clock menu more-arrow no yes mobile

Filed under:

Optimizing the Predators’ 2017-18 Lineup

A look at recent research on how offense is created and what that means for the Nashville Predators line combinations.

Anaheim Ducks v Nashville Predators - Game Six Photo by Sanford Myers/Getty Images

If you’ve read discussions of players I’ve done, I frequently include their passing profiles, which are the result of work done by Hockey Graphs writer Ryan Stimson on ways that offense is generated. Stimson’s research doesn’t stop at describing the style of a player’s game, though—he’s done that as part of looking at how using players with complementary styles can make lines more successful offensively. (His research also applies to defense pairings, but getting consistent offense from the defense isn’t something the Nashville Predators struggle with.)

With changes all over the lineup and questions about the second, third, and fourth lines, I’ve decided to take a look at the forwards who will be available to the Predators, their playing styles, and how they might fit together.

The Players and the Data:

It’s important to consider that although a fair number of games from 2015-16 and 2016-17 are included in this, the smaller and/or less current the sample size for the players—especially developing players like Kevin Fiala or players like Viktor Arvidsson who received a promotion in icetime and responsibility during the time covered—the less complete these profiles will be. This probably isn’t the ceiling for our younger skill players.

(The other player really noticeably affected here is P.K. Subban, who has under 250 minutes logged and looks surprisingly average in them. My best guess is that Subban playing in Therrien’s dump-and-chase system for the Canadiens, then playing injured for the Predators, are dragging his outputs down. Depending on how Stimson’s model handles data for players who change teams, that might play a role too.)

A player’s contributions are also weighted slightly by how he performed relative to his team. Nick Bonino, for example, is being considered partly in comparison to his former Penguins teammates, and might (or might not) do better once he’s not being compared to a pair of first-ballot hall-of-famers ahead of him in the depth chart.

Tony Bitetto was also a surprise in a small sample size.

(The numbers here are percentiles and are only for players at that position—so defensemen are only being compared to defensemen, and forwards to other forwards, instead of all players to all other players.)

A few standouts among the forwards include Craig Smith, who looks surprisingly well-rounded, Ryan Johansen, who delivers on his potential, and our own crown prince Filip Forsberg. If Mike Fisher were returning next year, he would have fit between Mattias Ekholm and Ryan Ellis on the overall depth chart. That’s a good place to be.

The Centers:

It’s not the best.

Instead, Fisher’s retirement leaves us with Johansen (playmaker), Bonino (balanced), Calle Järnkrok (dependent), and Colton Sissons (dependent) as the available centers. It’s extremely unfortunate for the Predators that Johansen is the only definite playmaker on the roster, although the type is rare enough that the team does still have their fair share.

It’s possible that Sissons, who has the third-fewest minutes tracked of any player in this sample, might be better than he looks here. But it could also be possible that Ryan Stimson and his Passing Project volunteers got a fairly representative sample of what Sissons brings.

Järnkrok and Sissons are weaker than the average forward at driving play forward into the offensive zone (as is Austin Watson, which may be a big part of why that line struggled last year). Fortunately, this is one way the Predators’ extraordinary defense corps may be able to help drive the offense without personally having to go in below the goal line. Ekholm and Roman Josi excel at helping generate controlled entries, Ellis is also very solid, and Subban is much better than his very limited sample shows him to be.

More concerning is that Johansen—who doesn’t shoot often—is actually the only Predators center in the top half of the league at contributing individual shot attempts. Bonino, Järnkrok, and Sissons will all need even more support from their wingers to make goals happen.

The Wingers:

I transcribed McLeod’s metrics first just to get it over with. Then I cleansed my eyes and my color palette by doing Smith’s.

I’m honestly mystified by Craig Smith, who works out as the best all-around Predators player at generating offense. We did notice during the 2015-16 playoffs that losing him to injury did not go well for the Predators, who lost all four games they had to play in without Smith (counting the game where he was injured during his second shift). He, like the late lamented Colin Wilson, seems to do a lot of things well.

One interesting thing is that none of the wingers, not even Cody McLeod, are terrible at making dangerous plays (which Stimson defines as a pass, or the shot it leads to, starting from below the goal line or crossing the slot). This is a nice surprise given the Predators’ perceived struggles with net-front play. Most of the wingers also shoot a lot relative to league average, which makes it easier to find shooting wingers to fill in around the team’s passing centers.

Putting It All Together:

In his article, Stimson analyzes the on-ice expected goals percentage that results from lines made up of each of the twenty combinations of forward types.

Sadly, the Predators do not have multiple playmakers (though that’d be another reason to hope for an Alex Galchenyuk trade). This rules out any of Stimson’s suggested best options that depend on having more than one. With the roster the Predators have, only sixteen of the twenty total combinations even could be possible, and some of them would be very silly.

Stimson also found that putting two dependent forwards, or a balanced and a dependent forward, on the same line is something that will usually end with that line getting outscored. The Predators don’t have the players they’d need to avoid that. Of the twelve forwards on the 2017-18 roster who have games tracked, one is a playmaker, three are shooters, three are balanced, and five are dependent.

I made the following assumptions when figuring out which players to arrange:

  • McLeod will not be playing. I’m aware he probably will in reality, but I’m playing with roster optimization here.
  • One of the three young forwards on the roster who doesn’t appear in Stimson’s data (Freddy Gaudreau, Vladislav Kamenev, or Pontus Åberg) will play instead of McLeod. I went with Åberg.
  • Miikka Salomäki will be playing. He might be replaced on the fourth line by Gaudreau (or McLeod) in reality, but I decided to use as few players I had no data for as possible.

Working with what we’ve got, I’ve put together the following:

Possible Lineup

Fiala (Balanced) Johansen (Playmaker) Forsberg (Shooter) 55%
Arvidsson (Shooter) Bonino (Balanced) Hartnell (Balanced) 50.90%
Åberg (???) Järnkrok (Dependent) Smith (Shooter) 44.7–52.1%
Watson (Dependent) Sissons (Dependent) Salomäki (Dependent) 41.80%

In addition to trying to find the lines with the best xG%s, I’ve done my best to group players with complementary strengths and weaknesses.

The First Line: Fiala–Johansen–Forsberg

JOFFI? FiJoFo? Fee Fi Fo Fum? We might have to go back to the simpler days of referring to lines with player numbers.

The JOFA line is a joy to watch, but it makes this team top-heavy. We’ve seen what happens when it can be neutralized, and Stimson’s work suggests that it’s not an efficient use of the Predators’ best players.

We know that Forsberg struggles away from a high-quality playmaking center, because we’ve seen what happens when he doesn’t have one. He was the obvious choice of Shooter to leave with Johansen.

Johansen’s hesitation to shoot is balanced by Fiala and especially Forsberg’s frequent shots. All three of these forwards work very well around the net, and Fiala’s current weaknesses are balanced well by his more experienced linemates.

Although Stimson doesn’t go into defense at all, this line should also be very good defensively.

The Second Line: Arvidsson–Bonino–Hartnell

Viktor “Shot On Goal” Arvidsson is probably not going to start passing too much anytime soon.

Scott Hartnell on the second line? I’m not sure myself. This line could struggle a little bit to create dangerous plays right around the net, but with Arvidsson keeping the goalie busy and Hartnell ready for rebounds I think it has a decent chance of working out. Bonino doesn’t shoot but is solid otherwise; Hartnell looks solid across the board. Arvidsson has higher highs and lower lows than the other two.

The Third Line: Åberg–Järnkrok–Smith

The next man up is still finding his stepstool.

Pontus Åberg, international man of mystery. What is he? I assume he isn’t a Playmaker, but he could be anything from a Shooter to another Dependent forward. He had an excellent year in Milwaukee, so I optimistically penciled him in as a Balanced player—one who doesn’t rely on his linemates to create—for now. If he does turn out to be a Shooter, a lot of this would need to be re-juggled; getting the chance to find out would be great.

If Åberg is even just exactly league-average in each of these metrics, this third line would be a decent second line. Smith is a frequent shooter who probably won’t completely miss the net as often this year as he did last year. He also balances out Järnkrok’s other weaknesses. This line’s biggest lack of strength could be in shot volume, but since Åberg averaged almost three shots on goal per game in the AHL last year he might be able to contribute as another shooting winger.

The Fourth Line: Sigh Watson–Sissons–Salomäki

May I repeat: sigh.

After the first three lines, this one’s not a fun drop-off. Someone on this line definitely needs to shoot more often. Salomäki, like the centers, tends to pass for a shot instead of just taking the shot. He’s also been by far the best player on this line at getting into the offensive zone. If someone else replaces Salomäki, they need to be able to help make controlled entries. We can’t expect the defense corps to do all the work.

Sissons’s relatively good rate here of making passes that lead directly to a shot on goal reminds me of his performance this past spring after Johansen’s injury, when the most notable thing about his playoff performance was his very high rate of giving primary assists on goals. If he can continue to develop that aspect of his game, while shooting a little more, things get less grim.


Overall, this lineup has an okay top line (which is probably better than it seems), a good second line, what could be a great third line, and a fourth line. Like I said, there’s a lot of guesswork, and some of it is math-driven. I think it’s an interesting starting point.

What would you change?