5 things: Quantifying hard work

1. What are we talking about?

People do not like when I refer to the Flames as a “bad team.” They tell me to look at the standings, because in sports, all people really care about is results. If a team could get away with being outshot every single night all season long — and by a margin that, say, as wide as what the Flames face on any given night — and still win all 82, then that would have been a good team. The process of how those wins are arrived-at matters not.

We know that this isn’t how things work, of course, and we know that outshooting your opponent more often than not is a pretty good way of likewise ensuring that you win more often than not. The numbers back all of this up, and even those who for some unfathomable reason remain skeptical of the efficacy of evaluating hockey teams using things as simple as their corsi percentage over the course of 20, 40, 60, even 82 games would have to agree that the best teams have the puck more than the worst ones. It’s a sliding scale, sure, and things don’t always match up 1-for-1, but it’s obvious that a good corsi share is key to a good record in far more cases than those for which the bad leads to a good record. We can all agree on this.

But the Flames are a team with a bad corsi share — a very bad one, in fact — and yet they continue to win most nights out. Often, they do it in improbable fashion. Much like other teams before them, who have forged playoff but not possession credibility over the bulk of the season, the Flames attribute this success not to fortunate bounces, but to hard work.

Hard work, we’re told, is not something you can quantify. But the thing is that if hard work shows up in the wins column, it necessarily has to show up in the goals column. And if it shows up in the goals column, it will show up somewhere in the “advanced” stats, whether it’s through corsi or PDO or Hextally. These “clutch” goals that have proven so crucial to the Flames’ success do not appear out of nowhere. If it’s as simple as working hard, there’s going to be a process behind it that generates the crucial scoring chances they’re converting and prevents them at the other end.

At this point I’m really and actually trying to understand the thought process for people who think this is a thing the Flames can do in perpetuity.

2. Possession

Okay, the Flames are a bad possession team. How bad? They’ve been third-worst in the league basically all year, ahead of only Buffalo (the worst possession team since people started tracking these stats by a mile) and Colorado (one of the other really really bad ones). Calgary is not historically bad, but it’s probably going to be in the bottom 10 for the salary cap era at the end of the season. So, yeah, bad.

(Please note that none of the following stats include last night’s game — and the four goals on seven shots over the last 40 minutes — and all of it is from War on Ice.)

But what’s interesting is that if you look at things on a period-by-period basis, the Flames get progressively better at holding onto the puck throughout the game. In the first period, they are a 44.6 percent possession team, which is third-worst in the league and about what you’d expect. In the second, that number actually drops somewhat to 43.5 percent, and is second-bottom in the NHL. In the third, it’s up to 45.1 percent, which is a little better but still pretty miserable. (Elite NHL teams usually post a CF% of about 55 percent, maybe a little higher, so in a best-case scenario, every team the Flames play all season long holds the puck like the Los Angeles Kings, on average.)

But this might not be what “hard work” means to people. The fact is the Flames typically end up trailing pretty early on in games. You don’t necessarily have to work hard when you’re trying to hold onto a win, but you have to do so when you’re trailing late. The Flames, not surprisingly, trail late rather a lot. But perhaps not as much as you’d expect. (Though the numbers have been getting worse as the season goes on, which is really something to think about.)

In terms of the number of minutes at 5-on-5 the Flames are actually behind in the third period, their 399.7 is “only” the sixth-largest total in the league, and actually a pretty large ways back of the cluster of worse-off teams like Arizona, Toronto, Columbus, and of course Buffalo and Edmonton. From a trailing position, they draw more penalties (plus-5) and start more shifts in the offensive zone (plus-30). Over the course of a season, this is a negligible amount; one extra zone start every two or so games, and one extra penalty every 12 or 13. It’s not nothing but it’s close.

And when they’re trailing in the third period, they do actually outpossess their opponents, with a CF% of 51.9 percent. That is, however, still third from the bottom in the entire National Hockey League. I’m very dubious indeed, then, that the hard work we’re looking for is in these numbers.

But perhaps all of that stuff about corsi not being related to hard work is correct, so we need to look elsewhere.

3. Goals

In terms of the number of goals scored during the entire game, the Flames are tied for eighth in the league at 2.8 per. That’s the same as Nashville, Minnesota, Pittsburgh, Anaheim, Vancouver, Ottawa, and San Jose. The point in listing all those teams is to show that they’re very firmly grouped in the upper middle of the league.

But we once again have to look at the breakdown. The Flames’ goals-for total in the first period is just 36, tied for 28th in the league. In the second period, it explodes to 58, which bumps the club up to 20th. In the third, it’s 84, and that’s second-best behind the offensive juggernaut Tampa Bay Lightning. In overtime, they have another nine, which is tops in the NHL.

So we’re looking at 93 goals in the third period or overtime, versus 94 in the first two periods. Which is significant because, at most, you’re looking at 25 minutes in the former versus 40 in the latter.

What about the goals they’re allowing? Well, the 59 they’ve conceded in the first period is the seventh-largest total in the league. The 60 in the second period is only 19th-biggest. In the third, it drops to just 48, which is the smallest in the NHL. In overtime, they’ve allowed just three goals, tied for 14th-most.

Let’s break that down by differential in the first-plus-second and third-plus-OT: They’re minus-25 in the opening 40 minutes, and plus-42 in the final 20-to-25.

At 5-on-5, which is when it’s easiest to judge team performance (but we’ll still cover special teams in a bit), the Flames are minus-17 in the first two periods and plus-11 in the third.

This incredible ability to “flip a switch” and outscore their opponents so dramatically is starting to show us something about the hard work, I suspect. If the people who are “on the fence” about analytics, or just outright reject them, are to be believed, that’s something they can replicate all season long despite obvious talent and possession deficiencies.

4. Shooting and saving

As you might imagine, when there’s a pretty stark improvement in the team’s ability to score goals at any point in the season — sustainable or not, that’s reflected in their percentages. No surprise here, but the Flames’ PDO (shooting percentage plus save percentage) is a third-in-the-league 103.6 in the third period. When they trail, it’s basically unchanged at 103.5, but that number is actually first in the league.

And here’s the difference as far as that goes: Teams that trail tend to shoot more often, but that generally leads to a lower shooting percentage because they are typically taking lower-quality shots. For the Flames, their trailing-in-the-third shooting percentage is 10.3 percentage, down a little from the 10.9 percent in all score situations. Which suggests that even the Flames aren’t immune to this portion of hockey math.

Calgary continues to generate some of the fewest shot attempts in the league in these trailing situations (27th in corsi for per 60) and continue to allow a bunch as well (29th in corsi against per 60).

So it stands to reason that they’re generating more scoring chances at this time, right? Well, yes and no. On the one hand, the number of attempts they make that actually wind up on net are at some of the lowest levels in the league (47 percent is second-worst, actually) but it’s because they’re still getting a large percentage of their shot attempts blocked. The number of unblocked shot attempts that get on net is actually 71.6 percent, and that’s much more middle of the pack.

The Flames are actually pretty decent at generating shots on goal in these situations, but they still get outshot (174 for, 176 against) when they trail.

A lot of this has ignored how they’ve done in their own zone, though. At least in terms of keeping the puck out of their own net. Their 5-on-5 save percentage when they’re trailing in the third is seventh in the league (.932) which compares favorably with their far more middling overall ESsv% of .922 (18th).

Here, there’s a lot more of what a coach would call “attention to detail.” While the Flames have blocked 1,074 shot attempts the Flames have at 5-on-5 this season, only 85 have been when they’re down a man (6.7 percent of the season’s total). That’s the second-largest total in the league. They block nearly 1 in 4 shot attempts they face at this time, but that’s actually down from 1 in 3 in all score situations.

Unfortunately War on Ice doesn’t have the ability to show save percentages by period, but we can look at the Flames’ netminders overall versus when the team is trailing. What’s pertinent here is the quality of chances they’re giving up, and how well the goaltenders are stopping them. So when it comes to shots that have a high probability of going in, Jonas Hiller and Kari Ramo stop .843 and .837, respectively. These numbers aren’t bad, but they’re not good either. That’s on a combined 350 high-quality shots all season long. But when the Flames are trailing, they’ve faced 150 such shots, and in these situations their save percentages explode to .880 and .873, respectively. 

Teams actually tend to take higher-quality shots when they’re ahead, because while they’re being conservative, they’re not going to pass up a free lunch. The Flames give out lots of free lunches and, for some reason or another, Ramo and Hiller are much better at keeping the Flames in relatively small arrears than in the black.

5. Special teams

Special teams, of course, were not included in the above data, except in the overall goal totals. As you can see, a pretty significant portion of the Flames’ scoring this season has come on the power play. In fact, 35 goals on the power play makes up a little less than 1 in every 5 they score.

But what’s interesting is that the Flames shoot only a little bit better on the power play (11.8 percent) than they do at 5-on-5, and this is something that just doesn’t happen very often. Typically, you shoot substantially better, by a few points, than you do at 5-on-5. They don’t generate a particularly large number of shots (12th in the league, so not a bad number either) but a lot of them sure do go in. They also don’t draw a ton of penalties relative to anyone else in the league.

The Flames aren’t a great penalty killing team, either. Their save percentage is among the worst in the league. But what they do well is not-take penalties in the first place (they’ve been shorthanded the lowest number of times in the league) and suppress shot attempts pretty well when that happens, as their corsi against per 60 when they’re shorthanded is fifth in the NHL. These two things don’t usually follow for a team that gets badly outpossessed every single night, but it’s certainly happened for the Flames this year and last year. 

And here’s what’s really, really, really interesting, and isn’t usually included in “special teams” consideration: The Flames are having an incredible season when they pull the goalie. The Flames have scored nine extra-attacker goals this season on just 50 shots. The shot generation is middling, but the success rate is third-best in the league. And the other thing is, they don’t give up a ton of empty net goals. They’ve conceded nine of them, but that’s a pretty small number for a team that’s spent 48.5 minutes with their own crease unoccupied.

So the question is this: Is any of this repeatable? History suggests that this much shooting success doesn’t happen over long periods, including if you’re Alex Ovechkin. As a team, the Flames have shot like Alex Ovechkin, and you can’t have that continued success, whether it’s at 5-on-5 or in special teams situations.

In their own end, they’re still giving up a ton of shots but they’re saving a decent enough amount. I’d think that can keep happening for Hiller because of his long track record, and maybe Ramo.

If there’s one thing that can keep this going — and I doubt that it can, but they have to hope — it’s that they’re very stingy in giving up power plays. For a club with a lot of guys who aren’t very good, one skill they do seem to have is the ability to avoid the penalty box. I’m not sure if that qualifies as “hard work,” but so far it’s the only thing I’ve found that they do well and reliably.

  • Craig

    Fun games to watch, but the real indicator will be the extent that the Flames improve next year. If they continue to win while getting outshot then maybe we’ve found a new style or something. If they collapse then it’s the same old story.

  • CutterMcAwesome

    There has to be something said about where the shots are being allowed from. Are they from high percentage scoring areas? or are they low percentage?
    I get that a corsi event can happen from anywhere.

    What is the percentage of the flames and their opponents shots that actually come from a high percentage area?

    Calgary seems extremely adept at taking their shots from high percentage areas, which would lead to their higher shooting percentages.

    They also seem to do a half decent job of limiting the other teams shots to the lower percentage areas.

    So possession matters yes, but so does shot quality. If the flames can continue to limit their oppositions quality of possession. Then they can continue to win games.

  • Lambert, I do respect that you try to dissect the Flames, and I do believe there is a place for analytics in hockey. However, it seems that you no longer even enjoy the games, and the way you write about the Flames makes me think that you visibly get upset whenever they score, or win.

    Its definitely important to maintain our expectations and figure out how the team could improve, but please enjoy the game while its exciting. You may not think the Flames are good, but they are doing well, and they are exceptionally fun to watch. Have a beer, and have fun watching them!

  • I don’t know enough about the actual mathematics that work into generating Corsi percentages to qualify my opinion….but here is my opinion.

    The Flames have several players that are bad (below league average) these players do irrepairible damage to the Flame’s Corsi numbers while they are on the ice I.e Mr. Engelland. This damage is then mitigated by the very good Flames players when it is their turn…but because these players are still young and emerging they are not yet at a point where their “good” can outweigh the “bad” done by other players.

    I also believe that the Flames number one line has two not good BUT VERY GOOD…dare I say elite young men playing on it at this point. Our top line is difficult to play against and perhaps their shooting percentage is sustainable and we just don’t realize how good number 23 and 13 are at this point??

  • RedMan

    “At this point I’m really and actually trying to understand the thought process for people who think this is a thing the Flames can do in perpetuity.”

    I hear you – you are trying to understand these people.

    Only thing is – where are they??? i read every article hear at FN, and, while I cannot say these people are nonexistent, they are so far outnumbered by the rest of us who appreciate the perspective that CORSI and other new stats brings, and also enjoy the team wins and the miracle season they have had. Miracle or luck – depending on one’s view.

    Who are you arguing with??? looking through the responses above, i counted one that indicated there could be sustainability.

    My personal hope is that, as this team develops and grows from it’s rebuilding infancy – that the possession stats improve.

    So, when you say they are a bad team, my only request is this: please add the caveat that they are a bad team statistically speaking regarding CORSI and whatever it is CORSI tells us.

    The fact is the Flames are a great team, when measuring fan fun, entertainment value, fan experience and playoff races. its been nothing short of an incredibly fun year for everyone else – hasn’t it been fun for you?????

    possesion wise

  • SoCalFlamesFan

    I’m pretty sure that I would quantify hard work with

    – Number of completed checks leading to…
    – Number of takeaways leading to open ice scoring chances
    – Cleared rebounds by skaters off of opposing shot attempts

    But sure, let’s make every damn thing in the universe about corsi and SH% as if they exist in a vacuum.

  • Derzie

    “People do not like when I refer to the Flames as a “bad team.””

    No, people find the good/bad dichotomy to be coarse and wholly uninformative to the point of being absurd. I would add I find it apparently intentionally misleading myself.

    “…it’s obvious that a good corsi share is key to a good record in far more cases than those for which the bad leads to a good record. We can all agree on this.”

    No, we cannot all agree on this. ALL other things being equal, we can all agree on this. But that’s not what you said. And ALL other things are never equal, as you damn well know.

    “and if it shows up in the goals column, it will show up somewhere in the “advanced” stats, whether it’s through corsi or PDO or Hextally.”

    This is circular reasoning. You have to demonstrate this. You cannot just assume it in the course of vindicating the advanced stats explanation for the “badness” of the team.

    There’s more but I have to go do more important things. Peace.

  • piscera.infada

    The one thing I wonder about with regards to Corsi and the Flames is about the sample-size(s)–I hope someone far more well versed in the area can enlighten me. How do we determine with any real accuracy the “predictive success” of Corsi as a team statistic when the roster has seen so much consistent turnover through the course of 67 games?

    This question isn’t meant to take away from the analysis at all, but rather something I’ve been trying to wrap my head around. I understand that the Flames are a poor possession team on a prima facie level (I still firmly believe that in some games a zone-time-possession statistic would make them appear “less poor”–case in point the second period last night–although likely still “poor”, but I digress).

    I wonder how much the turbulence of the first several months of the season negatively impacted possession through the fact that the Flames lost 3 NHL regulars upfront and 2 of their better veteran centres (Stajan and Backlund)–this forced a lot of players that likely wouldn’t be ready to jump into a “bus-driving” role to sink or swim. Moreover, we saw incredible turnover out of Adirondak with players coming in and out of the line-up ad nausea, as Hartley tried to find players to fill the gaps. Both of these situations caused the Flames lines to be consistently juggled–one has to wonder how much Corsi is affected by line mates on the micro-individual level; now imagine that on the macro level. Then, their best possession driving player and defenseman goes down for the season, forcing their second best possession driving player and defenseman (rightly or wrongly) to play with a far less adept possession driver (read: horrible possession player).

    None of this is to say the Flames are a good possession team, or that they have found a way around puck possession. I simply wonder how we can reconcile a clouded sample size. It seems well and good to talk about possession power-houses (like the Kings) perennially because the core of their roster largely stays static–aside from periodic injuries and/or other personal issues.

    The sample-size question posed above is a product of a young rebuilding team still trying to find it’s way. As such, I’m not too sure how much credence we can even give the answer. I’m almost certain we all went into this season assuming the Flames would be poor at possession. Why? Because their only possession drivers are a top-defensive pairing, two players under 21 years old, and Backlund–two of those players would not have been pegged as a “driver” to start the season.

  • RedMan

    I sure wish the play by play guys would start calling CORSI real time, that would be exciting!!!

    Lombardias: “Mikael Backlund just absolutely CRUSHED Getslaf with a heavy CORSI!!!”

  • RedMan

    what was that lambert quote?

    I’m sure somebody remembers but it was something like success can’t be measured by wins. funniest comment I’ve seen in ages. wish I could find it again.

    • TheoForever

      Yes, Lambert did make a comment like this, it reminded me of MacT school of thinking. MacT said something that Eakins performed all his coaching duties exceptionally well and was excellent coach, except for that winning part.

      Visually speaking Lambert would look well in the Oiler organization.

  • CutterMcAwesome

    God Lambert, you must be the most worthless tw@t to have ever tried to write about hockey. And that’s saying something.

    It’s embarrassing having a blowhard like you writing about as special of a team as this year’s incarnation of the Calgary Flames. You don’t deserve any sort of voice, let alone one in this arena.

    I look forward to the progression and acceptance of hockey analytics; as much as anything, so that more competent writers arrive on the scene and numbskulls like you are back flipping burgers.

    “Yes, I want cheese on it. That’s what a cheeseburger means, Lambert.”

  • Shooter 5567

    Thanks Ryan for your stats update and your personal conclusion that the Flames are a bad team.

    There is no point right now to debate whether you are right or wrong based on the data. The Flames are what they are. And most important their success is based on that fact which they all know who and what they are and play that way night after night. This lends well to the entertainment side of the game for most fans. Hartley even said they are a run and gun team and encourages that style.

    If they improve their Corsi stats in the future this DNA will have to change in order to change the stats. If you think they are a bad team because of who they are then this bad team will stay this way until the end of the season or the last playoff round which ever comes first.

    As for next year, they have as according to the stats a bright future with something to build on.

  • smith

    I am unsure how corsi can be called a predictive stat when nobody can tell me when the flames will regress. I am absolutely sure corsi has value, but obviously not enough by itself to show everything that is going on.

    Calgary is not worse than Phoenix or Edmonton.

  • smith

    -I think it’s fair to say that the Flames are severely over achieving.
    -however calling them ‘bad’ is a very one-dimensional way of viewing their succes.
    -I think the cohesiveness of the group is really outstanding.
    -I think the group has a lot of strengths and a lot of weaknesses.
    -There have been a huge number of statistically dominant teams that have never displayed the intangibles necessary to win championships.
    -Management has treated this year exactly the way every statistician hoped they would. Building a solid base of draft picks without hurting the team too much short term.
    – Our core is incredibly young and going to improve at generating shots and preventing them.

      • DoubleDIon

        I love the “you people.” Corsi is a tool. It needs to be contextualized. If you use it along with a whole bunch of variables then it has great validity.

        If you isolate it then you have an incomplete picture.

        TJ Brodie had a pretty good +/- when he was paired with Giordano. I’m guessing it won’t be so good with Engelland.

        Calgary has a terrible 5-5 Corsi. They also block a lot of shots and have a huge positive penalty differential. That helps to explain why they are in a playoff spot.

        No one is lucky enough to actually be the 3rd worst team in the league, worse than Phoenix or Edmonton and make the playoffs.

        Reality is, we might be the 20th or 21st place team if not for a decent amount of luck. We’re definitely not at the Phoenix, Buffalo or Edmonton level even if our luck was terrible.

      • RedMan

        Hi Justin!

        I just wanted to thank you for having never submitted another article since you got flamed for that embarrassment you tried to pass off as writing. Much appreciated. 🙂

  • piscera.infada

    The piece were all missing to understand this is quality of scoring chances… will we ever get there with stats maybe not… already stats like hits and turnovers are quite subjective.

    But I do want to point out some of the things that, for instance, Monahan is doing. Last night that goal he scored was fantastic. He held on to the puck surroundex by 3 ducks and picked his spot. Not many players would be able to do that.

    Im sure luck is somewhat involved in Calgarys succesz but you make your own luck and you can conciously make the most out of your scoring chances.

    The Flames are creating great scoring chances and finishing them. That is a combination of skill and attention to detail.

  • piscera.infada

    @Justin Azevedo

    Channeling your inner Lambert again? Patronize the readership instead of engaging in a discussion. Classy. Keep your stick on the ice bud.

  • DoubleDIon

    Ok. We have an incredibly in-depth look at the Flames here. No hyperbole and no wild statements.

    You keep writing like this and you become a Kent Wildon and even make the jump to conventional media one day.

    This almost feels like a light switches on moment to me. I hope it is!

    Good job.

  • Reidja

    Seems that difference between top team and bottom team is incredibly small what with league parity through Draft and team total salary caps and what with every team sniffing in the same bonanza Gold creeks, ie international, university, junior, prison, etc.

    With league frenzy for equal chance hasn’t corsi been ‘knee capped ‘ by league ‘powers to be’ or if it hasn’t than maybe we should have corsi for evaluating league success in achieving ‘equal chance’ or not?

    Maybe corsi is only relevant when league ‘higher ups’ fail at ‘equal chance’ in which case corsi has never measured team potential to each other but only league administration failure!