clock menu more-arrow no yes

Filed under:

Dion Phaneuf and Cody Ceci Must be Split Up

New, comments

Guy Boucher’s shut down pairing is failing on historical proportions

Colorado Avalanche v Ottawa Senators Photo by Jana Chytilova/Freestyle Photography/Getty Images

Ever since Guy Boucher took over as coach last season, the Ottawa Senators’ top two defence pairings have remained relatively the same. Erik Karlsson handles the big minutes playing next to someone else, and Dion Phaneuf is paired with Cody Ceci.

Unlike the forward corps, which Boucher likes to blend by the shift, the pairings on defence have consistently remained intact. Since the start of 2016-17, Phaneuf and Ceci have been the NHL’s 15th most used pairing at 5v5, 14th in all situations. Their usage has remained consistent as well, with their 2017-18 zone start ratio (comparison of offensive and defensive zone draws) of 36.67% ranking 7th lowest in the league (minimum 100 minutes).

I’ve argued in past that the pairing should be split up, although with the results they’ve currently been producing, the decision is long overdue.

The Stats

Looking at the surface, their results don’t seem to promising. Their raw Corsi% is well below 50 at 38.1%, and their expected goals percentage (weighted by shot location) is even worse at 28.35%. This obviously isn’t helped by their heavy amount of defensive usage, but when adjusting for those factors, their results are still in the basement. Their Corsi% improves to 40.04% and expected goals is up slightly to 29.55%, although those are still far from results you would expect a pairing that’s regularly playing 20 minutes a night. All further stats will be using these same adjustments.

The weak spot of the pairing is their abilities in the defensive zone. Expected Fenwick Save Percentage (xFSv%) is a useful stat for measuring shot suppression, as it weighs shots based on location, angle and type, then outputs the expected save percentage that the player or pairing would experience given what they’ve allowed when on the ice. For reference, the league average xFSv% is roughly 94%.

Phaneuf and Ceci together have experienced a 91.96% — last in the league. Using all available data going back to 2007-08, they rank 6th last amongst all pairings with 100+ minutes.

Looking at a shot location visualization from HockeyViz.com, you can see why this season’s been extremely rough for the duo. Red spots represent more shots from that area compared to league average. Their net-front presence couldn’t be more porous.

Last season wasn’t as bad (94.23%) and ranked right at team average, although their lack of offence brought them down to a 45.48% xGF%. This season’s results alone should be enough for Boucher to switch things up.

The Changes

Although it doesn’t look like Phaneuf and Ceci will be broken up any time soon, that doesn’t mean we can’t create an ideal scenario. I’ve used chemistry scores to create lineups in the past, and we’ll look at some early season data from 2017-18. In short, it measures the efficiency of a pairing by looking at their Corsi% is like together compared to when they’re apart.

Here’s a visualization of Ottawa’s pairings and their chemistry, where each box is coloured by chemistry score and sized by time on ice. An interactive version can be found here.

Looking for the darker blues on the chart with larger sample sizes, three pairs jump out to make an ideal combo:

Claesson - Karlsson
Phaneuf - Wideman
Borowiecki - Ceci

This gives each pairing a positive chemistry score, although because it’s early in the season, it’s worth being cautious with the sample sizes (Boro and Ceci have only played 30 minutes together). This would eliminate the Phaneuf-Ceci pairing, provide Karlsson with a suitable partner, and give the lineup slightly more balance.

This is just a suggestion, however, with a multitude of other possibilities. What do you think Guy Boucher should do with these pairings? Leave a comment below.