The "Turtle" Tendency
and why moving away from the Provorov-Gudbranson pairing may be in Pascal Vincent's best interest
The Blue Jackets have had quite a terrible run of performance in the month of November. While Pascal Vincent has noted that the underlying process has been improving, he also noted that timely mistakes were killing his team’s chances at winning.
I wrote about that recently, but after reviewing the tape of the game vs Carolina, where the Blue Jackets blew a 2-0 lead in the final 10 minutes of the 3rd period, a more startling pattern began to emerge.
Before we dive to deep into that, it’s important to note that there is a widely known pattern in hockey games. The team that is currently leading in a game has a tendency to give up more shots. These phenomenon are called score effects, and have been examined rigorously since the dawn of analytics.
Now, you can find websites and algorithms that seek to clean and adjust on-ice performance taking these metrics into account. NaturalStatTrick has a filter for their data that includes venue and score adjusted data, HockeyViz takes these effects into account when analyzing sG and career shot heatmaps and Evolving-Hockey certainly takes these into account in their RAPM models.
Micah Blake McCurdy, of HockeyViz.com, concluded that score effects are primarily driven by the leading team sitting back and forming a “defensive shell” rather than from the trailing team benefiting from aggressive play.
Cam Charron, then of The Athletic but previously as an analyst for the Toronto Maple Leafs and now an analyst for the Pittsburgh Penguins, wrote a detailed piece with some film analysis of some blown leads. His conclusion was that the Penguins were blowing leads because they decreased their forechecking pressure which allowed good teams to make organized neutral zone attacks.
Over the past couple of interviews, Pascal Vincent has mentioned that this “defensive shell” isn’t an intended part of Blue Jackets hockey, but is driven by players being passive as an attempt to not be the one who loses the game.
While it would be admirable for Pascal Vincent to coach these habits out of the players, that may be difficult because the primary culprits for this issue are NHL veterans and their playtime in late-game leading situations is well within his control.
The Ivan Provorov-Erik Gudbranson Pairing
Ivan Provorov and Erik Gudbranson haven’t really been paired together all that much at 5v5 this season. However, they are deployed as the top penalty-killers and the pairing has emerged late in periods or games when the Blue Jackets are trying to defend a lead.
We can see the first of that information here. Zach Werenski, Erik Gudbranson and Ivan Provorov all see a bump in available ice-time when CBJ have a 1 goal lead. Damon Severson, Adam Boqvist, David Jiricek and Jake Bean all generally see their workload decrease.
Erik Gudbranson and Ivan Provorov have played a total of 109 minutes together across all-situations across 22 games this season, according to NaturalStatTrick.com.
That would be the 5th most played D pair for the Columbus Blue Jackets(behind the Gudbranson-Bean, Severson-Werenski, Provorov-Jiricek and Severson-Provorov pairs), primarily a result of the two getting a lion’s-share of available penalty-kill minutes with 62 (the next closest pair being Werenski-Severson with 18:40 on the PK).
Still, at even-strength (which includes empty-net situations) the two have played 46:05. When trailing, 9-44 have played 11:21 and when leading the two have played 24:47. Pascal Vincent uses these two in high leverage defensive situations. The pair is designed to protect leads.
In their 18:44 of 5v5 icetime with the Blue Jackets in the lead, the pair have controlled 20.83 CF%, have been outscored 0-2 and control 19.23% of scoring chances (5 scoring chances for to 21 against).
In addition to protecting leads at 5v5, the pair also leads the team in time played against an Empty Net where their results are among the worst in the NHL. They have 6:44 played against an empty net, the next closest pairs being Gudbranson-Severson with 2:33 and Severson Werenski with 1:29. In that time, 9-44 have 1 goal for and 3 goals against.
It is worth mentioning that these are extremely small minute samples. This performance may not be repeatable across large measures and could regress to the mean.
Still, among all NHL pairs who played against empty nets, Provorov-Gudbranson ranks 4th in ice-time. Among all NHL pairs who have played against an empty net, Provorov-Gudbranson is one of two to have been scored on 3 times (David Savard-Mike Matheson being the other).
The NHL leading pair in ice-time against empty nets is Devon Toews-Cale Makar with 9:30, who have 6 goals for and 0 goals against. The other two pairs with more time than 9-44 are Derek Forbort-Brandon Carlo and Seth Jones Alex Vlasic and each pair has scored 2 goals and conceded 1.
This pair is the primary cause for “turtling” and I would strongly advise Pascal Vincent to start breaking in new blood in these high-leverage situations, but that may not be possible with the roster as currently arranged.
The “Deterioration” of Erik Gudbranson
Ivan Provorov may be catching some unnecessary heat. By simply being assocaited with Erik Gudbranson in high-leverage situations, he is being set up to fail.
Over the last 7 games at even-strength, Erik Gudbranson has been on the ice for 3 GF and 12 GA. That’s a 20 GF% across nearly 121 minutes (rocking a 36.47 CF% all the while). While he started the year as a solid shot-suppressor and crease-clearer, he has quickly reverted to the same performance he turned in last year.