Mapping the Bruins return to 5 on 5 greatness

I’m going to be talking about stats here. Not at a lot, but I’ll be mentioning 3 specific ones a lot. Here’s a primer on them:

CF%: Corsi-for percentage. The percentage of all shot attempts controlled by your team. Example: If I had 4 shot attempts and you had 6, my CF% would be 40%

FF%: Fenwick-for percentage. CF% but without any blocked shots, giving teams credit for blocking shots. Example: If I had 4 shot attempts and you had 6, but I blocked 2 of your shots, my FF% would be 50%

xGF%: Expected goals-for percentage. FF% but with every shot weighted on it’s likelihood of being a goal, i.e. it’s expected shooting percentage. We’re using Corsica’s model, which bases expected shooting percentage on the shot’s distance, angle, and a number of other factors (Was it a rebound? Did it come off a rush? Wrist shot? One-timer? etc). Example: If we each had 1 unblocked shot on goal, but my shot was twice as likely to be a goal based on it’s expected shooting percentage, I’d have an xGF% of 66.6%.

CF% is the most repeatable at smaller sample sizes, FF% is the second most, and xGF% is the least.

Still here? Good.

For 4 years, between the start of the 2010-11 season and the end of the 2013-14 season, the Bruins were far and away the best team in the NHL at 5 on 5. It wasn’t even close. Per 60 minutes, they’d score 2.67 goals and allow only 1.91, meaning that the Bruins had 58.3% of all goals when they played. The next best team, Chicago, had only 54.1%. Meaning that the Bruins were over twice as far above 50% as the next best team. They were truly dominant.

Then, something happened. The Bruins got a lot less good. The last two seasons, they’ve been middling, at best. Entirely average at 5 on 5. And there’s a long list of reasons as to why. But we’re not here to talk about that. Possible reasons as to why the Bruins might have become crap has been well documented, on this site and others.

What we are going to talk about is the Bruins sudden surge in 5 on 5 play and what’s led to it. Because, make no mistake, there has been a surge. From Corsica, so far this season, the Bruins are #1 in the NHL in score-adjusted CF%. They’re #1 in score-adjusted FF%. They’re #1 in score-adjusted xGF%. Over the past 25 games, the Bruins score-adjusted CF% is the highest it’s been during any 25 game period since the 2013-14 season where the Bruins won the president’s trophy. Here’s a graph of that, from Corsica. That giant 10% jump at the end is what we’re going to talk about.

Now, is there any guarantee that this is going to last? Well, no. They’ve played 16 games this year. The Bruins are just coming off of wins against Arizona and Colorado, two of the worst 5 on 5 teams in the league. Even still, it’s clear that something good has happened. It’s possible that the Bruins may be returning back to their old ways of 5-on-5 dominance, and we’re going to look at the steps they took to get here.

Step one: Draining the swamp

Just as a goal against is as important as a goal for, your worst players can be just as important as your best players. A guy who adds +5 goals to your team’s overall goal differential at the end of the season will have all his work reversed by a guy who adds -5 to it. That was the story of the Bruins at 5 on 5 for the past two years. Despite having easily the best LW-C combo in the league at driving possession, the Bruins were around an average possession team.

Here’s all the players in the past two season who had CF%s lower than 48.5% (CF% is the number in bold):

These are the guys who were dragging the team down in recent years. Let’s use 14-15 Bergeron as a benchmark for how bad these guys were. Statistically, Bergeron had one of the greatest defensive 5v5 seasons of all time in 2014-15. That year he had a CF% of 58.0%, so 8% above 50%, and played ~ 1000 minutes at 5 on 5. Using him as a benchmark, in 2014-15 the Bruins had Gregory Campbell play about 70% of the TOI Bergeron did. In that time, the team had a CF% of 41.3%, or -8.7% beneath 50%. In other words, the Bruins were almost just as bad with Campbell on the ice as they were with Bergeron on the ice, and they played almost as many minutes!

CF%, of course, isn’t the end-all-be-all of 5 on 5 play. But it’s really, really hard to be a good team when you have guys who are almost as bad as Bergeron is good on the ice for almost as much time as Bergeron is. In 15-16, the problem only got worse. Instead of having a few bad players and just one truly atrocious player like in 14-15, you had some players who didn’t do that great and about FOUR who were truly atrocious. None of them individually played as much as Campbell did in 14-15, but Ferraro came close, and Talbot and Rinaldo were somehow even worse than Campbell. Again, these aren’t definite rankings of how good players are, etc, etc. But I want you to appreciate just how hard it is to be a good team when you have players like this playing the minutes like that.

Now, here’s the list of players with a CF% < 48.5% in the 2016-17 season so far:

Keep in mind a few things. The sample size on these guys is very small. I had to cut the TOI requirement I used in the past two seasons in half, from 200 minutes to 100. Also, CF% is no where near the only important thing when it comes to 5 on 5 play (despite being the most repeatable one at sample sizes like this).

Still, the Bruins current worst players still look much, much better than they did during the past few years. Jimmy Hayes hasn’t been that good for the Bruins since about halfway through last season, but he’s young and has scored in this league before. He’s a 26 year old who was 1 goal away from hitting 20 goals 2 years ago when he was 24. No, he hasn’t been the cheap middle six scoring forward the Bruins hoped they’d get when they traded for him. But if he’s your worst player? You’re probably doing pretty all right.

When it comes to Schaller and Moore, the raw CF% definitely doesn’t tell the whole story. They’ve taken lots of defensive zone starts, but more importantly they’ve been exceptionally good at influencing shot quality. Looking at the xGF% of all lines that have played 50 or more minutes together at 5 on 5 and adjusting for score effects and zone starts, the Schaller-Moore-Accairi line was 9th in the ENTIRE NHL, with an xGF% over 60%.

Don’t bet on their stats lasting, as xGF% is a stat that’s very prone to variance in small sample sizes. But don’t be surprised if the Bruins 4th line is good at supressing goals this year. Schaller and Accairi both suppressed goals at an elite rate in small sample sizes last season. While +/- isn’t a stat you should pay attention to, Tim Schaller was hilariously enough one of only 4 players on last year’s Sabres to play more than 10 games and have a positive +/-. Beyond that, Moore is talented on faceoffs and all 3 of the players on that line are good penalty killers.

Having a great defensive line you can trust for important draws against touch competition is a powerful tool. Anyone who’s watched the Blackhawks in the playoffs the past few years should know this. The player in the NHL with the toughest deployments over the past few years is Marcus Kruger. The player with the easiest? Patrick Kane. That’s not a coincidence. Both are one dimensional players, but playing both to their strengths helps the team tremendously overall.

On the Bruins, having a line like Schaller-Moore-Accairi helps guys like Spooner and Vatrano (when healthy), who are good at scoring but not that great at everything else. I’d wager that easier minutes, due to those players, is part of the reason why Spooner’s 5v5 stats this year have improved. The C depth the Bruins have now in their bottom 6 also helped Spooner’s stats, since he can now play wing. Yes, it’s hard to watch Spooner play wing at 5v5 when you’ve seen what he’s capable of in the offensive zone as a C. But I’d argue that it was even harder to watch him last year in the defensive zone and neutral zone where he’d sometimes look flat-out lost in his center duties. Beyond that, his speed on his current line has undoubtedly helped make up for David Krejci’s declining footspeed. You get to play him there only because you have so many natural centers at the bottom of your roster, with Moore, Nash, Accairi, Schaller, Kuraly, and Czarnik all spending most of their careers as pivots.

Removing flat-out bad players and replacing them with niche, but good and useful players has definitely been the biggest contributor to the Bruins recent 5v5 success. The bottom of the roster matters. It matters a lot, and that should be clear now to Bruins fans. It’s the difference between your best players carrying an average team, or leading a good team. In 14-16, the Bruins’ best players carried an average team. But so far this year, they’ve been leading a good team.

Step two: Bend-but-don’t-break defense

I talked a lot about the fourth line, and forwards in general, in the first section. If you’ve watched the Bruins the past couple years, you’d know the fourth line was a problem. But, it wasn’t the problem.

The problem was the defense.

The problem hasn’t necessarily been solved. The Bruins still don’t really have any long term solutions on defense. Sure, Carlo may turn into that, but if Chara retired tomorrow? The Bruins would be screwed. It’s still hard to argue that the Bruins have a “good” defense.

What they do have, however, is a “good enough” defense.

Compared to playoff teams last year, the Bruins don’t have a good first pairing. Neither Chara or Carlo are going to win the Norris. They play about 22-23 minutes a night, neither really contributing on the PP. Overall, they look more like a good second pairing on a playoff team than a good first pairing. Think “Ellis-Ekholm”, not “Hedman-Stralman”.

Krug-McQuaid is probably on par with most playoff team’s second pairings, with the two of them playing around 20 minutes a night on average. The thing that sets the Bruins apart, however, is that their 3rd pairing has been better than lots of team’s second pairings. The key for the Bruins defense is that there isn’t much of a drop off between pairs. They’re all “good”. None are “great”, but none are bad, either.

And that’s fine. You don’t need an elite first pairing to have a good defense. The Bruins blueline in 2014-15 was still probably their biggest weakness. They were, overall, just unable to drive possession enough for the forwards to score. This was despite having one of the best defensive pairings in the league that year in Chara-Hamilton (R.I.P. in Peace). The problem was the drop off past that pairing was horrible. The Bruins might have made the playoffs that year if they had 3 “good” pairings instead of the one “great” and two “awful” ones they did have.

Let’s look at why this current scheme is working where prior schemes failed. Last season, and the season before that, the Bruins would spend many shifts just stuck in their own zone. The problem was that the team often just wouldn’t have enough players who could clear the zone. Now, that very rarely happens. As with everything related to possession, the solution starts with the blueline.

The Bruins 6 most played defenseman this year, in order by 5v5 TOI, are Chara, Carlo, Krug, Liles, Chiller, and McQuaid, with McQuaid having played over 3x as many minutes as the 7th most played D. If you were to ask me to rank these 6 players by how good they are at clearing the zone, and nothing else, I’d say:

1. Krug
2. Chara
3. Liles
4. Chiller
5. Carlo
6. McQuaid

When healthy, the Bruins have almost always played Chara-Carlo, Krug-McQuaid, and Liles-Chiller as their 3 defensive pairings. The Bruins always have one of their 3 best D men at clearing the zone on the ice. Beyond that, they’re playing their best at clearing the zone with their worst, their second best with their second worst, and so on.

Overall, this speaks to another trend in the Bruins defensive pairings so far this year. The Bruins really don’t have any defensemen who are good at everything. Everyone has clear faults. But teams are having a tough time taking advantage of those faults because of how the defensive pairs are set up.

Chara’s footspeed is an issue. But whenever Chara gets flat out beat in a race, Carlo is always a little bit behind the play and can use his great skating to get back and cover the guy. Carlo isn’t the best at making a pass out of his zone, and can’t really do anything with the puck in the offensive zone once he’s got it. But that’s OK, because he passes it to Chara, who does know how to do all those things.

Krug’s biggest weaknesses as a player is that he’s bad in front of the net and isn’t that great at generating defensive zone turnovers off of a backcheck or board battle. But, guess what, those are literally the only things McQuaid is good at, and he’s done a good job of making up for Krug’s faults. His job is, essentially, to get the puck to Krug and then try to not get in the way. That may sound condescending, but the pair is working really well so far, and that’s important. In today’s NHL, the thing that contributes the most to winning and losing is how well your team plays in the neutral zone. Krug is arguably the team’s best defenseman in the neutral zone, so it’s important that he’s put into positions where he can be the difference maker he has the potential to be.

Liles is 36, and it’s clear that his play has declined and that he’s no longer the top pairing D he once was. His skating isn’t what it used to be and he never really had much size and strength. But, he’s still a top 4 D on most teams, and that’s due to the things age can’t really take away from a player. He’s one of the best decision makers and passers on the Bruins blueline. He’s usually paired with Colin Miller, who, on the flip side, is a young, fast, strong, big blueliner who needs help on his passing and decision making.

All 3 pairings are made up of polar opposites. And while teams have, at times, figured out how to exploit that (with the Liles-Chiller pairing being the most often exposed), it’s nowhere near as bad as it was in the past two years. There’s no Dennis Seidenberg slowly meandering around in his own zone for 45 seconds at a time, pointed vaguely in the direction of the play. There’s no Morrow-McQuaid pairing, where neither player can ever seem to make a zone exit. Krug can play defense all on his own now. You don’t have to see Chara looking dejected after a goal when he and his defensive partner just get flat out beat by a skater neither of them had a chance to catch.

Yes, it’s a very tentative set up. One good injury and the set up is likely ruined. McQuaid, and maybe Chiller, are probably the only two D who the Bruins can afford to lose for an extended period of time. They’re not perfect. They will bend. But, unlike bluelines of years past, if everything holds this defense will not break.

Step three: Retaining Claude Julien

Claude Julien has always been a controversial figure among Bruins fans, even when they were doing well. After missing the playoffs for two years in a row, the detractors got louder than they ever had been before. There have always been two main accusations against Julien. The first is that he values defense over offense to a fault, and the second being that he seems to not give new kids much of a chance.

The first accusation has always been pretty bullshit, at least as far as 5 on 5 play is concerned. From 2010-14, no team scored more goals at 5 on 5. It’s very, very hard to be the best at something you don’t value. As far as the powerplay is concerned, that’s a different story. The Bruins have rarely had a good PP under Claude Julien. That’s why, despite leading the league in goals from 10-14, no individual Bruin scored more than 30 goals or even hit 70 points. But as a team? The Bruins have scored plenty under Claude Julien.

The second accusation used to have a bit more weight, especially after 14-15. That year, blatantly bad vets like Gregory Campbell who were far below replacement level played game in and game out. But last year, 15-16, Claude played young kids almost to a fault. Guys like Ryan Spooner, Landon Ferraro, and Joe Morrow probably got a lot more minutes than they deserved. On top of that, the Bruins were the kind of high-flying offensive team with a great PP that most said Claude would never play.

Common sense says you can’t give a coach credit for a team’s success while simultaneously excusing him when the team does bad. If he was the reason they did well, he must have been the reason for them not doing well. But anyone who saw the talent level Julien was given to work with in recent years knows that’s not a valid line of reasoning. The Bruins were the best team in the league at 5 on 5 by a landslide for almost half a decade under Julien. They didn’t do it off of the backs of lots of super star skaters. Yes, Bergeron, Chara, Marchand, etc are all good. But they alone didn’t make the team as great as they were. The Bruins were great at 5 on 5 because they played great as a team at 5 on 5. Julien had 4 lines and 3 pairs of good players, and he made them into the best team in the league.

But that strength of Julien’s system is also it’s biggest weakness. Julien’s system is good at elevating a collection of good players into a great team. He can’t turn some good players and some bad players into a great team, or even a good team. Everyone on Claude’s teams needs to be reliable. And, after 13-14, talent left the team. Some due to free agency, some due to trades, some just gone from players aging. It was most obvious on the blueline and in the team’s bottom 6 forwards. The Bruins attempts at replacing lost talent were failures. For 2 years, the team flat out had too many bad players, and his systems faults were made clear.

Now, the 16-17 team is starting to look like the Claude Julien teams of old. The advanced metrics show that the team really doesn’t have that many “bad” players left. Everyone is good. Everyone is reliable. It’s becoming clear that the coach was never the problem. In fact, Claude probably made the team look better at 5 on 5 from 14-16 than they really were.

If everything holds, the 2016-17 team will be another great 5 on 5 Claude Julien team.