I had a lot of feedback and good discussion on social media from the last post. After writing this one, I intend to get back into the defensemen and slotting in general again because I think there was a bad implication present last time. Francois Beauchemin is a legitimate top four defenseman and should see those minutes; he just shouldn't be blocking Cam Fowler's ice time, which isn't on him but rather on his coach. Ben Lovejoy is a legitimate third pair defenseman (the good kind) and would better balance this blue line if asked to stabilize those minutes. My next post will get into this some more.
Today, let's get back to the inefficiencies I saw when looking at head coach Bruce Boudreau's ice share preferences based on score. Again, these charts are courtesy of my buddy Micah Blake McCurdy. First up I'll present just the offense, but I'll throw up both the defense and offense down below.
A quick word: I wrote most of this before the Detroit Red Wings game, which both illustrated some of my points (in the extreme) and also tried one of my repeated suggestions. Additionally, any reference to Devante Smith-Pelly is obviously no longer relevant.
I see three inefficiencies with this chart immediately. First, Boudreau is not rolling four lines very well. Second, Nate Thompson plays entirely too many minutes when the score is tied or plus-one. Third, there's a clear lack of "rewarding good players" with many of the lineup decisions. The problems with the third point are more closely related to the first, so let's dive in.
The Ducks broadcasts love to tout Boudreau as a coach who rolls four lines and gets everyone involved, but how he splits his ice share suggests otherwise. This chart verifies what many of us see each game: the coach has a 1A line of Ryan Getzlaf and a 1B line of Ryan Kesler, leaving a 4A line of Thompson and a 4B line of Rickard Rakell. The result is that the top lines play too many minutes and the bottom lines play too few minutes 5v5 in most situations.
This is more or less what Boudreau did last season with Getzlaf and Saku Koivu as well, and the results were identical. The Ducks start off incredibly strong, on the backs of the top players seeing the most ice by a considerable margin, but then fade really hard when those players are simply fatigued. Since Anaheim's offensive success is largely from two players and both are less effective for the playoffs, it is no wonder the team fizzles out given their ice share.
Some of the reliance on the team's top players comes from Boudreau playing them far more than everyone else, however. Asking for secondary scoring on a team that trots out the same two or three guys over and over, with everyone else's legs freezing up on the bench, is a much tougher get. Depth players have a fewer share of ice time to attempt to provide or generate production. This has helped raise the profile of guys like Kyle Palmieri, who has a wicked shot and a hardcore points/60, but has done very little to help out the likes of Emerson Etem, Smith-Pelly, or Patrick Maroon.
Take last night's Detroit game. The first period featured a heavy 5v5 reliance on Getzlaf, Perry, Jakob Silfverberg, Kesler, Hampus Lindholm, and Beauchemin . Not only were they tired by the second, the rest of the team was barely warmed up. After a dominant power play (tons of shots and chances but no payoff) with Rakell and Palmieri playing most of it, Boudreau went back to his well of top players only to see the team get slaughtered in shot attempts against. The Red Wings just flat out beat them to every puck on the ice. Not coincidentally, Detroit coach Mike Babcock played his roster more evenly in the first period, so he had the fresher troops!
This has to change if the Ducks want to have any shot at getting into at least the third round. When the playoffs come around, loading up with the best guys makes a lot of sense. But riding the top guys for an 82-game grind is an excellent way to get beat in the first round. Boudreau needs to better split his ice time among lines to stop loading up so heavily on the top players, giving them needed rest down the final stretch.
Another inefficiency is (or was) Thompson's play when the score is tied or plus-one. I like Thompson, I think he's a perfectly good fourth line specialty player, which is where he should slot. He is fast on the forecheck, he hits everything he can, he battles hard in the corner, he kills penalties, and he absolutely can't score or defend at even strength worth a lick. And you know what? On the fourth line, that isn't the worst sin. But on average, especially early on in the season, he was the third line center.
Not only was he seeing third line minutes early on, but when the Ducks went up by one goal, Thompson's minutes went up. (Note: I think he's become the fourth line center in recent weeks, and I haven't checked to see how that's affected his plus-one play time.) But prior to now, this was a big-time inefficiency because he simply isn't the caliber of playing to be trusted in those minutes.
Some stats tell us why. When the score is tied, the center's on ice shot attempts differential is 49.9% in 237 minutes played. (That is CF% or SAT% on 5v5, for all you abbreviations lovers.) When the Ducks are up by one goal, Thompson's CF/SAT% drops more than most players' numbers already do (because score effects). He's 41.9 CF/SAT% when Anaheim is up one in 147 minutes played.
Boudreau is giving a massive amount of ice time to a guy who is on the ice for only 40% of shot attempts for, which is why Anaheim is basically awful at holding leads on the year. When 60% of the shot attempts taken are against your goalie in those in those minutes played, the chances of all of them staying out of the net is lowered. That isn't an advanced stat, that's just common sense.
So compare Thompson's usage by his coach next to Anaheim's possession metrics, because the trends coming from the bench tend to fallout with the on ice product. The Ducks score adjusted CF/SAT% when tied is 51.6, according to puckon.net. That's good for 11th in the NHL, which isn't bad per se but also isn't even in the top 10. Anaheim's CF/SAT% when up by one is a lowly 43.1, or good for 21st.
The Ducks are the ninth worst team in controlling play when holding a lead. That's ugly. That's inefficient. That's a sign that this team shells way too hard when it should keep pressing and trying to score. Something like that happens when a player like Thompson, who is very good in a limited role, is given priority minutes when up by one.
Why does Thompson's share of ice time increase when up by one? It's 100% because Thompson is good at taking face-offs. To the stats uninformed, which Boudreau is increasingly making the case to be, winning a face-off is akin to starting with possession. It visually appears that way, right? But seven to 10 seconds later, give or take, the statistical effects of a face-off (won or loss) is more or less neutralized. Thompson doesn't help his team retain possession despite winning face-offs pretty often. The head coach would know this, however, if he recognized that winning a face-off is not giving the team meaningful possession of the puck.
Similar to the defense post that seemed to pile on Beauchemin and Lovejoy, I don't want anyone to walk away thinking Thompson is terrible. He's not. He's a serviceable role player in limited minutes. And his reduced possession metrics and the rise in ice share when leading by one are probably tied together somewhat, because trailing teams play their better players in those instances. All players have lower possession numbers when up by one.
Thompson's possession numbers are just lower than normal. Furthermore, he has basically zero offensive upside, so the minutes he does play when up by one aren't threatening to opposing teams. This incentivizes opponents to play even more risky when he's on the ice, because he can't hurt them going the other way.
Now ironically, and going back to the first point I made, the share of ice time Thompson sees when the Ducks are up by one is right where Boudreau's normal third line center should be playing in all situations. This would balance Anaheim's lines, allowing the bottom players a bit more ice time and potentially evening out some of the reliance the team has on Getzlaf and Perry. But as you could guess from the numbers, Thompson isn't the third line center Boudreau should be going with.
The third issue I have with Boudreau's ice share allotment is in its rewards system. We know the coach loves to blend lines. In fact, I am a huge proponent of line blending on the whole, so that's not the issue for me. But the coaches who do it well are those who have a pulse on the roster to such an extent that they know who is going one night and who isn't and are willing to alter course.
Boudreau tends to blend lines on a game-by-game basis and less so in-game. That being the case, it would make sense for the players who are pushing play offensively to continue being blended onto the two lines he tends to overuse. But that's not really the case, which is inefficient and harmful to the team's production overall.
The team's best possession forwards on the year, with a minimum of 300 (zone adjusted) minutes played, are Maroon, Silfverberg, Matt Beleskey, Getzlaf, Palmieri, and Rakell. When the game starts (tied), Getzlaf is consistently given the most prime minutes, which makes sense. But only Beleskey ends up with decent minutes, by virtue of being Kesler's wing, aside from him. Otherwise, the team's best players at either generating shots or suppressing opponent shots, or both, are clumped into the middle and tend to change centers based on the night.
This sort of gets to the heart of the matter of possession metrics over flat out production though, doesn't it? Shot attempts correlate well to goal scoring. The higher the attempts, the higher the goals scored. Conversely, the worse the shot attempts differential, the more goals are scored against. Teams can (and do, all the time) buck this trend in short samples, but eventually the correlation ends up being true. Certainly by the end of an 82-game season, there are enough measurable events to provide adequate predictability going into the playoffs.
Shot attempts are better measures than pure shots on goal because there are simply more events to measure. More data is always better for measurement. Some people look at a list of high end possession players, guys like Justin Williams on the Los Angeles Kings, and see a lack of corresponding production. To these people, he's not as good as other more productive players. Removing his outstanding playoff numbers here, even if he never produced at a rate commiserate his possession stats, ignoring how he gets those numbers shortchanges him as a player.
I mention Williams because he's been written about extensively before and many of you have probably read such pieces. Coming back to the Ducks, Silfverberg is the closest player to a Williams type that the team has. Here's where the Kings are more efficient, however: Williams leads his team in 5v5 play, whereas Silfverberg is fourth. In front of Silfverberg in 5v5 is Getzlaf, Kesler, and Andrew Cogliano, who is just behind Rakell in effectiveness turning play toward offense (so, slightly negative relatively).
While Silfverberg has not produced at a rate even he is pleased with, the team loses a lot of solid two way play without him on the ice. When the game is tied, he is playing third line minutes. When the Ducks have a lead, his ice share jumps up, which is excellent. But when Anaheim is down a goal, his share of time drops to fourth line levels.
It is inefficient to bench one of your best possession players (a guy who makes more shots go on the opposition's net more often) when you desperately need to create offense. Possessing the puck and creating offense go hand-in-hand. Players like Williams or Silfverberg aren't the guys who will produce the most, but they are the guys who put the productive players closet to the net the most consistently. That's needed!
Maroon is the same way: regardless of his lack of speed or production, when he's on the ice, the Ducks tend to be shooting the puck. Like Silfverberg, he's averaging third line minutes when the score is tied. He does see more ice when the team is trailing, which is where much of his ice time is made up overall and is a good thing. But he doesn't see the ice as much when the team is up, which is strange because he's still a positive relative player when that's the case. (Read: his numbers don't drop as hard as most of the roster's does, or basically, he's still pretty good at driving play.)
The Ducks post poor team numbers due to these, and others I'm sure, inefficiencies. The guys who top the team in possession should see favorable minutes in score close situations. "Results" (goals) are what we as fans see as the most obvious "good plays" to cheer for, but coaches are supposed to be the great thinkers driving results. As such, they are supposed to see the details we miss, they are supposed to preach "process."
This is bad process, in my mind. It isn't horrendous process, a la Randy Carlyle-coached systems. But there are enough inefficiencies in the share of ice times doled out to explain perfectly why the Ducks are so average (or below) in terms of possession metrics. These inefficiencies are also why I don't believe in Anaheim as a very serious contender.
By the time the Ducks get to the playoffs, Getzlaf, Kesler, Perry, and Beauchemin will have too many "miles" on those legs. Seeing as all four are among the best on the team, they should be saved for the playoffs, not exhausted to get there. Anaheim will face a similar problem as it has the last two playoff runs: once in, those top players will be easier for teams to neutralize (due to exhaustion), meaning the depth players who haven't been given the best opportunities during the year to grow into defined production roles will suddenly have to be the heroes. Not how I'd coach it...
To end this on a less fully discouraging note, I think the last 20 games can go much differently. Assume I possess Boudreau's brain and make his decisions for him. I would play these lines at some point:
Maroon-Getzlaf-Perry (spotted with Palmieri)
Beleskey/Etem-Rakell-Palmieri (spotted with Perry)
Thompson, Jackman, Bourque, Smith-Pelly, etc. bringing up the rear, in any configuration.
When Beleskey is set to return, I would probably keep Etem in on the fourth line so he can slot up and down. The main thing I would change is rolling the top three lines more regularly, slotting the fourth line in as zone starts allow. The top three players (Getzlaf, Kesler, Perry) need to be playing less overall, and that means more minutes for six other guys, at the very least.
Note: With the Jiri Sekac trade, I would try him with Rakell.
Lindholm-Beauchemin (spotted with Fowler when trailing or leading, say)
Like the offense, I would simply level out the minutes the Lindholm-Beauchemin pairing play while probably keeping them together. The point of it is to get Fowler on the ice more so that he can be a difference maker, as his equal defensive numbers (to Beauchemin) are backed by a much higher offensive upside. However, the sooner Lindholm and Fowler are explored together, the quicker the team can find out if it has a true shutdown first pairing to compete with the elite teams. (I think it does.)
Also I would pray to my favorite deity that Frederik Andersen stays completely healthy. Then I would eat like 12 ice creams, because I'm possessing Bruce Boudreau, and I suspect he likes Rocky Road.
Follow Kid Ish on Twitter: @kid_ish