Sunday night's breakdown against the Calgary Flames has got me thinking about what the causal factors are for something like that to happen. Is it Dave Cameron's system? Is it his player usage when the Sens are up a few goals? Do the Senators even have the personnel capable of protecting the lead? Well, it turns out that it's likely a combination of all three.
First, the system. Now unfortunately I don't have access to GameCentre anymore to go back through the game and take screenshots like I normally do, but I'm sure anyone who watched the Sens - Flames game on Sunday saw the formerly-aggressive Sens sit back in the third. Characteristic signs of this includes dumping the puck in more, having a passive forecheck, and exiting the zone by chucking the puck into the neutral zone - a turnover - instead of trying to exit with control to create offense on the other end. Here's an example from Nichols:
There was one play in 3rd that was emblematic of Sens sitting back: Ryan was on right side of neutral zone with puck, Hoffman came off bench— The 6th Sens (@6thSens) March 9, 2015
and streaked through middle of NZ with speed, Ryan looked him off and floated a long dump in on Hiller from red line that Hiller covered.— The 6th Sens (@6thSens) March 9, 2015
The problem is more philosophical and starts way before players get to the NHL. It's easier to destroy than to create. As a coach, it's easier to tell your players to suffocate the opposing team and not turn the puck over. There are still players whose imagination and creativity capture the Soviet spirit - Johnny Gaudreau in Calgary, Patrick Kane and Jonathan Toews in Chicago just to name a few. However, they are becoming exceptions to the rule. Many young players who are intelligent and can see the game four moves ahead are not valued. They're told "simple, simple, simple."
That mentality is kind of boring. Nobody wants to get fired. Nobody wants to get sent down to the minors. If you look at the coaches in Juniors and minor league hockey, many of them were not skill players. It's a lot of former enforcers and grinders who take these coaching jobs. Naturally, they tell their players to be just like them. Their players are 17, 18 years old - younger than I was when I joined the Red Army team. Say what you want about the Whiplash mentality (or the Soviet mentality), but if coaches are going to push kids at that age, why are they pushing them to play a simple game? Why aren't coaches pushing them to create a masterpiece?
We lose a lot of Pavel Datsyuks to the closed-minded nature of the AHL and NHL.
Although this quote is directly about the type of players that make it into the NHL and the contrast between North Americans and Soviets, the bolded part (added by me for emphasis) is directly related to the situation at hand. Everybody knows that the best thing to do when you have a lead is to keep piling it on - increase your lead and continue to play aggressively so you don't give the other team a chance to get back in it. But WHY doesn't this happen? Nobody wants to get fired. Nobody wants to be "that player" who goes for an aggressive play but ends up out of position while the opposing team gets an odd-man rush the other way and scores. Despite the likelihood of these kinds of plays happening LESS if you're being aggressive and limiting shot attempts against, everyone is scared of what I like to call the "Big Event" - the goal against that deflates your team, gives you the minus, and gets you a bunch of stares from your teammates. This is why, despite there being 20 teams generating 50%+ shot attempts when the game is tied, that number drops to 5 (!) when the team is leading by one and 0 (!!) when the team is leading by two.
Okay, let's hypothetically say that Cameron molds the Sens into one of those five teams that still keep the foot on the gas when the team is leading. Which players does he use in that situation right now? Well, thanks to the amazingly wonderful Micah Blake McCurdy (@IneffectiveMath), we have a couple of graphs to show this in a unique way. I've been absolutely captivated by score-deployment graphs because I never thought of it having an impact on player performance, but now I think it may have a similar or greater effect than the quality of competition a player faces, given that we know that teams are likely to be filled in possession (and goal) wise when they're up a goal (harder situation) and are likely to be playing a lot of offense trying to tie the game when they're down the goal (easier situation). Now, Micah provided graphs for the entire Senators season, but also split them up to account for the Senators coaching change. The first graph will show the Senators score deployment under Paul MacLean, and the second graph will show the Senators score deployment under Dave Cameron. We're only going to be looking at the defensive personnel because that's the focus of this article, though that's a limitation of my analysis at this time because teams do operate as a five-man unit.
The Y-axis is a measure of ice-time, and the x-axis shows the situation. Thus, we can see that when the score was tied (0), Erik Karlsson was on the ice approximately 45% of the time under Paul MacLean, and 40% of the time under Dave Cameron. This matches Cameron's comments which state that he's been looking to get the top blueliner's ice-time down a bit as he's been playing way too much, which leads to fatigue. We can see that in the absence of Marc Methot, Chris Phillips was playing the second-most minutes of the team unless they were leading by two goals, which is a strange thing to me and maybe speaks to Paul MacLean's wacky player usage: Phillips, who's *bad* at offense, was given the easier situation of having to play a lot when the Sens were trailing and looking to go on offense to tie the game, and was buried more when the Sens were leading. This is counter-intuitive to what we'd expect because strong "defensive" players should be playing when the Senators have a lead in order to protect that lead.
Anyways, lets focus our attention on Dave Cameron's graph (bottom), which makes a lot more sense. The team's only bonafide top-four defensemen - Erik Karlsson and Marc Methot - play a ton of minutes, averaging 5-15% more than any other Senators D. After that, we see a cluster of Senators defensemen, which speaks to me that the coach doesn't really know *who* he should use in those second-pair minutes because they're all not quite there yet (or anymore). We do still see some logical patterns though. Patrick Wiercioch plays the most of his minutes when the Sens are trailing, as they want him to generate offense - which he does despite his production as his possession numbers are through the roof - but plays the least amount of minutes when the Sens have to protect a lead by a fair margin, showing that the coaching staff doesn't have faith in his defensive abilities at all. In contrast, Mark Borowiecki, Eric Gryba, and Chris Phillips smartly don't play as much when the Sens need offense, but all find their ice-time elevated when the Senators have to protect the lead. Jared Cowen and Cody Ceci have very similar score-deployment lines, and I think that's because Cameron used these two as the second-pair for most of his tenure as that's what the organization needed from these two, but it clearly wasn't happening.
Now, can we fault Cameron for this usage? The one thing I'd change is to try playing the puck-moving pairing of Wiercioch - Ceci when the Senators need to defend a lead.
If you look at the shiftchart from the game against Calgary, courtesy of War-On-Ice, we can see Cameron's score-deployment in action. Wiercioch and Ceci played a TON when the Senators were trailing or leading (lots of long red bars signifying shifts), but as soon as the going got rough in the 3rd period, marked on the graph as "40" on the X-axis, we see that Wiercioch - Ceci had four shifts until overtime, when the Senators needed a goal again. Instead, the pairing of Mark Borowiecki and Eric Gryba played a ton, but got their lunch handed to them as they were unable to exit the zone with control or were coached to make the safe play. That brings us back again to the question of WHY this happens, and the Igor Larionov quote. I think Wiercioch and Ceci, as offensive players, aren't trusted in defensive situations because they aren't big, physical players, but we've seen time and time again this year that the best defense involves puck-moving. Hell, it's no coincidence that the Senators are playing their best hockey of the season with FOUR capable puck-movers in their lineup, instead of the TWO they had earlier in the season with Wiercioch inexplicably scratched and Methot injured. Coaches and folks in player evaluation need to get away from the false dichotomy of offense-defense and focus on players who are balanced - who can move the puck into the offensive zone because the point of hockey is to score more goals than the opposition. The more time the puck is in the offensive zone, the less time it's there in yours.
This brings us to the last part of this article: do the Senators even have the personnel to protect the lead? Given that Cameron and co. aren't likely going to read this article and suddenly change their mind on using Wiercioch and Ceci differently, we have to ask the hypothetical question of whether these two *could* protect the lead if trusted to. Now, although there's a lot more contextual factors here like zone exits that I don't have tracked, Micah sent me a graph that has a pretty curious bit of evidence that I'm going to show you.
On the Y-axis, we see Corsi against, which is the number of shot attempts directly heading towards Ottawa's net and our measure of "defense" right now. On the X-axis, we see Corsi for, which is the number of shot attempts directly heading towards the opposition's net and our measure of "offense" right now. Generally speaking, the pairings to the right of the red line have positive possession numbers when they're on the ice together, and the closer you are to the bottom right, the better the pairing is. That's why we see the pairing of Methot and Karlsson (3+65) close to the bottom right generating around 1.0 Corsi for a minute while only giving up 0.8. If we look at the pairings in question (46+5 and 62+74) we see that despite the Wiercioch - Ceci pairing generating way more offense and thus being positive possession players on the right side of the line, they give up the same amount of shot attempts against as Borowiecki - Gryba. The "better" defensive duo gives up the same amount of shot attempts against as Wiercioch - Ceci, and don't provide any offensive benefit to boot. What is a bit counterintuitive to the tone of this article though, is the fact that they do still give up a ton of shot attempts against, and thus, may still spend a fair bit of time in the defensive zone as we would like.
What if Wiercioch - Ceci struggle in those minutes, do the Senators have any other players available to potentially try there? Well, seeing that Chris Phillips is the only player to be paired with Karlsson and have negative possession numbers, he's out. Gryba does okay if he's with Wiercioch, but other than that, him and Borowiecki spend most of their time on the left side of the line. Lastly,
Shea Weber Jared Cowen isn't going to save anyone either, as he brings down every pair he's on and even though he's a positive possession player with Erik Karlsson, the amount of shot attempts they give up a minute is staggering.
What about the future? These players are young and the team has been improving as the season has gone on - both in terms of record and possession numbers - but the defensive personnel, other than Erik Karlsson and Marc Methot, haven't improved at all possession-wise. Micah (please follow him and thank him for me because I wouldn't have been able to produce this article otherwise) has created the graphs below called ribbon plots, which show the players possession numbers as a function of 5-on-5 score-adjusted Corsi For (offense) on the x-axis and 5-on-5 score-adjusted Corsi Against (defense) on the y-axis. Like the previous chart, if you're on the right side of the purple line in the middle, you're in positive possession territory and if you're on the left side, you're not doing so great. There's still competition usage factors to consider, but these graphs are great because they a) adjust for score (like the deployment graphs earlier!), and b) can show a player's possession results from the BEGINNING of the year, indicated by the letter S, and then follow the blue line to see where the same player is at RIGHT NOW, indicated by the letter E. Thus, we can see a player's possession numbers as the season has progressed.
As we can see, there is either very mild (hi Gryba!) or no improvement (Phillips, Borowiecki) as the season has gone on. Remember that the axes are a little different for each graph, so despite Wiercioch not improving much, he's still well to the right of the Purple Line of Competence whereas Chris Phillips is way to the left of it.
Thus, we're left with a troubling result. Dave Cameron is like many other NHL coaches - risk-averse with his systems and using his "defensive" players in defensive situations (when the Senators are leading) despite their mediocre performance. There is a little bit of hope, with Patrick Wiercioch and Cody Ceci perhaps able to handle the harder minutes of playing with a lead, as their extra offense may help the Senators score another goal, though there is room for skepticism as the duo gives up a similar amount of shot attempts against as the defensive pairing of Mark Borowiecki and Eric Gryba. It is also worrisome that these six players have been counted on to produce two top-four calibre defensemen, but have not improved much as the season has progressed, owing to the fact that the Senators increased play and possession numbers are likely due to better player usage on the forward end with skilled players Mike Hoffman, Mark Stone, and Mika Zibanejad getting prime minutes at the expense of others. Therefore, in conclusion, I think the Senators should really, really, really try to acquire a top-four defenseman in the offseason to help the coaching staff and team out. Having one would really stabilize the Senators ability to hold leads like the best teams in the league, and will allow the other, worse defensemen to play in a role that's more comfortable to them. In case you were wondering when I was citing the trailing possession numbers at the beginning of the article, the Senators have the 6th worst possession numbers when up by one, and the 10th best (!) when up by two - both well under 50% though. Thus, it seems to be that the Senators really struggle when they feel the pressure of defending a one-goal lead, and having another capable defender would certainly help with that.
Thanks for reading!