Tag Archives: hockey

ShiftChart 101 (Part 1): a visual model of a hockey game

As one of the creators of ShiftChart.com, I thought it might be helpful if I shared a few notes on how I use the tool, which, I realize can be a little overwhelming to the uninitiated.  In this post, I’ll focus on the main shift chart itself and only its oldest most basic features.

If you’re completely new to the site, I recommend hitting the Play button:

Play button

Press the Play button. Sit back and watch.

Colored rectangles scoot about, in sync with a skipping blue vertical line.  I’ll call this “Play mode”.  Let it run a minute or so and allow your eyes and mind to acclimate to all the activity.  Hit the Pause button at any time to stop (“Pause mode”).

[UPDATE 6/17/14 – 80-second video added below for Game 5 (Cup clinching), LA Kings vs. NY Rangers on 6/13/14]

What is it doing?

The site is simulating an NHL game from opening faceoff, except here, we’ve abstracted away all human-form elements and replaced them with rows of rectangles.  Further detail:

  • The blue vertical “clock line” hops in 15-second game-time intervals.
  • The on-ice skaters for each team slide to the top of the chart replacing those skaters who have gone to the bench.
  • Each “shift rectangle” represents a player’s shift:
    • Left edge = when the player got on the ice.
    • Right edge = marking when that player left the ice.

Why do this?

That question deserves its own entry, but for now, I will just say that there is information about a game that is more easily gleaned from this perspective.  In particular, the identities of the on-ice players are difficult to keep track of: the camera follows the puck and so player substitutions often occur off-screen.

A few things should become apparent in Play mode, which might be surprising if you don’t follow hockey:

  1. Players change very frequently.  In even strength, players are rarely on ice for more than a minute.  Typical NHL shifts range from 30-50 seconds.
  2. Star forwards only play about 1/3 of a game; star defensemen play more, but usually less than 1/2 the game.  Very different personnel dynamics operate here compared to sports like soccer or basketball!
  3. A team’s on-ice forwards typically substitute at the same time. Separately, a team’s two defensemen substitute simultaneously.  Later, these same sets of players return to the ice together.  This is because these sets of players are constructed units that practice together.  Each team is composed of and practices as 4 “lines” of three forwards and 3 pairs of defensemen (ignoring healthy scratches).
  4. The same players from each team face one another frequently.  This is standard coaching strategy and is referred to as a “matchup“, a topic I go into further detail here.  Matchups are front-and-center in the playoffs, with coaches sending out their “shut down” players to stop the opponents’ top scoring threats.  In the Kings-Blackhawks series, the Toews line often faced the Kopitar line in key moments of games.  A match up is usually one of:
    • forward line vs. forward line
    • forward line vs. forward line + D-pair
    • forward line vs. D-pair.

Analysis of specific game instances

With the game in Pause mode, the clock line can be drag-and-dropped to specific moments in a game.  I always stop at goals (clicking on a goal line will jump the clock line to it) to see if there are contributing factors such as advantageous matchups, poor substitutions, or long shifts.  In the deciding Game 6 of the Canadiens vs. Rangers series, the Rangers’ 4th line maintained a cycle deep in the Canadiens’ zone, leaving three Canadiens players stranded on ice for 1:31 each, an eternity in playoff hockey.  The goal proved to be the game winner.

On Dominic Moore's goal for the Rangers, three Canadiens were trapped on the ice for 1:31 each.

On Dominic Moore’s goal for the Rangers, three Canadiens (marked in orange) were trapped on the ice for 1:31 each during the 2nd period.

I also look for instances when the usual line combinations are tinkered with.  Chicago coach Joel Quenneville is known to do this when behind in games, but in Game 7 of Blackhawks vs. Kings, it was the Kings’ coach, Darryl Sutter, who made an adjustment to his forward lines after quickly going down 2-0 in the first period.  In particular, he split up “That 70s line” of Pearson-Carter-Toffoli which had been so instrumental earlier in the series.  The newly-assembled line of Carter-Toffoli-Brown scored a crucial goal shortly after, getting the Kings back within a goal before the 1st period ended.

Coach Sutter splits up "that 70s line" mid-period after going down 2-0 early in game 7 to Blackhawks.

Kings’ coach Sutter splits up “That 70s line” (highlighted players) mid-period after going down 2-0 early in game 7 to Blackhawks. The newly-assembled line of Carter-Toffoli-Brown pays off with a goal late in the 1st period.

Hockey is a complicated game and I won’t claim that this was the turning point (Kings would go on to win in overtime), but certainly the game changed character from one that had the beginnings of a Blackhawks runaway.

Off-ice players

So far, I’ve discussed the on-ice player composition, but the off-ice personnel is often the focus of my attention—it illuminates which players the coach is not utilizing. Here’s a snapshot from the 3rd period of the same Blackhawks-Kings Game 7 with 13:36 remaining in the game:

Kings-Blackhawks Game 7

3rd period of Kings-Blackhawks Game 7.  Off-ice players are organized by position and when their next shift is.

First, note our organization of the off-ice skaters for each team: forwards are grouped together, then defensemen.  This makes it easier to see the rotation of forward lines and defensive pairs, which are managed by separate coaches on the bench.  Then, within each group (forwards or defensemen), the players are sorted in order of next-on-ice.  With 13:36 left in the game, the Blackhawks have the Kruger-Sharp-Smith line up next for forwards and the Seabrook-Oduya for defense.  The Kings have Richards-Carter-Pearson up next with Greene and Martinez.  Below is a zoom in of the same screenshot, with some players highlighted:

Zoom in of Kings-Blackhawks Game 7, 3rd period and OT.

We can easily discern that #52 Bollig is done for the night for the Blackhawks (in fact, Quenneville did not use him after the 2nd period), and #13 Clifford is done for the Kings.  Neither player steps on the ice again.  In the row above each of these, we see that #23 Versteeg and #22 Lewis will get one shift each for the rest of the period (Versteeg goes on to have two shifts in OT).  #26 Handzus, the OT hero of game 5, also sees limited ice for the rest of the game.

On the defensemen side of the roster, we can see the limited usage of each team’s 3rd defensive pairing. For the Blackhawks, this is Leddy-Rozsival, for the Kings this is Greene-Martinez (the latter is the eventual OT hero).  Both pairs are frequently skipped over in rotation, making the D-substitution anything but a revolving door of players.

Conceptually, the off-ice portion of the chart illustrates team depth, or at least, the coach’s perception of his team’s depth, a concept I go into further detail here.  The “shortening of the bench” is understandable: no one wants to be the coach that has to answer questions of why he had his weakest players on ice when the winning goal was scored against his team (I’m reminded of Brian Burke’s dream about how 19-year old Seth Jones would cost the US a medal).  This is, in fact, what happened in the Anaheim Ducks’ elimination of the Dallas Stars in OT of the first round.  Dallas coach Lindy Ruff tried to play his 4th line (two rookies and a previously scratched Erik Cole) and got burned:

Ducks score OT winner against Stars' 4th line.

Ducks score OT winner against Stars’ 4th line.

The image of Ruff tearing up his bench notes is still etched in my mind.  Despite not being a Stars fan, it’s tough to watch..

Pittsburgh Penguins coach, Dan Bylsma would not make the same mistake.  In Game 7 against the New York Rangers, down 2-1 with just under 10 minutes to go, Bylsma essentially trimmed his bench to his top two forward lines and defensive pairs (highlighted players get little ice time):

Penguins' Coach Bylsma shortens bench in failed comeback effort against Rangers.

Penguins’ Coach Bylsma shortens the bench in a failed comeback effort against Rangers in Game 7. Orange players are those that saw little ice time in the final 10 minutes.

Although I have seen stats-based articles stating that Penguins’ depth was not an issue, it’s hard to argue from the above that at least Bylsma himself had little confidence in his bottom six forwards to produce the tying goal in the waning moments.  “Throw physic to the dogs; I’ll none of it.”

The problem, of course, is that Sidney Crosby, the world’s best player, is not the world’s best player when forced to skate every other shift for 10 minutes at the end of the 3rd period of a Game 7.  Okay, he still might be the best player, but certainly his effectiveness is diminished.  As rigorously as elite players like Crosby or P.K. Subban train in the off season, these are not video game characters and the advantage, when protecting a lead, often goes to the fresher legs.  I’m hardly alone in believing in team depth as critical to the Rangers’ playoff success.

Closing thoughts

In summary, a shift chart shows the coaches’ execution of player usage.  One can glean from this line combinations, matchups, mid-game adjustments, and team depth.  In my next post, I’ll discuss some of the other features we’ve built into the site.

Thanks for reading.  Would love to hear your thoughts.  Now let’s watch some hockey!