|
During the 1994 playoffs, Brian Leetch was a wonder to behold. He led the playoffs in scoring with 34 points, had a +19 rating and led the New York Rangers to their first Stanley Cup in 54 years. In fact, Leetch’s performance was so impressive that it is the greatest playoff performance by a non-goaltender in NHL history, just ahead of Wayne Gretzky’s 47-point performance in 1985 and Henrik Zetterberg’s Conn Smythe winning playoffs in 2008.
Huh?
While Zetterberg was certainly excellent during the 2008 playoffs, surely it’s a stretch to consider that he put in the third-best playoff performance of all time, no? Yet that’s what the numbers were telling me, so clearly the numbers were wrong. To find out why they were wrong, it’s useful to back up a bit.
Normalizing for the playoffs
This problem first jumped in my face when I posted the list of recent playoff GVT leaders in my column two weeks ago. For my readers’ benefit, here it is again:
Name Team P Season GP G/GAA A/GA Pts/SA +-/Sv GVT
Mario Lemieux Penguins F 90_91 23.0 16.00 28 44 14.000 13.1
Mario Lemieux Penguins F 91_92 15.0 16.00 18 34 6.000 12.5
Patrick Roy Canadiens G 92_93 21.6 2.13 46 647 0.929 21.2
Kirk McLean Canucks G 93_94 25.7 2.29 59 820 0.928 20.8
Martin Brodeur Devils G 94_95 20.4 1.67 34 463 0.927 19.6
John Vanbiesbrouck Panthers G 95_96 22.2 2.25 50 735 0.932 22.8
Eric Lindros Flyers F 96_97 19.0 12.00 14 26 7.000 14.6
Olaf Kolzig Capitals G 97_98 22.5 1.95 44 784 0.944 25.5
Dominik Hasek Sabres G 98_99 20.3 1.77 36 587 0.939 21.1
Ed Belfour Stars G 99_00 24.1 1.87 45 651 0.931 16.5
Patrick Roy Avalanche G 00_01 24.2 1.70 41 622 0.934 24.8
Peter Forsberg Avalanche F 01_02 20.0 9.00 18 27 8.000 16.0
J.S. Giguere Ducks G 02_03 23.4 1.62 38 697 0.945 32.3
Jarome Iginla Flames F 03_04 26.0 13.00 9 22 13.000 18.4
Dwayne Roloson Oilers G 05_06 19.3 2.33 45 618 0.927 20.7
Daniel Alfredsson Senators F 06_07 20.0 14.00 8 22 4.000 15.8
Marc-Andre Fleury Penguins G 07_08 20.9 1.97 41 610 0.933 19.3
Do you see the problem? Mario Lemieux’s two great playoff seasons sit dead last among the single-season GVT leaders, miles behind, among others, Jarome Iginla’s MVP-caliber performance in 2004. Granted, scoring was higher during the 1991 playoffs, 3.22 goals per game compared to 2.12 in 2004, but still! I checked the list of top forwards and defensemen in the playoffs:
Best playoff seasons by a forward or defenseman
Name Team P Season GP G/GAA A/GA Pts/SA +-/Sv GVT
1 Brian Leetch Rangers D 93_94 23 11 23 34 19 20.0
2 Wayne Gretzky Oilers F 84_85 18 17 30 47 28 19.6
3 Henrik Zetterberg Red Wings F 07_08 22 13 14 27 16 18.4
4 Jarome Iginla Flames F 03_04 26 13 9 22 13 18.4
5 Brad Richards Lightning F 03_04 23 12 14 26 5 17.3
6 Petr Sykora Devils F 00_01 25 10 12 22 15 16.7
7 Peter Forsberg Avalanche F 01_02 20 9 18 27 8 16.7
8 Joe Sakic Avalanche F 00_01 21 13 13 26 6 16.5
9 Paul Coffey Oilers D 84_85 18 12 25 37 26 16.0
10 Joe Sakic Avalanche F 95_96 22 18 16 34 10 15.8
11 Daniel Alfredsson Senators F 06_07 20 14 8 22 4 15.8
12 Mark Messier Rangers F 93_94 23 12 18 30 14 15.7
13 Martin St. Louis Lightning F 03_04 23 9 15 24 6 15.6
14 Patrik Elias Devils F 00_01 25 9 14 23 11 15.6
15 Cory Stillman Hurricanes F 05_06 25 9 17 26 12 15.3
16 Marian Hossa Penguins F 07_08 20 12 14 26 8 15.1
17 Brian Rafalski Devils D 00_01 25 7 11 18 10 14.9
18 Reggie Leach Flyers F 75_76 16 19 5 24 0 14.9
19 Wayne Gretzky Oilers F 83_84 19 13 22 35 18 14.8
20 Eric Lindros Flyers F 96_97 19 12 14 26 7 14.6
Fully 12 of these 20 seasons were in the last 8 years, and Mario is nowhere to be seen. Something was amiss here, and I knew what it was: I was normalizing incorrectly. How exactly do you normalize?
Normalizing for the regular season
In 1945, Maurice Richard became the first player to score 50 goals in a season. In 1953, his rival Gordie Howe nearly matched him by scoring 49 goals during the year. In 1982, Marcel Dionne scored 50 goals, reaching the 50-goal plateau for the 5th time in his career. Finally, this year, Alexander Ovechkin led the entire NHL with 56 goals.
All four of these players scored about 50 goals. Which one of these performances is the most impressive?
Obviously, we understand that goal scoring must be placed in the context of scoring levels and schedule length. Richard scored his 50 goals in only 50 games, but he did so in a watered-down league that was allowing 3.68 GPG (Goals Per Game). Howe’s performance came in 70 games, but at a scoring level of 2.40 GPG. Dionne scored his 50 goals during the highest-scoring year in league history, an 80-game schedule featuring 4.01 GPG, while Ovechkin got the longest schedule of all at 82 games, though he was in a league that only scored 2.85 GPG. After adjusting, Howe’s 49 goals become 82, the highest total of the four, Richard’s become 76, Ovechkin’s become 67 and Dionne only 44.
Everyone understands how to normalize during the regular season, but people rarely do it. Whenever a player is being discussed for the Hall of Fame, his career goals are always brought up, as if such a number can be discussed in a vacuum. Many players from the 80s who have been under consideration in recent years have benefited from their inflated totals (Cough! Glenn Anderson! Cough!).
Normalizing for the playoffs is even harder. One element, goals per game, is easy to take into account. However, what is the “schedule length” of the playoffs?
Several years ago, when I first applied GVT to the playoffs, I tried to come up with a method that was as fair as possible. My premise was the following: during the regular season, by normalizing for schedule length and scoring levels, each player gets the same opportunity to accumulate GVT. Therefore, in the playoffs, each player should have the chance, on average, to accumulate the same as well. The way to do this was to normalize the total number of games in the playoffs by the number of teams in the league, and I brought this back to an average of 7 games per team, which has been about the historical average of playoff length in the NHL.
In 1991, there were 184 games played on total in the playoffs (92 games * 2 teams), and there were 21 teams in the league, so Mario’s 1991 playoffs were normalized by 7 * 21 / 184 = 0.80. By contrast, during the 2004 playoffs, there were 178 games played, but 30 teams in the league, so Iginla’s 2004 numbers were normalized by 7 * 30 / 178 = 1.18.
Now some of my readers may find this idiotic, but in fact from a career point of view this makes perfect sense. In a 30-team league, you will make the playoffs less often than in a 21-team league, and you will go deep into the playoffs less often (good luck putting up big career playoff numbers if you don’t play for the Red Wings). However, from a single-season point of view, this normalization is wrong, and the reason is sample size.
Some people consider Terry Sawchuk’s 1952 playoffs, where he allowed 5 goals in 8 games (on 229 shots!) to be the greatest playoff performance of all time. I respectfully disagree. While Sawchuk’s play was no doubt phenomenal, and his save percentage of 0.978 otherworldly, he only kept it up for 8 games. Jean-Sebastien Giguere managed a save percentage of 0.945 over 23 games in 2003, which is better in my book. While a player’s level of play during a 50-game regular season is probably a good indicator of what he would have put up over 80 games, it’s clearly a stretch to take an 8-game playoff and extrapolate that to 20 games. Is it fair to Sawchuk that he never got to show his stuff over four seven-game series? No, but that’s life. We can only judge him on the data we have. So if Mario played in 23 games in 1991 and Iginla 26 games in 2004, then that’s what we’ll count.
Plugging the new numbers in, here is what we get:
Best playoff seasons by a forward or defenseman
Name Team P Season GP G A Pts +-/ GVT
1 Brian Leetch Rangers D 93_94 23 11 23 34 19 18.5
2 Wayne Gretzky Oilers F 84_85 18 17 30 47 28 17.3
3 Mario Lemieux Penguins F 90_91 23 16 28 44 14 15.9
4 Joe Sakic Avalanche F 95_96 22 18 16 34 10 14.7
5 Mark Messier Rangers F 93_94 23 12 18 30 14 14.7
6 Brad Richards Lightning F 03_04 23 12 14 26 5 14.5
7 Jarome Iginla Flames F 03_04 26 13 9 22 13 14.4
8 Paul Coffey Oilers D 84_85 18 12 25 37 26 14.4
9 Wayne Gretzky Kings F 92_93 24 15 25 40 6 14.0
10 Henrik Zetterberg Red Wings F 07_08 22 13 14 27 16 14.0
11 Pavel Bure Canucks F 93_94 24 16 15 31 8 13.7
12 Mario Lemieux Penguins F 91_92 15 16 18 34 6 13.6
13 Joe Sakic Avalanche F 00_01 21 13 13 26 6 13.5
14 Wayne Gretzky Oilers F 83_84 19 13 22 35 18 13.5
15 Peter Forsberg Avalanche F 01_02 20 9 18 27 8 13.4
16 Martin St. Louis Lightning F 03_04 23 9 15 24 6 12.9
17 Eric Lindros Flyers F 96_97 19 12 14 26 7 12.9
18 Wayne Gretzky Oilers F 87_88 19 12 31 43 9 12.6
19 Al MacInnis Flames D 88_89 22 7 24 31 6 12.6
20 Petr Sykora Devils F 00_01 25 10 12 22 15 12.5
Much better! Mario takes his rightful place in 3rd, while Zetterberg’s superb 2008 season still shows up in 10th place overall. Granted, this list is still dominated by players, from the 80s and later, when the playoffs got longer, but that is only normal.
There are several morals to this story. One is that when a schedule is short, you cannot normalize for schedule length. The other is that numbers must always pass a “feel” test: if numbers don’t feel right, find out why. Maybe they really are telling you something you don’t know, or maybe they are flawed.
2009 Conn Smythe Watch
Name Team P GP Val OG Val D Total
1 Sidney Crosby Penguins F 17.0 8.8 1.5 10.3
2 Evgeni Malkin Penguins F 17.0 8.5 0.6 9.1
3 Johan Franzen Red Wings F 16.0 5.2 1.0 6.1
4 Henrik Zetterberg Red Wings F 16.0 4.7 1.0 5.8
5 Chris Osgood Red Wings G 16.5 5.0 0.0 5.0
6 Daniel Cleary Red Wings F 16.0 3.3 1.7 4.9
7 Bill Guerin Penguins F 17.0 3.3 1.4 4.6
8 Nicklas Lidstrom Red Wings D 14.0 3.4 1.0 4.5
9 Eric Staal Hurricanes F 18.0 3.7 0.3 4.0
10 Martin Havlat Blackhawks F 16.0 3.4 0.5 3.9
11 Sergei Gonchar Penguins D 15.0 2.9 0.8 3.7
12 Brent Seabrook Blackhawks D 17.0 2.6 0.8 3.3
13 Jonathan Toews Blackhawks F 17.0 2.9 0.4 3.3
14 Patrick Kane Blackhawks F 16.0 3.7 -0.4 3.3
15 Valtteri Filppula Red Wings F 16.0 2.2 1.0 3.2
This will be the last time you will be seeing the Hurricanes or Blackhawks players here; as befits two teams that mostly dominated their respective conference finals, the Penguins and Red Wings players dominate the top of the list as well. I have filtered out players from teams that have already been eliminated, although Jonas Hiller and Alexander Ovechkin would still rank 1st and 4th if they were included.
Should the Penguins lose in the final, we will likely end up with yet another situation like those I described two weeks ago where the best player in the playoffs played for the finalist rather than the winner. Barring a complete domination of the series by Detroit or an insane final from somebody like Franzen or Osgood, it’s likely that Crosby or Malkin will be deserving of the Conn Smythe regardless of the series result.
Tom Awad is an author of Hockey Prospectus.
You can contact Tom by clicking here or click here to see Tom's other articles.
| 
0 comments have been left for this article. 
Conn Smythe Watch: Lundqvist Leads