Analytics and lineups: Why Vigneault should keep Nash with Zuccarello

I have had a significant difference of opinion with Alain Vigneault over the choice of forward lines since the start of the season, but just as I prepared to unleash my statistics and arguments upon the Internet the team went on a winning streak and it felt silly to criticize the coach’s choices at the time.

Now the team is still 4-1-0 in their past five games, but Vigneault made the one move I’ve been waiting for in the game against the Minnesota Wild: putting Rick Nash on a line with Mats Zuccarello and Derick Brassard. Because while Rick Nash has been scoring like there is no tomorrow to start this season, his possession numbers are the 2^nd worst on the team (ahead of Tanner Glass) – which doesn’t bode well for the future.

Shot generation and suppression

While analytics people generally talk about possession (usually in the form of Corsi Rel. for individual players), there are actually two elements which determine a player’s ability to drive possession: shot generation and shot suppression. The two are related, since if you have the puck a lot you are going to generate more shots while the opposition won’t have opportunities to shoot, but they aren’t one and the same which means some players are a lot stronger in one or the other – and Rick Nash is one of them.

Those of you who follow me on Twitter know that I post charts that break down how players have contributed in different areas of the game. These charts have six categories: three for point scoring, and three for possession. The ones for scoring are just regular goals/assists/points per 60 minutes in the different game states (5v5, 5v4, 4v5), while the possession-related ones are slightly more complicated than that.

For measuring a player’s contribution to his team’s possession game I use my own metric which I have named "vs. Exp.", since it measures how a player has performed versus the expected scenario based on the quality of his teammates, competition, and his zone starts (or rather, the zone starts are factored out by only counting events 10+ seconds after OZ/DZ face offs). The shot metric that I adjust is Fenwick (shots on goal + missed shots) per 60 minutes. FF = Fenwick For, FA= Fenwick Against, and FD = Fenwick Differential.

This is how Rick Nash has measured up over the past two regular seasons as a Ranger:

The six categories are measured in standard scores to make them comparable; anything over 1 standard deviation above the mean is very good (roughly top 15%) while anything under 1 St. Dev. below the mean is very bad (roughly bottom 15%). Nash’s 1.355 score in 5v5 P/60 places him 26^th in the league during this period, just ahead of James Neal and Patrice Bergeron, while his 2.916 score in 5v5 G/60 puts him 2^nd behind Gustav Nyquist.

As we can see Rick Nash has been a positive factor for the Rangers’ puck possession during his tenure in New York, but it is solely down to his excellent shot generation skills – in fact he is downright poor at suppressing shots. So Nash has been an incredibly potent offensive weapon 5v5, but he doesn’t come without weaknesses.

This is how Nash has fared so far in 2014-15 (though a very small sample):

While his scoring is literally off the charts, he has been extremely poor at generating possession for the team. Some of that has to be blamed on Nash himself, but I believe that the main culprit is Vigneault’s choice of linemates. Let’s inspect his two most common linemates so far this season, Chris Kreider and Martin St. Louis:

Kreider only has one statistically significant season in his career, but it tells us that he is a good shot generator and possession driver, while being average or slightly above average in the other categories. Overall he was a player in a similar mold to Nash as his primary strengths were generating shots and scoring goals. A good player, but perhaps not the best complement to Nash’s skillset.

With St. Louis you can clearly see a trend of his 5v5 play the past four seasons; he is an excellent and well-balanced point scorer – but he is a clear detriment to the possession game. We can also see that, apart from the shortened 2012-13 season, his scoring is in decline while his contributions to the possession game are pretty consistent.

Forming a line with two poor and one merely average shot suppressors, especially with one playing out of position as a center, never looked like a good idea on paper – and while the line scored goals (at an unsustainable pace at that) it was mostly spending time chasing the puck in their own end.

When I look at the characteristics of Nash as a player, two things stand out:

1. He needs good shot suppressors on his line to make up for his defensive deficiencies
2. He would fit well with pass-first players that get him the puck

So which player does immediately come to mind? Mats Zuccarello, who is the 7th best shot suppressor in the league over the past four seasons (out of players with 1500+ minutes) and perhaps the most creative passer on the team.

Zuccarello and Brassard seem like a package deal under the coaching of Alain Vigneault, so while Brassard isn’t the perfect center for our top scorer – it seems like a necessity if he is to play on the same line as Zuccarello. The chart clearly shows how Zuccarello and Nash are nearly the perfect complements for each other as they cover for their respective weaknesses while both are great 5v5 players in their own right. Brassard is merely average, but he is a well-balanced average and that is all it takes to be a good fit on this line.

However, building a lineup isn’t just about optimizing one line – it is about fielding four (or at least three) well-balanced lines that can be trusted in most situations. This is how I’d form the rest of the lines, assuming a fully healthy roster:

This is another great, well-balanced line that, at least on paper, would be one of the better 5v5 lines in the league. Stepan and Hagelin are well above average in every 5v5 category and can cover for St. Louis’ possession deficiencies while every player on the line is a potent 5v5 point scorer.

Kreider-Hayes-Duclair

I had a hard time deciding whether Kreider or Hagelin was to be the one playing with the two rookies, but in the end I decided to go with Hagelin for the top line due to his track record of significantly improving star players (Gaborik, Nash) and his better 5v5 stats in a larger sample. I also have the feeling that Kreider is better at creating his own opportunities and is therefore less reliant on linemates for his 5v5 scoring. Hayes and Duclair have both performed well 5v5 so far, and I have no doubt that this line could perform, especially since the other two scoring lines likely would draw the tough defensive matchups leaving these three to feast on lesser competition. There is no comparison chart as the data samples are too small for Hayes and Duclair.

Stempniak has been playing well so far and it may seem harsh to demote him to the 4^th line, however that is where he belongs. First of all his scoring isn’t sustainable, Stempniak has been nothing more than an average 5v5 scorer over the past four seasons – so his sudden increase in scoring pace is likely to regress back towards the mean. Stempniak, due to his poor shot suppression, isn’t ideal for an Alain Vigneault 4^th line but putting the 1^st line caliber (5v5) Carl Hagelin on the 4^th line because his shot suppression is better than Stempniak’s is to me a bit like taking the engine out of your LaFerrari and putting it in your Prius.

Dominic Moore is an excellent 4^th line center, but you have serious problems if he is in your top-9 as he is a purely defensive player. Malone is well-balanced and could likely be plugged onto any line without being a detriment.

Tanner Glass should obviously sit as he unfortunately isn’t an NHL caliber player.

In terms of general quality I’d say this 4^th line matches up well with the 4^th line of the 2013-14 Rangers, but in terms of fit for the role it is a step back as the latter were far better shot suppressors:

Conclusions

With the emergence of Kevin Hayes and Anthony Duclair the Rangers have the personnel to form a potent and well-balanced set of forward lines, but due to the specific skillset of some of the players the right combinations have to be chosen to eliminate weaknesses. Rick Nash with Martin St. Louis was the wrong combination. Rick Nash with Mats Zuccarello should be the right combination.