Blasphemy, isn’t it? Who doesn’t love the JOFA line? The truth is we all love that line. Don’t try me for heresy yet, but lets take a look at why splitting up JOFA might be a good idea for the Nashville Predators.
Player Clustering
As the world of hockey analytics expands, the collective community will draw new and refine conclusions as we get better at collecting, analyzing and interpreting the data. One of the more recent developments devised comes from Ryan Stimson. Using data from Corey Sznajder’s passing project which is tremendous, Stimson has identified clusters of players by playing style.
The four playing styles identified for forwards are: playmaker, shooter, balanced and dependent. Each playing style has a collective expected amount of goals as seen below.
There are no considerations made about which players are better or worse in their particular roles, which is an admitted weakness of using the same simple expected goals rate (xG) for each player who fits that particular category. To use a Predators example, Filip Forsberg and Craig Smith are both categorized as shooters, but it’s evident not each will have an expected goals rate of 52%. Forsberg is a far more dangerous player than Craig Smith. Nonetheless, it still gives us an idea of which playing styles lead to better outcomes. The less dependent players a team has, the better it will perform throughout the season.
Break Up The Top Line?
Correspondingly, Dawson Sprigings presented at the RIT Hockey Analytics Conference last fall, and concluded that teams would be better off spreading their top forwards among separate lines. The pertinent example given by Sprigings was the first line for the Boston Bruins. Putting Marchand-Bergeron-Pastrnak together on one line left a serious lack of scoring down the lineup.
Alex Novet made a similar conclusion about hockey being a strong link game. The conclusion here is simple: if you put strong players on different lines, it makes it that much harder on the opposition. Most NHL teams are able to scrap together something that resembles a top pairing on defense, but it’s much harder to assemble a strong Top-4. Other than Nashville, Calgary, Carolina and Anaheim, most teams have significant weaknesses in their Top-4, leaving an area to exploit by opposing coaches.
Optimal Line
Sean Tierney is one of my favorite analysts. He writes for The Athletic and for Hockey Graphs. His work is what has spurred me to get better at using visuals. Throughout the summer he’s produced graphics on each team’s optimal lines based on the data from Sprigings and Stimson. Assuming Nick Bonino is healthy, Tierney’s chart of optimal lines has the JOFA line up top with a second line of Fiala-Bonino-Smith.
The bottom six is pretty rough, with five dependent players and only one balanced player in Scott Hartnell. Use caution with some of these bottom six guys, because a lack of data may be underselling them as dependent players, when is reality they could become something more once they log more games and more we gather more data. Although the bottom six looks pretty brutal, with expected goal rates below 45% for both lines, in reality they may be able to fair marginally better.
Can The Preds Do Better?
Is the above lineup optimal? The most optimal single line any team could ice would be a mix of three playmakers, with an expected goal rate of 58.9%. Being the rarest type of player, these guys are at a premium throughout the league. The Predators only have one such player in Ryan Johansen, who last year had an expected goal rate of 54.8%, near the established norm for playmakers. The top single line the Preds could produce is indeed the JOFA line with a 55.1% expected goal rate (two shooters and a playmaker).
If we flip one player from Tierney’s graphic from the second line to the first, the Preds theoretically could ice a top six with an expected goal rate of 55.0% and 54.7%, far better than 55.1% and 50.9% (barring a drastic TOI difference between line one and line two). Now the question becomes, which player should make that transition?
The Case For Fiala
If we are going to make a move between wingers on the first and second line, we have four players to work with. Either Forsberg or Arvidsson must be swapped with Fiala or Smith. An argument could be made for Craig Smith, but he fits the shooter prototype, which wouldn’t do anything for expected goals. The clear answer then is move Fiala to line one. Let’s make the case by dividing up each aspect of offensive play and come up with an optimal deployment.
Playing Styles
Ryan Johansen, being the teams’ one playmaker should remain on the top line for the foreseeable future. Even if the Preds top center prospect Vladislav Kamenev reaches his ceiling, he’s likely a good second line center, not a first line talent. RyJo will be unchallenged as the 1C. Newly acquired Nick Bonino served as the third line center in Pittsburgh, but will likely play line two in Nashville once he returns to health.
Using another visual from Stimson, we can compare the Predators three most productive wingers (not counting Craig Smith). In terms of total shot contributions, shot volume and transition play, Forsberg contributes at an elite rate. His passing is underrated, producing at a near elite rate in primary shot assists and build up play. Arvidsson represented in red is similar to Forsberg in transition play, shot volume and shot contributions, while not as dangerous in passing results. Kevin Fiala excels in dangerous shot contributions, as seen in purple, and does well in shot volume and transition play.
Once the puck is in the zone, Laviolette can rely on both Forsberg and Arvidsson to get the puck to the net, with Forsberg or Fiala providing some dangerous shot contributions.
What goes unmentioned is just how good of a playmaker Filip Forsberg can be. Once the puck is in the zone, we can expect adept passing skills from Forsberg, making him arguably the second best playmaking forward on the team. Although shot generation will remain the strongest part of his game, he has the ability to set up his teammates as well or better than anyone not named Ryan Johansen.
Evaluating this trio by the eye test: Forsberg is an all-around talent with a great shot and underrated passing ability. Arvidsson is a wolverine on the forecheck with a shot volume that rivals anyone in the league (his 19.02 individual shot attempts generated per sixty minutes ranked 7th among all forwards, ahead of names like Tarasenko, Pastrnak, Matthews and Pacioretty). Fiala on the other hand can create down low. He’s not afraid to get to the front of the net.
Fiala can score on his own and would likely be able to set up scoring teammates based on his dangerous shot contributions. Three years removed from his draft year, Fiala just turned 21 over the summer, and he’s now entering the wheelhouse for forwards where his production should really start to take off.
Zone Entries
Carolina Hurricanes director of analytics Eric Tulsky pointed out a few years ago how much more effective carrying the puck into the zone is compared to dump-ins. His league-wide data from 2013 stated that teams generated 0.62 shots per carry-in and only 0.28 shots per possession when they dumped the puck in.
Although these numbers are from three seasons ago, there is no reason to expect these rates to have changed much. If we use 0.62 shots per carry-in as league average, the Preds had six forwards last year generate more offense per carry-in than league average (essentially everyone above Mike Fisher in the visual below).
What is particularly meaningful here however is the rate of carry-ins. Nashville had four forwards on the roster generate controlled entries over 64% of the time: Smith, Fiala, Forsberg and Johansen. The more a team can enter the zone with control, the greater volume of shots they will generate. As hockey continues to evolve with meaningful data we will see teams work to generate more of these types of plays. Eventually neutral zone defense will change to meet this shift, which could lead to yet another evolution. Such is coaching and tactics.
Player Comps
Hero Charts is a great evaluative and comparison tool. If we compare Fiala to Arvidsson then Fiala to Forsberg, it’s becomes evident which player he compares to best.
Starting with Arvidsson, both players scored at similar rates last year, but Arvidsson contributed more to first assists than did Fiala. This is due mostly to Fiala’s two most common linemates last season being Craig Smith who shot just 7.7% and Mike Ribeiro who never shoots. He bounced around the lineup later in the year skating on a line with Mike Fisher and Craig Smith; then one of James Neal and Calle Jarnkrok.
Skating most of the season down the lineup, Fiala was left to produce on his own, which he was able to do. Fiala’s shot generation rates were very comparable to Arvidsson’s. Arvy generated 122 shot attempts per sixty minutes, while Fiala generated around 112.
Comparing Forsberg to Fiala, they are nearly identical possession players with similar goal scoring rates. The only difference is in generating primary assists, which was likely due to having very different linemates, and ice-time, something Fiala has no control over.
Bring in Bonino
The best comparison we can make for Nick Bonino is that he’s Mike Fisher with hopefully a little more offense. Playing center in Pittsburgh and not being Sidney Crosby or Evgeni Malkin means that you get mostly defensive zone starts. Bonino’s possession numbers aren’t anything to write home about, but when taken in context he should be decent (not outstanding) in a second line role. Without being a shooter in his own right, slotting Bonino with Forsberg and Smith gives him two volume shooters to play with; freeing up Fiala to play on line one.
Defense
Fiala’s defensive results went underappreciated. He ended the season with a team best 51.1 shot attempts allowed per sixty minutes, which also gave him the best CorsiRel (how many more or less shots does he allow) among Nashville forwards. There is no need to staple him to the hip of Nick Bonino. He’s providing the evidence that he’s ready for a bigger role on the team, and the data tells us he can be responsible in his own end.
The Psychology of The Thing
In a totally scientific poll, I asked my followers on Twitter which player they thought was a more dangerous offensive player, Filip Forsberg or Viktor Arvidsson. The results came in at 84% Forsberg.
Opposing coaches presumably see Forsberg as the most dangerous forward on the team as well. Taking him away from the team’s best playmaker may seem counter-intuitive, but if you are standing on the opposing bench, when are you sending your top pairing over the boards? Would it be against a Fiala-Johansen-Arvidsson line or a Forsberg-Bonino-Smith line?
Remember 2015-2016?
Two years ago, before Preds fans were treated to a heavy dose of the JOFA line Forsberg played a bulk of his time with Craig Smith. During that campaign the Forsberg-Smith combo dominated possession at a 58.5% shot attempt rate and outscored their opponents at even strength by eight (34-26) with goals for percentage of 56.7%.
Wrapping It Up
Player clustering helps us identify playing styles, along with which player combinations perform best together. Data by Sprigings and Stimson indicates that splitting a teams top forwards creates more scoring. Think of Crosby-Malkin in Pittsburgh or Kane-Toews in Chicago. Coach Q would keep them apart all year long. The only time you’d see them together was on the power-play or in the playoffs if the Blackhawks were down. It’s the same discussion happening in Edmonton right now between McDavid and Draisaitl.
Optimizing these lines for the Predators indicates that removing a shooter off the JOFA line and replacing him with a balanced player like Fiala does not significantly affect the scoring ability of the first line. But it does significantly improve the scoring prowess of the second line.
Amid their top forward group (Forsberg-Arvidsson, Fiala and Smith) the Predators have four players who can carry the puck into the zone at high rates. Once in the zone both Johansen and Forsberg have demonstrated the ability to be good at distributing the puck to their teammates. Forsberg and Arvidsson have shot rates among the highest in the league. All three of Johansen, Forsberg and Fiala can create down low and unload dangerous passes (from below the goal line or across the royal road) to waiting linemates. Fiala can compliment Johansen and Arvidsson without needing the defensive help of a guy like Bonino.
Moving Forsberg to line two would let he and Smith rekindle the magic from two seasons ago. Forsberg is the Preds most productive 5-on-5 player in primary points (goals and first assists) per sixty minutes at 1.80. Pairing him with Bonino, who has some creativity on his own should elevate Bonino’s point totals into the 40’s. Don’t expect the Crosby effect here, but Forsberg’s relative rates in terms of shots, goals and scoring chances are all positive. Craig Smith is tired of not scoring, because we all know he wants to. Maybe it’s Forsberg who can help make that happen.
Splitting up JOFA and giving first line minutes to Fiala would give the Predators a more balanced and more dangerous top six. This is the year for Fiala to make that leap. He just turned 21 and is among a group of players in the same draft class who’s teams are waiting for them to emerge as true talents. Many have said that Fiala could be the next breakout player in Nashville, and there would be no better way to mark his arrival than first line minutes with Johansen and Arvidsson.