Lemieux, who as an 18-year-old some years prior had also emerged very quickly as one of the game’s best.  In Lemieux’s final NHL season, he suited up with a rookie named Sidney Crosby, who was the NHL’s sixth-leading scorer as an 18-year-old.  Crosby finished second in Calder voting that year behind a 20-year-old rookie named Alex Ovechkin, whose 52 goals were good for third in the NHL.

I can keep going on like this, tying the player just mentioned to a new one (Ovechkin to Nicklas Backstrom, who finished second in Calder voting behind Patrick Kane, who blah blah blah blah blah.)  But it’s laboring the point a bit to just keep doing that, although I could probably get another 25 players in before having enough.

A rookie coming in and doing anything at a high level, particularly becoming the top player on their team in their first season, by itself starts to disprove the idea of where the best players in the world are playing their hockey.  Mario Lemieux and Sidney Crosby and Dale Hawerchuk as 17-year-olds in the QMJHL, Teemu Selanne as a 21-year-old in the SM-Liiga, Alex Ovechkin as a 19-year-old in the Russian SuperLeague…these were very clearly among the best players in the world.  That they happened to not be in the NHL at that moment in time is absolutely, 100% irrelevant.

Taken through to its natural conclusion, the argument that the NHL possesses the best xxx number of players in the world can be pared down to this: the best player not in the NHL is a worse hockey player than the worst player currently in the NHL.  I don’t need charts or data or spreadsheets to demonstrate exactly how stupid of an argument that is.

A Brief List

Have you ever heard of the following players?

  • Erik Andersson
  • Jan Benda
  • Brad Church
  • Brian Felsner
  • Ryan Huska
  • Jeff Libby
  • Johan Lindbom
  • Jeff Mitchell
  • Ryan Mulhern
  • Kirk Nielsen
  • Terran Sandwith
  • Ronnie Sundin
  • Petri Varis

If that’s too esoteric, what about the following?

  • Philippe Audet
  • Ryan Bast
  • Karel Betik
  • Fredrik Bremberg
  • Jeff Buchanan
  • Sylvain Cloutier
  • Jonathan Delisle
  • Robb Gordon
  • Trevor Halverson
  • Victor Ignatjev
  • Mario Larocque
  • Patrice Lefebvre
  • Jan Mertzig
  • Peter Nordstrom
  • Remi Royer
  • Ryan Savoia
  • Dmitri Tertyshny
  • Brian Wesenberg
  • Brian White

What about these:

  • Keith Aldridge
  • Maxim Balmochnykh
  • Ryan Bonni
  • Paul Comrie
  • Dave Duerden
  • Miika Elomo
  • Erich Goldmann
  • Jorgen Jonsson
  • Geordie Kinnear
  • Jeff Lank
  • Petr Mika
  • David Moravec
  • Marc Rodgers
  • Dale Rominski
  • David Van Drunen
  • Alexei Vasiliev
  • Alexandre Volchkov
  • Dmitri Yakushin
  • B.J. Young
  • Jeff Zehr

These players represent a motley collection:

  • First-round picks who didn’t pan out for one reason or another
  • European legends who didn’t stick in the NHL
  • Minor league players who had a brief callup
  • Capitals prospects
  • An absolutely tragic case in Dmitri Tertyshny, who died after his rookie year in an offseason boating accident

But the question of what they all have in common is twofold: they all played games in just one single NHL season, and that one season was (chronologically by list) one of 1997-98, 1998-99, or 1999-00.

Why those three years?  Because after years of being beaten up physically and mentally on the ice, Mario Lemieux decided to retire.  And after a three-year retirement that covered the above years, he decided to come back in 2000-01.

In the 43 games that he played that year, Lemieux chalked up 35 goals and 41 assists for 76 points.  And the Penguins, who were 15-14-6 when he came back, went 27-14-3 the rest of the way and ended up in the conference finals.  And Mario Lemieux, who played just 43 games, finished 2nd in voting for the Hart Trophy as league MVP.

There lived a brilliant scientific mind named Erwin Schrodinger, who is best known (outside of scientific circles) today for a thought experiment which involved a cat that was simultaneously alive and dead.  I’m not a scientist, but my understanding of it is this:  A particular theory was proposed over the state of being.  Schrodinger felt that it wasn’t right, so he followed the theory through to its natural conclusion, which yielded results that were complete nonsense.  To demonstrate that it was nonsense, he came up with a demonstration in which is that it would be possible to have a cat in a box who is simultaneously alive and dead.  It wasn’t designed to be some mind-blowing thing to demonstrate his own theory, it was a way to skewer a flawed argument.

One of my favorite people from history is the Greek philosopher Zeno, who today is best known for a series of baffling and humorous paradoxes.  Some of them claimed that motion was an illusion – after Zeno’s death, Diogenes (another of my favorite people) was taught Zeno’s paradoxes.  In response, Diogenes stood up and walked away to simply demonstrate the absurdity of Zeno’s claims.  I don’t believe for a second that Zeno didn’t believe that motion actually did not exist, but he had a unique gift for making people think.

I, too, can be an annoying SOB like Zeno and Diogenes, which is probably why their various antics crack me up.  Sometimes I bring data and facts to the table, sometimes I skewer bad arguments in order to demonstrate their absurdity or to gently hint that the argument needs cleaned up.  But I think a simple litmus test of “what happens if we follow this all the way through?” is kind of important.

If the NHL is filled with nothing but the best xxx number of hockey players, then it stands to reason that during the entire span season of 1997-98 through 1999-00, Mario Lemieux was a worse hockey player than the fifty-two players listed above….as well as a couple hundred more who played those years and had short careers.  That Lemieux came back and dominated to an extent very similar to what he had before would thus seem to be irrelevant.

The Best “How many?!?!?” Players in the World?

Since everything referred to (so far) with the player database has involved skaters, I’m doing the same here.

The actual dressed player limit has changed a great deal over the years.  At times there has been a minimum, sometimes there has not been.  But there’s a maximum number of skaters that there should be – dressed roster limit times number of teams.

Of course, this is ludicrous.  In the span of any given season:

  • Teams carry more players than they can dress in a game
  • Players get injured
  • Players get sent to the minors
  • Players retire mid-season
  • Players are traded for other players who are not in the NHL but are then called up
  • Players are suspended

The following chart shows how many skaters played so much as a single game in a given season.  Expansion years are in orange, non-expansion (or early years that may have included expansion teams) are in blue.  As expected, we see jumps in expansion years.


(Click to enlarge)

We see a few notable things on here:

  • During the “Original Six era” (1942-43 through 1966-67), we see fluctuations in the number of dressed skaters.  It’s nothing terribly dramatic, but there is movement on a year-to-year basis.
  • The most skaters ever dressed in a season was not an expansion year – it was 2003-04 (916 skaters).
  • The five seasons with the highest number of skaters were not expansion years – 2003-04, 2018-19 (906), 2015-16 (898), 2011-12 (894), and 2010-11 (891).
  • There are sharp drops in both the lockout-shortened seasons of 1994-95 and 2012-13.  I believe the reasons are twofold.  With 34 fewer games played per team, more players were able to stay healthy, or at least play through injuries that they might not in a full season, and coaches and GMs were less willing to shake things up as long as the team was in a playoff hunt – which lasted through to the very end of the season.

Now let’s take a look at now many additional players were dressed compared to the official (roster spots) x (teams).

In this graph, we can see the split between the dressed limit and the “extra” skaters who suited up.  In non-expansion years, the roster limit is in blue and the extra are in green; in expansion years, the roster limit is orange and the extra are maroon.  The first year is 1925-26 due to discrepancies in whether there was a roster limit, or even a roster minimum, before that.


(Click to enlarge)

There is a negative year in there (1929-30).  At the time, there was no minimum roster, so a couple teams dressed significantly fewer players than might be expected.

We’re getting somewhere, but it’s still not quite illuminating yet.

Maybe the better question is this: what is the percentage of extra skaters beyond the dressed limit, and does it move up or down with expansion?

If the talent pool is diluted by expansion, we should see a clear move up: more players being churned through rosters in an attempt to find actual NHL-caliber skaters in a changing landscape.  Normal years and early expansion years are in blue, modern (1967-present) expansion years are in red.

The formula used here is a basic growth: (number of skaters – dressed limit) / (dressed limit).


(Click to enlarge)

To be perfectly honest, I’m not certain what to make of this.  We see the same drops in the lockout-shortened seasons (obviously), and the highest year on record is 1991-92 – which is an expansion year.  On the other hand, it was immediately preceded by the four next-highest years on record, and it immediately dropped back down.

My expectation would be that the red expansion years would be noticeably higher than the blue years surrounding them, but the opposite seems to be the case.  One would expect that teams would churn through a greater number of players than before, both unencumbered by any type of loyalty to a (departed) player and also looking for someone to replace what they just lost.  And as for the expansion team(s), the search for talent in the expansion draft has traditionally yielded a roster full of wild cards; they tend to go through a lot of players to find out who can play at that level.

More research is needed to really delve into this and exactly what it means.

NEXT: Part 7 (Can’t these guys at least LOOK competitive?)