Advanced Stats: Should Rangers part with Stepan in Nash deal

At this point it’s anyone’s guess whether the long anticipated trade of Rick Nash will ever actually go down. This story has dragged on since February and it appears as if we are still no closer to a resolution. While countless rumored return packages have been proposed in the media, then dissected and criticized by the fans, until a trade is finally consummated or Scott Howson reveals publicly exactly what players he wants in return for his franchise player (which would be a violation of the league’s tampering rules) no one knows for sure how high a price Columbus is demanding.

However some close to the situation, including Aaron Portzline of the Columbus Dispatch, would seem to have a better idea than most. Portzline made an appearance on WFAN on July 7th (as noted here on SNYRangersblog) and indicated a deal could get done for Nash if the Rangers offered Brandon Dubinsky, Derek Stepan and a first-round pick. Since a trade has not been made it would seem the Rangers are unwilling to move Stepan. That belief is supported by Larry Brooks who wrote as much in the NY Post (again noted on this site) last month.

Obviously we know at this stage of their respective careers Rick Nash is a superior player to Derek Stepan. On that alone the Rangers could be considered foolish to not make the deal as outlined by Portzline. However it’s also evident Rangers management feels Stepan has only scratched the surface of his potential otherwise a deal would have already been completed. I can’t imagine the inclusion of either the first-round pick, hopefully a low one in 2013, or Dubinsky is a deal-breaker for the Blue Shirts in their pursuit of Nash. Clearly the club projects bigger and better things from Stepan and that’s why they are reluctant to make the deal.

Are the Rangers justified in their refusal to include Stepan in a potential swap for Nash? Advanced stats can help us make that determination.

Since we can’t tell the future we have to make educated guesses as to how each player will perform down the line. Finding players historically who produced similar numbers at similar stages in their careers as Stepan and Nash is paramount in this exercise. Once enough of those comparable players are identified we can project how Stepan and Nash might perform in the coming seasons by using the performances of their respective comparables as a gauge.

For Stepan I searched out players who began their NHL careers around the age of 20 or 21, produced offensive numbers more or less in line with that of a second-line center and who wasn’t a premier first-round draft pick. The latter of those criteria would eliminate guys like Messier, Gretzky, Lemieux, Crosby and Lecavalier.

With regard to Nash I was looking for big, goal scoring wingers. Now by big I guess I should clarify that they don’t necessarily have to be big but they have to play big. An example of that would be a guy like Pat Verbeek. Verbeek was just 5’9”, 192 but scored 522 career goals and earned an impressive 2,905 minutes in penalties. The comparables I was looking for didn’t have to necessarily be tough guys or prolific fighters but they needed to be guys who are unafraid to play physically to give themselves room on the ice. Perhaps the term “power forward” is fitting.

In addition to combing through my brain in search of possible fits I used Hockey Reference’s Similarity Scores to identify potential candidates. The Similarity Scores use adjusted point shares (explanation here) over the course of a player’s career and lists other players who have recorded similar totals throughout their own careers. From there I pared the list down using the criteria I listed above.

Here’s what I was left with (Note: I included Brad Richards as Torts has often said Stepan reminds him of a young Richards. The numbers might not bear that out but if Torts feels that strongly about a player then I felt I should include him):

Stepan – Richards, Derek Roy, Stephen Weiss, David Legwand, Doug Weight, Bernie Federko, Steve Sullivan, Andrew Cassels and Saku Koivu.

Nash – Jarome Iginla, Brendan Shanahan, Keith Tkachuk, Bill Guerin, John LeClair, Gary Roberts, Owen Nolan, Pat Verbeek and Dave Andreychuk.

For this part of the exercise we will examine the scoring totals and rates plus the players’ GVT totals and rates for their first two NHL seasons (since that’s how long Stepan has been in the NHL). Also, since these players played during different eras I normalized their scoring rates. For example; scoring was significantly higher in the 1980’s and early 1990’s than it is today. Scoring is up in today’s game versus what it was in the late 1990’s and up to the lockout. During Weight’s first two NHL seasons the average team total of goals was 291.5. In 2010-2011 and 2011-2012, Stepan’s first two seasons, the average team potted 226.5 goals. Dividing 226.5 by 291.5 we get 0.777. Multiplying 0.777 by Weight’s point totals puts his numbers on an even playing field with Stepan in this era.

Player

Age

Gms

G

A

Pts

Pts/Gm

Factor

Adj. Pts/Gm

GVT

GVT/Gm

Stepan

20

82

21

24

45

0.55

 

 

7

0.085

 

21

82

17

34

51

0.62

 

 

11

0.134

 

Average

82

19

29

48

0.59

-

0.59

9

0.110

Richards

20

82

21

41

62

0.76

 

 

6.9

0.084

 

21

82

20

42

62

0.76

 

 

7.9

0.096

 

Average

82

20.5

41.5

62

0.76

1.027000

0.78

7.4

0.090

Roy

20

49

9

10

19

0.39

 

 

2.6

0.053

 

22

70

18

28

46

0.66

 

 

7.8

0.111

 

Average

59.5

13.5

19

32.5

0.55

0.976300

0.53

5.2

0.087

Weiss

20

50

12

17

29

0.58

 

 

2.7

0.054

 

22

41

9

12

21

0.51

 

 

1.8

0.044

 

Average

45.5

10.5

14.5

25

0.55

0.976300

0.54

2.25

0.049

Legwand

20

81

13

28

41

0.51

 

 

4.6

0.057

 

21

63

11

19

30

0.48

 

 

4.1

0.065

 

Average

72

12

23.5

35.5

0.49

1.027000

0.51

4.35

0.060

Weight

21

53

8

22

30

0.57

 

 

5.8

0.109

 

22

78

17

31

48

0.62

 

 

8.1

0.104

 

Average

65.5

12.5

26.5

39

0.60

0.777000

0.46

6.95

0.106

Federko

20

31

14

9

23

0.74

 

 

3.9

0.126

 

21

72

17

24

41

0.43

 

 

2.1

0.029

 

Average

51.5

15.5

16.5

32

0.52

0.854700

0.45

3

0.058

Sullivan

21

16

5

4

9

0.56

 

 

1.1

0.069

 

22

54

13

25

38

0.70

 

 

8.2

0.152

 

Average

35

9

14.5

23.5

0.67

0.911470

0.61

4.65

0.133

Cassels

20

6

2

0

2

0.33

 

 

0.5

0.083

 

21

54

6

19

25

0.46

 

 

2.7

0.050

 

Average

30

4

9.5

13.5

0.45

0.793345

0.36

1.6

0.053

Koivu

21

82

20

25

45

0.55

 

 

7

0.085

 

22

50

17

39

56

1.12

 

 

11.1

0.222

Average

66

18.5

32

50.5

0.77

0.911470

0.70

9.05

0.137

 

The next step is to eliminate the highest and lowest rated comparables. Since we’d like to be conservative with our projections removing the top and bottom players from the study lessens the impact of any outliers, positive or negative.

The player with the highest GVT/Gm rating and the second best scoring rate is Saku Koivu. Since we want to use Richards, who has the highest scoring rate in this study, I won’t preclude his presence. So Koivu is out.

Andrew Cassels posted the lowest scoring and GVT rates. He’s gone too.

Next I took the GVT ratings for each of the remaining players for seasons three through eight (I used six seasons since that’s the remaining term on Nash’s contract) of their careers and listed them as total, average per season and average per game played.

GVT Next 6 Yrs.

GVT/Yr Next 6 Yrs.

GVT/Gm N6Y

Richards

82.9

13.82

0.182

Roy

68.7

11.45

0.160

Weiss

45.7

7.62

0.099

Legwand

43.7

7.28

0.105

Weight

66.3

11.05

0.159

Federko

75.1

12.52

0.164

Sullivan

69.1

11.52

0.155

Averages:

64.5

10.75

0.146

 

If we believe Stepan’s career follows a path similar to that of our average comparable he will be a 12 GVT player per season over the next six.

Now we move to Nash.

Gms

G

A

Pts

Pts/Gm

Factor

Adj Pts/Gm

GVT

GVT/Gm

Nash

674

289

258

547

0.81

1.0000

0.81

102.1

0.151

 

 

 

 

 

 

 

 

 

 

Iginla

708

285

285

570

0.81

1.0267

0.83

114.7

0.162

 

 

 

 

 

 

 

 

 

 

Shanahan

632

288

310

598

0.95

0.8221

0.78

93.7

0.148

 

 

 

 

 

 

 

 

 

 

Tkachuk

576

294

258

552

0.96

0.9023

0.86

110.3

0.191

 

 

 

 

 

 

 

 

 

 

Guerin

570

175

178

353

0.62

0.9203

0.57

64

0.112

 

 

 

 

 

 

 

 

 

 

LeClair

583

269

269

538

0.92

0.8988

0.83

127.4

0.219

 

 

 

 

 

 

 

 

 

 

Roberts

550

235

228

463

0.84

0.8105

0.68

95.9

0.174

 

 

 

 

 

 

 

 

 

 

Nolan

565

210

224

434

0.77

0.8919

0.69

80

0.142

 

 

 

 

 

 

 

 

 

 

Verbeek

623

257

235

492

0.79

0.7649

0.60

59.8

0.096

 

 

 

 

 

 

 

 

 

 

Andreychuk

631

278

341

619

0.98

0.7649

0.75

82.3

0.130

 

Repeating the process from earlier, we will remove John LeClair and Pat Verbeek from any more consideration.

Next 6 Years GVT

N6Y GVT/Gm

Iginla

106.2

0.221

Shanahan

101.7

0.214

Tkachuk

78.5

0.196

Guerin

80.9

0.177

Roberts

56.9

0.145

Nolan

65.6

0.154

Andreychuk

81.9

0.181

Average:

13.6

0.185

 

An average year from our comparables, which we are using to project Nash’s future performance, is a 15.2 GVT player. The difference in performance between Nash and Stepan according to projected GVT is 3.2 per season.

Now in addition to Stepan the Rangers would also have to include Dubinsky and a 1st round draft pick for Nash. Dubinsky had an off year last season but for his career he has averaged a 0.088 GVT per game, or 7.24 GVT per 82-game season. It’s safe to assume given Dubinsky’s age that he is not in the midst of a career regression and thus should progress back to his career GVT mean. Therefore, combining the average projected GVT values of Stepan and Dubinsky for the next six seasons we would have a total GVT of 19.24. Nash is worth just 15.2. Without even gauging the value of the first-round pick, the return package of Dubinsky and Stepan is likely to best the performance of Nash in terms of GVT over the next six seasons. But what if the Rangers acknowledge that likelihood but feel an incremental improvement (i.e. Nash) in the next couple of seasons will yield the game’s ultimate prize, a Stanley Cup championship?

With that in mind, let’s project the performances of Nash and Stepan just for the next two seasons. The following table shows the projected averages in GVT for the next two seasons using the historical performances of the comparable players.

 

Total GVT

GVT/Yr.

GVT/Gm

GVT/82 Games

Nash

32.8

16.4

0.226

18.57

Stepan

24.48

12.24

0.144

11.81

There is a greater disparity (6.76 GVT) now between the respective projections for Nash and Stepan for the next two seasons. Although if we assume Dubinsky pitches in with his career average it still tilts the balance in the favor of not making the proposed deal.

I imagine this is similar to the type of analysis which has already been done by the Rangers front office. Truthfully, I would have preferred to have included more comparable players but I must face the reality and the responsibility of working a regular job 50 hours a week thus leaving me little time to do this type of study. As it was I spent many hours on this.

Of course this is just part of the overall evaluation process every team must go through before making a move of this magnitude. Scouting obviously plays a big role in these decisions. It’s possible that moving to a better team will cause a spike in the performance of Nash. Scouts and evaluators like his game a lot and if surrounded by a better supporting cast Nash’s performance could explode. He also might have an impact on the players around him in New York. Perhaps his presence could help elevate the performances of some of his potentially future teammates. That’s a question the evaluators need to answer.

There you have it, a statistical argument against including Derek Stepan in any trade for Nash. The drop off from the projected performances of Nash to Stepan is not great enough to improve the Rangers either short-term or long-term when the additional assets required to make the deal are also accounted for.