After Martin St. Louis was traded to the New York Rangers on March 5, he appeared to struggle to produce in his new role. In 19 regular season games with the Rangers, he scored one goal (shorthanded) and had eight points. Things are looking up for him these playoffs (six goals and thirteen points in 18 games), but for a while there it looked pretty bad.
I have been curious about whether this was just an extended slump driven by a (very) low shooting percentage or if something about his usage in New York had changed things for him. His TOI/60, for instance, went from 21.3 with the Lightning to 18.4 with the Rangers, and the decrease came across the boad (EV, PP, & SH). [All stats are 5v5, for the regular season only, and are are from extraskater.com and stats.hockeyanalysis.com unless otherwise noted.]
What has changed for St. Louis?
-
Time on ice (as noted)
-
Zone start % has increased, indicating that he has been getting more favorable starts with the Rangers than with the Lightning. This isn’t surprising as Alain Vigneault is known for optimizing player usage across his lineup.
-
His personal shooting percentage has dropped from 17.4% in his 62 Lightning games this season to a pitiful 2.7% with the Rangers, and that can explain a large portion of the drop in goals. His shooting percentage for the season as a whole is 14.7%, which is a bit high compared to his career rates that hover around 11%.
-
His shots per 60 minutes at 5v5 (regular season) are down from 7.4 with the Lightning to 6 with the Rangers. It appears that this is mostly because he’s getting credited with more missed shots than he used to, which could be a function of a difference in how the scorers at Madison Square Garden count on-goal shots compared to the Tampa Bay Times Forum. The Forum is known for overcounting shots. It could also be attributed to other differences between the two teams, of course. But his missed shots have increased by 0.8 shot attempts per 60 minutes (about 4 games). There’s still a decrease of about half a shot on goal per 60 minutes that this doesn’t account for.
-
He isn’t one of the top shot attempt generators on his team anymore. While his CF/60 and CA/60 rates are very similar on both teams, his relative corsi differential is much lower on the Rangers than it was on the Lightning. This means that the Rangers generated relatively more shot attempts while St. Louis was off the ice than they did while he was on. This was far less stark when he was on the Lightning.
-
His personal share of the shots, shot attempts, and points generated while he’s on the ice is decreasing. He’s being relied on more for what he can do off the puck than he was when on the Lightning.
|
St. Louis’s personal share of on-ice generated |
||||||
|
|
Goals |
Assists |
Points |
Shots |
Corsi |
Fenwick |
|
Lightning |
34.00% |
57.60% |
72.00% |
24.80% |
22.70% |
23.10% |
|
Rangers |
0.00% |
50.00% |
50.00% |
19.20% |
20.90% |
22.20% |
In essence, the majority of the change in St. Louis’s personal scoring can be attributed to a decreased shooting percentage, but there are other things happening that provide context to that. He’s getting more favorable territorial assignments, so it’s very hard to understand why his scoring isn’t better than it is.
He is shooting slightly less and his linemates are shooting more, so his role on his line appears to be changing. As his shooting percentage returns to a more reasonable number (somewhere closer to 9 or 10% than to 3%) the goals will come. It does seem, however, that the slight changes in usage could potentially continue to depress his scoring from what it was with the Lightning.
I then went in to look at St. Louis’s WOWY (“With Or Without You”) numbers. These help to decipher some of the impact an individual player has on the teammates he plays with most. There are some caveats to using these numbers. Unlike the data at ExtraSkater, the WOWY data at HockeyAnalysis is not broken down by team. That means that this data compares St. Louis’s season as a whole to, for instance, Tyler Johnson’s season as a whole. [Data from Martin St. Louis’s WOWY page at stats.hockeyanalysis.com.]
In the end, I feel that looking at what St. Louis’s teammates did with or without him allowed me to escape the worst of this sampling issue. If he played with a player, he was on that team. If a player was on the ice without him, it didn’t make an enormous difference whether that was because he was on the team or simply on the bench. (It does make some difference, but not too much to make this unusable.) This is not true when looking at, for instance, Tyler Johnson’s effect on St. Louis’s numbers, so I skipped that part of the analysis.
I chose the top 6 forwards and top 4 defensemen from each team in terms of how much time they played with St. Louis. Here are the 20 players in order by TOI with St. Louis.
|
Player |
Pos |
|
Tyler Johnson |
C |
|
L |
|
|
D |
|
|
D |
|
|
D |
|
|
D |
|
|
C |
|
|
L |
|
|
C |
|
|
C |
|
|
C |
|
|
D |
|
|
D |
|
|
D |
|
|
D |
|
|
C |
|
|
C |
|
|
C |
|
|
L |
|
|
L |
What do the WOWY numbers show?
In goals:
With one exception all Lightning players did better with St. Louis than without when it came to generating goals. That is to say that St. Louis’s top Lightning linemates got an offensive (goals) bump playing with him. The exception was Valtteri Filppula, who saw a 58.7 On-Ice Goals For % without St. Louis but only 42.9% with him. The majority of the difference for Filppula-St. Louis came in decreased Goals For, but Goals Against also went up when they were together.
All Rangers did better in goal prevention when on ice with St. Louis than without him. Brad Richards, Anton Stralman, Carl Hagelin, and Rick Nash all generated fewer goals with St. Louis than without him but all other Rangers had an offensive bump. Richards did particularly bad goal-wise with MSL. The pair failed to generate a single 5v5 goal when on ice together in 19 games. This was a loss of nearly a goal per 20 minutes of 5v5 ice time for Richards (0.901).
In shot attempts (Corsi):
Four Lightning players saw more Shot Attempts For while on ice with St. Louis: Tyler Johnson, Ondrej Palat, Matt Carle, and Valtteri Filppula. All Bolts except Filppula saw more Shot Attempts Against with St. Louis than without him. Only Carle and Filppula treaded water in CF% but for both the difference was very small. In other words, these Lightning players generally did better, especially defensively (shot attempts), without Marty St. Louis than with him.
As for the Rangers, Brad Richards, Dan Girardi, Derek Stepan, and Rick Nash all saw attempts for increase when on ice with St. Louis. The rest saw a decrease in attempts for. Richards, Girardi, Hagelin, and Nash also saw defensive bonuses. Those four players ended up with overall better shot attempt ratios (CF%). All other Rangers did worse with St. Louis.
|
Difference between Player’s performance with MSL vs without |
|||||||
|
Player |
Pos |
GF20 |
GA20 |
GF% |
CF20 |
CA20 |
CF% |
|
JOHNSON, TYLER |
C |
0.402 |
0.099 |
6 |
1.34 |
2.28 |
-1.4 |
|
PALAT, ONDREJ |
L |
0.307 |
0.142 |
1.4 |
0.860 |
1.49 |
-0.8 |
|
CARLE, MATT |
D |
0.432 |
-0.01 |
11.7 |
1.210 |
0.95 |
0.4 |
|
HEDMAN, VICTOR |
D |
0.419 |
0.199 |
4.3 |
-0.270 |
0.61 |
-1.2 |
|
GUDAS, RADKO |
D |
0.641 |
0.079 |
14.4 |
-2.370 |
1.35 |
-4.8 |
|
SALO, SAMI |
D |
0.550 |
0.067 |
10.2 |
-0.790 |
2.62 |
-4.8 |
|
STAMKOS, STEVEN |
C |
0.599 |
-0.128 |
15.2 |
-2.840 |
0.4 |
-4.1 |
|
MALONE, RYAN |
L |
0.169 |
-0.669 |
30 |
-0.920 |
2.57 |
-4.6 |
|
KILLORN, ALEX |
C |
0.394 |
0.023 |
7.8 |
-1.520 |
0.48 |
-2.7 |
|
FILPPULA, VALTTERI |
C |
-0.420 |
0.011 |
-15.8 |
1.250 |
-0.01 |
1.6 |
|
RICHARDS, BRAD |
C |
-0.901 |
-0.442 |
-51.9 |
0.510 |
-4.09 |
6.3 |
|
GIRARDI, DAN |
D |
0.447 |
-0.175 |
15.6 |
0.670 |
-0.28 |
1.2 |
|
STAAL, MARC |
D |
0.187 |
-0.306 |
18.1 |
-1.190 |
1.19 |
-3.2 |
|
STRALMAN, ANTON |
D |
-0.342 |
-0.501 |
13.8 |
-1.980 |
1.16 |
-4.2 |
|
MCDONAGH, RYAN |
D |
0.094 |
-0.303 |
13.5 |
-0.550 |
0.87 |
-1.7 |
|
BRASSARD, DERICK |
C |
1.120 |
-0.149 |
25.8 |
-2.540 |
1.17 |
-5 |
|
ZUCCARELLO, MATS |
C |
0.690 |
-0.007 |
13.3 |
-1.960 |
4.32 |
-8.4 |
|
STEPAN, DEREK |
C |
0.117 |
-0.256 |
14.4 |
0.240 |
3.48 |
-3.9 |
|
HAGELIN, CARL |
L |
-0.575 |
-0.469 |
-3 |
2.640 |
-2.49 |
6.2 |
|
NASH, RICK |
L |
-0.210 |
-0.33 |
8.4 |
0.190 |
-2.25 |
3.2 |
[Bolded items are offensive or defensive bumps gained while playing with St. Louis.]
What conclusions can be drawn from all this? Keeping in mind that WOWY data doesn’t untangle questions of usage or competition and that it’s impossible to know how much the change in teams is affecting the data, there are still some trends we can identify.
Mainly, both teams were seeing an offensive boost from the contributions of Martin St. Louis, but the Rangers have gotten a greater defensive boost from him than the Lightning did.
St. Louis and Brad Richards may not have gotten any goals while on ice together, but they played well enough together otherwise (59.2 CF%) to think that the scoring slump was probably temporary. And their playoff performance appears to be bearing that out. Playoffs are different–seeing the same opponents several games in a row will skew the numbers to some extent–but it’s not unlikely that this pairing will be productive in 2014-15.
St. Louis had better Corsi For % with three Rangers than with any Lightning players. Those players are Richards (59.2), Hagelin (60), and Nash (56.6). While sample size is a bit of an issue with Hagelin and Nash, it’s clear that linemates are not particularly holding St. Louis back. Overall his CF numbers were better with his top Rangers linemates than they were with his top Lightning linemates.
All in all, the underlying WOWY numbers put this trade in a different perspective than the drama surrounding it. While no one should be happy to lose Martin St. Louis, who is still a very productive player, he does appear to have a slightly better effect for the Rangers than he did for the Lightning–at least when one takes the vagaries of shooting percentage into account.
At the same time, St. Louis’s overall possession statistics are slightly concerning, particularly in terms of future performance. His 50.4 CF% (5v5) for the season put him in a five-way tie for 221st among the 435 skaters with 62 or more games played this season. While this is by no means a bad number for an NHL player these days, it is not an elite number. Nine Lightning players and thirteen Rangers, including Ryan Callahan, put up a better possession performance this season.
This is a higher number than St. Louis has generated over the past three seasons–largely because the Lightning as a whole were terrible at generating shot attempts over that time. This matters because past shot attempt differential (corsi) does a better job of predicting future goal differentials than past goal differential does. St. Louis has done quite well at scoring while operating at or below 50% CF in the past. It’s to be expected that as his shooting percentage regresses upward his scoring rates will come closer to what is expected from a player making $5.625M against the cap.
It is possible, of course, that he will never completely regain his form. He is aging, after all. It remains to be seen just how far St. Louis can push the envelope as he approaches 40. We should assume a decrease in ice time, although how much of a drop is yet to be seen. This is certainly something to keep an eye on as time passes. Overall, however, he remains a valuable asset who seems to be fitting in well with Alain Vigneault’s Rangers, even if his regular season production didn’t show it.
