Quoting: MitchMarnerElite
There was an article written about why Ceci & Rielly were horrible when paired together. The issue was, Rielly’s inability to prevent scoring chances & stop the offensive zone cycle sunk them. Obviously it doesn’t help with Ceci’s inability to move the puck, but he has actually been quite good at preventing scoring chances.
The article looks at statistics and tries to fit ideas to what the statistics are telling the author. I think that it gets things very much wrong because it is not analyzing games to see if there are alternative reasons that explain why the statistics say what they say. When an author uses statistics to show that Ceci was great defensively this season, but had been very poor previously to this season chances are it is going off the rails. I don't think that Ceci was as good defensively as the article claims (he was still good) or that Ceci was close to as bad offensively as the article claims. But the article uses statistics to show that Rielly and Ceci should never be played together. Certainly the Leafs have all those statistics and a whole lot more. They have Rielly and Ceci together in camp so far.
I am not a big fan of Ceci. I think that he has done what the team asked him to do quite admirably, but he doesn't bring the physicality or the meanness that I want in a defender who plays the role that Ceci was given. Now that is a personal bias and the team may or may not care about a defender in that role bringing those other elements. I am not looking for physicality and meanness in every Leafs' D, but do think that 2, or perhaps 3, D are needed in that fashion similar to what Colorado has right now.
I will use CF/60 and CA/60 for no other reason than CF and CA were one of the metrics that the article used. The other metrics would be similar.
For the first 23 games of the season (those under Babcock) Ceci had a CF/60 of 62.6 while playing with Rielly and a CF/60 of 68.4 when playing without Rielly (those numbers would rank him 10th and 1st among all D who played at least 500 minutes TOI this season). His CA/60 was 57.5 while playing with Rielly and 60.6 without Rielly. After the coaching change he was 56.1 CF/60 and 63.9 CA/60 while playing with Rielly and 48.6 CF/60 and 50.6 CA/60 without Rielly.
That is a massive change, both when playing with Rielly and even more so when playing without Rielly, that occurred with the coaching change, but if you watch the games both from early in the season and from later on you can see that there was a massive change in how Ceci was deployed. Before, surprisingly as it was with Babcock, Ceci played a normal shift. After the change, Ceci's deployment changed to a very defensive role. This had a lot to do with trying to get Barrie as much of an offensive role as possible. This meant that Ceci had a tendency to still play with Rielly when the Leafs were either protecting a lead or when they were going up against a particularly strong offensive opponent (ie Tampa Bay in the final game of the season). But when not partnered with Rielly, the goal was very much to get Barrie out in offensive situations whenever and however possible. What I mean by this is that Ceci could be out on the ice in his own zone early in a shift and when the Leafs' get the puck and start heading up the ice Ceci goes off and Barrie goes on. This inflates Barrie's offensive advanced stats and deflates Ceci's offensive advanced stats. And this very much happened after the coaching change. It was a change that makes sense, but it also makes the advanced stats misleading.
Before the coaching change Ceci and Barrie were as likely to be out on the ice as the other if the Leafs' had the lead or were trailing or were tied. All three were within 4% and two of the three were within 1% which is completely extraordinary - but not a good indication of coaching to the players' strengths. The CF/60 for those three game states were also very close for Barrie and Ceci.
After the coaching change Barrie played 35% more when the Leafs' were trailing, and Ceci played almost as much more than Barrie when the Leafs' had the lead. But the CF/60 were very interesting. When the Leafs had the lead Barrie's CF/60 was 52. Ceci's CF/60 was also 52. When the game was tied Barrie's CF/60 was 69, Ceci's was 53. When the Leafs' were trailing Barrie's CF/60 was 70 and Ceci's was 45. From even, to Barrie being up by 16 to Barrie being up by 25. Ceci didn't become worse offensively when the Leafs' were trailing - the deployment changed significantly.
I am using statistics to make my point, but I also do so by re-watching and analyzing player deployment and how they influence advanced statistics. In the case of Ceci and Barrie, I did so a couple months ago, but I try to do so with as many situations as I can that I come across where the advanced stats community is at odds with coaching staff of a team. My assessment of Ceci. He is a top-4 D and will be on whatever team is on next year. He is pretty good, but not great defensively. He does things in the defensive zone the Leafs' lack (which is why Sandin looked better while playing with Ceci than he did with other D like Barrie or Liljegren). He is not a good offensive D, but the statistical appearance that he is dragging offensive numbers down is due to deployment and timing of on-the-fly shift changes (which was visible when watching games if you were watching for this issue and recording when it happens) as was outlined (not specific to Ceci, but as a method in general) by Tyler Dellow in The Athletic several years ago.