Scoring Chances, Corsi, Fenwick and Whatnots

Screen Shot 2015-02-09 at 12.21.28 PM

The Calgary Flames play their 54th game of the season later today. In anticipation of that contest, let’s review some numbers, shall we?

There are three commonly-used possession statistics: Corsi, Fenwick and Scoring Chances. By sheer coincidence, all can be found on the helpful War On Ice site.

  • Corsi percentage is the percentage of all shots, blocked shots and missed shots towards the opponent’s net, no matter where they’re taken.
  • Fenwick percentage is the percentage of all shots and missed shots, ignoring blocked shots.
  • Scoring Chances, per War on Ice’s explanation: (detailed with images at numbers at the link)
    “In the low danger zone, unblocked rebounds and rush shots only.
    In the medium danger zone, all unblocked shots.
    In the high danger zone, all shot attempts (since blocked shots
    taken here may be more representative of more “wide-open nets”, though
    we don’t know this for sure.)”
  • And when I tabulate Scoring Chances here, I use all Corsi events (shots/blocks/misses) from the “home plate” area.

Let’s compare what the three numbers tell us.

FORWARDS

Player CF% FF% SC% FN
SC%
Backlund 47.6 48.2 50.6 53.81
Hudler 47.5 48.4 46.6 52.88
Jones 45.1 46.3 47.8 52.48
Jooris 46.5 49.7 49.6 51.76
Granlund 41.4 43.7 44.7 51.71
Glencross 44.9 44.6 46.1 51.33
Gaudreau 46.9 48.5 46.5 50.72
Byron 46.6 47.5 47.8 48.99
Colborne 41.7 42.0 42.0 46.95
Monahan 46.0 46.0 44.5 46.92
Stajan 46.3 45.2 45.6 46.11
Bouma 41.6 43.3 42.4 45.35
Bollig 40.5 41.8 37.9 43.46
Raymond 44.8 45.8 43.7 41.46

Mikael Backlund leads the way in Corsi, Scoring Chances and FN Scoring Chances. He’s really good, doncha know? When looking at Fenwick, Backlund is leaped over by Hudler, Gaudreau and Jooris, who leads the team. If you had said that two rookies would be amongst the best Flames in driving the play and creating offense, we all would’ve been very impressed. Which we are, by the by.

On the other hand, who should hang their heads in possession shame? Bollig is the worst regular in terms of Corsi, Fenwick and Scoring Chances, while Mason Raymond dwindles in FN Scoring Chances.

Most forwards get better – a lot better in Granlund’s case – in FN Scoring Chances compared to Corsi, as they are basically Corsi close to the net. That means most of ’em can drive the net and get shots in the high-percentage areas. I’m hesitant to say this contributes tremendously to Calgary out-kicking its coverage possession-wise this season because (a) it probably doesn’t and (b) the data’s not exhaustive enough to support that claim yet.

Forwards who are worse in FN Scoring Chances versus Corsi? Mostly just Mason Raymond, who has visibly shied away from contact since returning from a shoulder injury. Stajan’s numbers are close enough together to be considered even, and considering he’s the fourth line center and used in a defensive role, it makes complete sense.

DEFENSE

Player CF% FF% SC% FN
SC%
Diaz 44.0 45.9 43.3 56.74
Giordano 47.8 47.5 49.6 53.65
Brodie 47.2 46.9 48.6 51.94
Engelland 42.0 44.1 41.3 46.75
Wideman 43.9 45.1 44.2 45.59
Smid 42.0 44.9 44.0 45.53
Russell 43.7 45.2 42.3 43.41

Diaz has played a lot less than Brodie and Giordano and against lesser competition, so everyone chill. Beyond that, everything more or less unfolds the way you’d expect it to. Diaz leads in FN Scoring Chances while Russell lags behind. Giordano leads in Corsi, Fenwick and Scoring Chances, while Engelland lags behind in all three.

SUM IT UP

Team-wide, the Flames generate FN Scoring Chances at a 47.83% rate. This compares with a 45.3 Corsi percentage, 46.3 Fenwick percentage and 45.8 Scoring Chance percentage. So no matter how you slice it, the Flames are getting out-possessed and out-chanced.

Which makes the handful of players that do drive chances and possession very valuable for the team going forward.

  • Lordmork

    Okay, so. Let’s assume we all see the value of high possession players as good hockey players. How do you make the Flames a better possession team? Is it enough to replace the obvious boat anchors (Smid, Engelland, Bollig)? Do they also need to replace guys like Wideman, Russell, Colborne, Stajan, Raymond?

    • mattyc

      Need to take into account their role/opposition, but yes.

      Obviously it’s more nuanced than “Russell is less than 50% corsi, Byron is 53%, keep Byron”, but when you look at guys in similar roles, (Bollig and Byron), and one bleeds chances, while the other doesn’t it should make you question what the other is really bringing to the table game-in game-out

      • piscera.infada

        This is where the stats mentality loses me when making these kind of player decisions. You can’t make a comparison between what Bollig or Colborne brings to the table & role & what Byron does. You will very quickly have a tiny team of players that are skilled at carrying the puck & not getting a whiff of prime scoring areas, but hey, the possession numbers are great, the luck will turn & they should start to win games.

        This stats as a whole is having trouble explaining success of teams with possession numbers that have no business having that success. Too many unmeasurable factors come into play before you start getting into player decisions of punt him & keep him because he has a 2.0 % higher corsi.
        Stick to what stats should have value, scouting & drafting, player development, assistance to player performance to help assist them getting out of scoring slumps & stuff like that.

        • mattyc

          maybe size doesn’t matter (as much as you think it does).

          Also, and maybe more to your point, if size is really what allows you to get in good scoring areas, then why do Bollig and Colborne get so many less chances, and give up so many more compared to Byron, Gaudreau, etc.

          • Lordmork

            Size doesn’t matter???? This is where we part agreeing to disagree. Tell that to Glen Sather when he utilized two of the best players in the world with Dave Semenko. Back at you, how many times has guys like Gaudreau & Byron banged it up in a corner with a big Dman & walk away with the puck? Possibly because that isn’t what their role is on the team. It’s a team game. Maybe we should look at each assist of Colborne & Byron & how it was earned.

            You see Byron as a better piece than guys like Colborne based on some Corsi numbers. Except Colborne has finish & can score on opportunities. The Kevy corsi says punt that Byron because he can’t put the puck into a 20′ wide net on a breakaway. I want guys to score when they get those scoring opportunities. How many times has Byron squandered an opportunity to put us back in a game & misses? How many shootout goals hasByron have versus Colborne. Probably because Hartley knows he can’t score on a breakaway anyways & doesn’t use him in shootouts.
            I bet you would say Martin Gelinas has unsustainable high shooting percentages if he were playing today. This argument will never end & we will not come to an accord as to what conclusions should be accepted as gospel.

          • Lordmork

            gaudreau does it all the time. size does not make a player good.

            you have a team full of colbornes, you’d never get the chances for a team full of byron’s to take because you would be in your own zone all of the time. see why corsi is important? a guy thats only getting luck on scoring gives the team basically nothing as he depends on everyone else to set it up. they can’t set it up, he doesn’t get the chance.

            colborne is the perfect example, which makes it funny you chose him. he doesn’t do anything except get lucky once in a while.

          • Parallex

            Justin, you seem so blinded by Corsi-love you just throw out garbage and think people will accept it at face value. Colborne doesn’t do anything but get lucky once in a while? Are you smoking something? At the start of the season he lead the league in assists and was perhaps the Flames’ best set-up man. Then he got injured. Lately he’s been turning it up again.

            A team of Byrons may have the puck a lot but they’d rarely score. If you haven’t learned anything watching the debacle of the Oilers the last few years perhaps you should start watching them. It’s instructive. You can’t just throw supposed talent on the ice and get results. You have to build a team. You need a variety of players. Building that team, Corsi is about the very last thing that is important for building that team….

          • mattyc

            Johnny Gaudreau wasn’t even born when Dave Semenko stopped playing. Back then, if a player was too skilled Bobby Clarke just broke his ankle. Like, beyond the fact that they still use pucks, it’s basically a different sport.

    • SmellOfVictory

      Replacing Smid, Engelland, and a couple of bottom six forwards would likely do wonders for the team. It’s painfully obvious when the Flames’ bottom pair is on the ice because they just sink anyone they’re with.

      I honestly think that, were the Flames to substitute good (for that role) players for Smid, Engelland, Bollig, and Colborne, they’d be right around a 50% CF/FF team. Not only would the possession of that pairing and third/fourth line improve substantially, but I think there’d be a domino effect; better low-end guys means less overall defensive zone starts and (maybe) a bit less of a hard matchup workload on the top end guys, which means everyone’s possession is likely to improve to some degree.

    • Parallex

      Preferably with a longterm contract extension in one hand and a pen in the other.

      On the articule I have one quibble… is a corsi event near the net really a “scoring chance”? I mean, shouldn’t a scoring chance be something that actually had a chance to score (misses and blocked shots having a 0% chance of directly scoring afterall). It’s not that I don’t think measuring possession stats in high score probability zones has value… I just think that terminology is poor. Really I think weighted possession stat analysis has got to be where advanced stats go from here now that they’ve hit mainstream.

  • Matty Franchise Jr

    Do you think that these numbers are affected by zone starts and quality of competition? It seems logical that a player getting easier ice time would have better numbers, but maybe it’s counter intuitive.

    • mattyc

      yes. Specifically, I think it hurts Kris Russell a lot because he doesn’t play with the strongest FWDs either.

      Edit: I’m wrong, Russell’s most common linemates (from BTN) DENNISWIDEMAN JIRIHUDLER JOHNGAUDREAU SEANMONAHAN PAULBYRON

      • piscera.infada

        I’m still of the opinion Russell spends most of his time covering for Wideman’s shortcomings (especially in the defensive zone). Ideally, Russell would be your #4 defenseman paired with a solid number 3. In fact, from what I’ve seen from him in his Flames career, I think he’d be great if that were the case. Instead, he’s a de facto #3 defenseman (and part-time babysitter in his own zone), and I think that hurts his numbers as well–the second the puck moves “south” of centre, he’s basically a one-man show.

        The odd time he ends up on the ice with Brodie or Giordano, he looks very good.

        • mattyc

          I want to agree with you, and that’s my gut, but why are Russell’s #s worse than Wideman? The one indication I can see, is that Wideman’s posession #s are worse without Russell than with, although that’s only 170ish minutes…

          • piscera.infada

            Frankly, it doesn’t concern me. Watching Wideman, I see the deficiencies–he virtually never actually gets the puck out of the zone when pressured, he’s not very active in the corners (both with and without the puck), he doesn’t make the smart pass first pass half the time, and he gets caught pinching too often for my liking. I’m not here to bash Wideman, as I like some of the things he does in the offensive zone, but watching the two players together on the ice, it’s fairly easy to see who’s doing what, when, where, and how.

  • prendrefeu

    Thanks for this post. I do recognize, and fully appreciate, that you (Mr. Pike) are doing the majority of writing on this site, and clearly you and other FN writers are putting a lot of time, and effort, into writing.

    That said, this was the basic conclusion:


    DOOMED!!! THE FLAMES ARE DOOOOOOOOOOOOOMED!!!

    Right? I mean, no conclusion or suggestion about what could be done to improve things in terms of a trade, change, etc:. Just that the Flames suck, apparently.

    Or maybe, just maybe… if you really open your mind to it….
    they could be just a really hard working team playing better than the stats can keep up with (and no, it isn’t ‘luck’**), a team going through a very strong rebuild period, a team with a really bright future, and one of the most exciting teams to watch all season, win or lose.

    I’m wondering what FN would look like if the Flames were #1 in the league. Would there still be posts about how the stats war is being lost? Or would the FN writers prefer a team at the bottom of the table – but gosh darn those Flames, they’re winning the stats war and that’s all that matters, right?

    I know that negativity sells. It works for your local news as much as it works for Lambert and getting ‘hits’ to the site, which translate into ad revenue…

    But… yeah… maybe some other topic might be both more interesting and unique, thereby getting those much needed site-numbers up and keeping the FN readers somewhat interested. Just a thought.

    Also, can we stop using Corsi? Its quite crap. Corgi is much better. Thanks.

    ** Re: Luck. Seriously, please stop. It’s just making any person with a brain in their skulls look pathetic when they point to ‘luck’. See the Black Swan Theory. If you’re pointing at anything as just ‘luck’ you aren’t digging deep enough. Have some dignity and just say “actually, we haven’t figured out what is going on, but clearly something is and we’re working on re-adjusting our approach to have a more comprehensive understanding of things”

    • mattyc

      an unequivocal no. you will continue to see stats and corsi. this is how the game is analyzed now. if you want platitudes and narratives go to the sun. we base our writing on the performance of the team and corsi is the most correlated stat when it comes to predicting future wins. the war is over; accept it.

      • MonsterPod

        If you’re saying the war is over and you’ll just throw out Corsi because it’s the most-correlated stat out there I think you’re just displaying your naivity and lack of understanding of the value and use of stats. What this season has proven, more than others is that Corsi explains virtually nothing, is poorly correlative to actual team success and teams focusing on it as a tool to improve (eg Edmonton & Toronto) are proving the naysayers’ points that it is virtually useless.

        Stats can be advanced and valuable but you’ve got to get good data and be open to analyzing to the level necessary not only to see the obvious trends but also to understand the anomalies. Corsi fails at that. Dig deeper.

          • Burnward

            Mattyc,

            Well your entry isn’t adding much to the discussion…

            If Corsi is supposedly so indicative to great play, and correlative to success than I guess it’s pretty clear from the article’s own numbers(above) that the Flames don’t have even a single decent D man, much less a Forward. Heck, we don’t have a single player that even reaches the low threshold of just run-of-the-mill average of 50% Corsi.

            Wow! How pathetic! (Sarcasm all). What’s really pathetic though is that some of you stats nerds actually believe it. Dig Deeper.

          • mattyc

            No one is arguing corsi should be a catch-all for who is good or not. It’s just unfortunate that your misunderstanding and objectively wrong claims pollute and inhibit intelligent discussion.

          • Aussie Flame

            I understand Corsi perfectly well, and intelligent conversation using advanced stats is exactly what I’m looking for. Unfortunately I don’t see much of it.

        • Burnward

          you assume that we will not continue to dig deeper now that corsi is where it is. you are wrong in this assumption, and if you actually understood this community the way you say you do you would have known that already. when we have the ability to track further – which we will if sportvu can figure its stuff out – we will have new stats that are better than corsi and we will use them.

      • Burnward

        War. Hilarious.

        If analytics are supposed to be a scientific look at the game, you should accept that these stats are simply a hypothesis on how winning occurs.

        One that is refuted commonly by good goaltending etc.

        Just because you can’t accept other points of view, don’t turn this into an us vs them on here. The majority here like and understand these stats but can disagree intelligently on their validity.

        • Burnward

          you seem to think that this stuff can be questioned. about 20% of it, maybe. but that’s far and away the best of anything we have in terms of player evaluation.

          I’m not sure why it is unreasonable to say that sometimes teams get luckier or unluckier than they should be. that’s what we are saying with corsi. there are other ways to evaluate, yes – but none of them seem to be as good as corsi is.

          • Burnward

            I have no problem with Corsi if it’s taken in context. I understand how it relates to winning. But if it was perfect the team with the best Corsi each night would win.

            I believe there are always reasons why a team doesn’t fit into a norm, but luck seems lazy.

            And when you factor in all the other stats where the team does well, it just feels like there’s more to offer on the stats end.

            And are you talking player evaluation or team evaluation?

            Also, when does luck exceed a sample size? And how do you adjust Corsi for different teams playing styles? I just wish the conversation gave a little more respect to those raising questions.

          • SmellOfVictory

            fine. it might be lazy, but it’s also accurate at this point in time. we haven’t been able to find what that percentage is made up of but that doesn’t mean we aren’t trying. I don’t see why you are quibbling about 20% when you could be quibbling about 50% or 60% or 100%.

            both players and teams can be evaluated using this. players need a lot more context. the team doesn’t do well in much except shot suppression but that is cancelled out by awful shot generation.

            why are “leadership” and “toughness” not lazy?

          • Burnward

            but no one can prove that and until they can I will say that they are so minute in deciding what team wins that it doesn’t makes sense to build a team around the concepts; because that’s the evidence I have.

          • Burnward

            because their sv% this season is like below .900. with an average sv (between .915-.920) they’d be 23-31 goals better and probably at least 5 wins better.

          • Burnward

            it doesn’t really matter when it comes. those 10 extra goals give a team 3-4 more points over the course of a season based simply on gf/ga. not getting those extra shots means you don’t get any of those 10: is the potential for one better than zero? to me it is.

          • Burnward

            Here’s the thing though. If you were tracking lucky goals and when they came, I could get behind it a bit more.

            But the blanket the team has been lucky doesn’t jive.

          • Burnward

            okay, but the reason they are lucky is because their overall percentage of shots that go in is higher than usual. that’s the evaluation at this point: I am assuming that the reason there is a discrepancy is because we (“stats guys”) can’t tell you why that difference exists.

            I’m sure they have “lucky” goals but they also have chances where they should have scored, right? my feeling is that the amount of lucky goals and the amount of robberies probably equals out over the course of the season, it’s the higher preponderance of normal goals that makes the team lucky. hopefully that makes sense.

          • Burnward

            I get this. However, I contend that
            the Flames create more “lucky” goals due to their continued effort to get to the dirty areas which allows for chance to take over. This comes down to heart and desire for me.

            You can fire all the pucks from the outside you want and inflate a Corsi stat to look good, but if you aren’t putting in the effort to get to areas where those shots matter, who cares?

          • Burnward

            that’s fair and logical, but every study along those lines I’ve ever seen shows that over the course of the season those teams that get in the dirty areas by and large don’t have a better shot quality than they otherwise would have.

            players who are just firing pucks to inflate corsi stats will inevitably have lower corsi stats because they have less time with the puck than other teams. pitt the past couple years under bylsma was a great example of this; lots of shots but always having to rush back and defend more than they should whenever cros and malkin weren’t on the ice and as a result a lower ov corsi than they should have had with those two putting up like 57% and 55% respectively.

          • Burnward

            if I had to guess, I’d say the breakdown is this:

            ~35% save percentage
            ~35% shooting percentage
            ~30% “unmeasurable/undiscovered”

            penalty kill and power play also have a lot to do with it but since those vary from team to team and situation to situation I wouldn’t count them.

          • Burnward

            Though, you can’t discount it because it is part of the game. Calgary adds to their ability to win because they draw way more penalties than they take.

            If they had a better PP, they might be 5-6 points higher in the standings.

          • Burnward

            sorry, that’s not what I meant – they are obviously very important to the team and the ability to win. I just wouldn’t put them in the “luck” category because I believe pp/pk are skill based much like the ability to possess the puck. the sh% and sv% are basically 50/50 (i.e. two players to make a play) whereas pp/pk takes a lot more than that.

          • Burnward

            Right. So if Calgary plays more of a counter style, with disciplined play as a backbone to balance out a lack of ability from their forwards to play that possession, cycle game to increase Corsi…why is that luck?

            I consider that intelligent use of the assets at their disposal.

            And then when you look at actual shots allowed, they aren’t too bad at all.

            How can one stat reconcile all the variables?

          • Burnward

            All I’m saying is that Corsi is a good stat, but please don’t think that because some of us don’t think it’s the greatest thing in the world we’re idiots.

            Thanks for the conversation though…and I think we can all agree on this…GO FLAMES GO!!!

          • Parallex

            That would be a whole lot easier if most of the folk that “don’t think it’s the greatest thing in the world” (and just to be clear I’m not talking about you, or even just commentators here, I mean the online fandom at large) would…

            A: Put the work in instead of just crapping on the people who do put the work in, by which I mean creating, testing, and presenting alternative measures

            B: Didn’t whine about it constantly,

            C: Weren’t basing it in blatant raging homerism.

            Yes, folk ought not get treated like idiots… but with so many of those people (again not you) doing their best Stan Gable impersonation (yeah that’s right… I’m pulling out a 30y.o. pop culture reference) sitting behind their computers (probably in letterman jackets) going “NERDS!” doesn’t help.

          • Burnward

            Hahaha! No offense taken. I understand your point.

            There is a lot of that out there, but I like to think this place is a much more understanding venue for these discussions. And I really hope that legitimate conversations continue to move these new analytics forward. They’re great, but still in their infancy I think we can agree.

            Now where’s my “Members Only” jacket….

          • Parallex

            Not really in infancy anymore. When many of it’s proponents are being publically snapped up for employment by teams, the NHL has talked about adding them to their stats page, and it has gotten the attention of the MSM (beyond guys like Steve Simmons and Mollberg hate raging on it) I think it’s fair to say it’s out of it’s infancy.

            I don’t think it’s not a fully formed adult yet (to continue on with the human development metaphor) but it’s now well past infancy. I’ve heard sabermetrically inclined baseball writers/commentators say it’s about 10 years behind baseball… hope we see someone publish Moneypuck sometime in the not to distant future.

            Edit: If anyone here hasn’t already read Moneyball I highly recommend it. Don’t just watch the movie the book is way better.

          • Burnward

            I can agree with that. I wonder if it’s even possible though.

            Hitter vs. batter. Nothing really changes except for a defensive shift maybe and the pitcher’s selection?

            I get the idea behind bringing hockey to that level, but with so many moving parts it’s a tough thing to do. Interesting exercise though.

      • Aussie Flame

        Actually, I think you will find Fenwick close is the most correlated stat when it comes to predicting future wins. Before you make smart remarks espousing how good these stats are, I think you should actually attempt to understand them.
        To my eyes, the biggest issue people have with these stats are not actually the stats, they are all simple and make sense, it is the people who endorse them.

    • mattyc

      ** Re: Luck. Seriously, please stop. It’s just making any person with a brain in their skulls look pathetic when they point to ‘luck’. See the Black Swan Theory. If you’re pointing at anything as just ‘luck’ you aren’t digging deep enough. Have some dignity and just say “actually, we haven’t figured out what is going on, but clearly something is and we’re working on re-adjusting our approach to have a more comprehensive understanding of things”

      I forward a FN motion that we replace the word ‘luck’ with destiny.

  • MonsterPod

    I am trying with this stuff, but I still can’t wrap my head around it.

    Brandon Bollig has low Corsi and Fenwick, etc. meaning he doesn’t shoot much, whether his shots hit the net, miss the net, or get blocked. Okay, got it.

    But Bollig is a fourth line fighter, hitter, visual deterrent — or a least he’s supposed to be. I’m not surprised his advanced stats suck and I’m not bothered by it, because it is not his role/job to generate offense.

    Last year everyone was cooing about Backlund’s Corsi/Fenwick, meaning he had a lot of shots and attempts and even scoring chances. And then he ended up with 18 goals.

    Monahan had apparently poor Corsi/Fenwick and an abnormally high shooting %, and he was poo-pooed here and by some wanker on OilersNation. Yet, he ended up with 22 goals.

    What this shows me is that Monahan is simply a better shooter than Backlund. Monahan has better finish than Backlund. This has been the case again this year. See shootouts and breakaways for further evidence.

    Where do these stats leave the playmakers? When Johnny G gets the puck and cycles around the net, and bobs and weaves, and puts it on Hudler’s stick, and Hudler shoots the puck — whose Corsi goes up?

    When Gretzky played for the Kings, and he would put up about 30 goals and 150 assists in a season, would his Corsi/Fenwick have stunk then?

    When a team goes up 3-0 in the first, and for the next two periods just gains the red line and dumps it in, but then wins 3-1 (hooray!), would their Corsi/Fenwick stink at the end of that win?

    It seems that, in theory, Player X could be trigger happy, play with good playmakers, shoot the puck all the time, and rarely score because his shot is a muffin. But his Corsi/Fenwick would be awesome and the Moneyball guys would be singing his praises, no?

    • Burnward

      you are drawing the wrong conclusions from the data. higher sh% does not mean “better” shooter, especially when we only have one year’s worth of data. monahan is shooting 12.9% this year – how do you explain that?

      you people focus too much on the number and not enough on the context – corsi explains possession. it is irrelevant how good the corsi chance is because it doesn’t matter – study after study have shown us that over the course of a season all “shot quality” basically equals out. teams that have the puck more often than not are good teams. corsi shows us that.

      every player has the opportunity to be a positive corsi player. unless you are starting a disproportionate amount of shifts in the d-zone or against the best players in the world you should have a positive corsi rel. bollig doesn’t do that and from my experience the other stuff he brings to the table is worth jack squat. you can disagree with me but you can’t disagree with the data that shows us the same results over and over.

      • MonsterPod

        I wasn’t saying a high shooting % makes for a good shooter. I remember people were criticizing Monahan for it last year, saying it was unsustainable.

        And I’m not crapping all over advanced stats. I come here for exposure to them. I heard Kent interviewed by Steinberg the other day and it’s all good. I’m glad NHL.com is creating a page for it.

        I’m just trying to figure out how shots (missed, hit, blocked, or otherwise) explain ‘possession,’ especially in terms of the examples I gave of Gaudreau and Gretzky.

        And what about when you control the play, own the zone, cycle the puck for a minute but only get one shot off? Then the other team comes down, and fires off two quick shots. Who gets the higher Corsi?

        To my mind, shots don’t entirely explain possession. And it seems clear that if the Flames want to raise their Corsi/Fenwick and overall possession, they need to stop being one of the worst faceoff teams year after year.

      • MonsterPod

        Okay, (dot dot dot) so you want to make it even more cryptic.

        “Corsi percentage is the percentage of all shots, blocked shots and missed shots towards the opponent’s net, no matter where they’re taken.”

        If it’s not this — what the article has led me to understand — then what is it?

        • Parallex

          For starters you’ve missed that it’s not a “sole source” stat… meaning when a Corsi event occurs credit (or debit) isn’t accumulated by just one person.

          To use one of your quarries…

          “Where do these stats leave the playmakers? When Johnny G gets the puck and cycles around the net, and bobs and weaves, and puts it on Hudler’s stick, and Hudler shoots the puck — whose Corsi goes up?”

          … the answer is everybody on the ice for the team that is attempting the shot when that occurs (opposingly everybody on the ice for the other team when that happens has their Corsi go down).

  • MonsterPod

    I like reading these articles. They’re kinda like the slap in the face we all need to keep our excitement in check. Don’t get me wrong, I’m loving what they’re doing and really hope they make the playoffs and get caught up sometimes too, but we gotta understand that they are playing well “against odds”.

    That said, I think this season is more or less understood…a hard working team that is somehow capitalizing on chances at a really high rate and playing above their expected level.

    What I’d be really interested in seeing is how the Flames project for next year. With the progress of the rookies and steady play of veterans, is it likely the Flames gonna be out possessed next year? It seems like they need defence upgrades to improve next year but what if we get stuck with the same old? Or maybe add Wootherspoon… Do the Flames continue to get out chanced as badly as they are?

  • Lordmork

    I see people questioning stats and I feel as though they should read this. Last paragraph is most relevant.

    Why Do Many Reasonable People Doubt Science? – National Geographic

    “The trouble goes way back, of course. The scientific method leads us to truths that are less than self-evident, often mind-blowing, and sometimes hard to swallow. In the early 17th century, when Galileo claimed that the Earth spins on its axis and orbits the sun, he wasn’t just rejecting church doctrine. He was asking people to believe something that defied common sense—because it sure looks like the sun’s going around the Earth, and you can’t feel the Earth spinning. Galileo was put on trial and forced to recant. Two centuries later Charles Darwin escaped that fate. But his idea that all life on Earth evolved from a primordial ancestor and that we humans are distant cousins of apes, whales, and even deep-sea mollusks is still a big ask for a lot of people. So is another 19th-century notion: that carbon dioxide, an invisible gas that we all exhale all the time and that makes up less than a tenth of one percent of the atmosphere, could be affecting Earth’s climate.

    Even when we intellectually accept these precepts of science, we subconsciously cling to our intuitions—what researchers call our naive beliefs. A recent study by Andrew Shtulman of Occidental College showed that even students with an advanced science education had a hitch in their mental gait when asked to affirm or deny that humans are descended from sea animals or that Earth goes around the sun. Both truths are counterintuitive. The students, even those who correctly marked “true,” were slower to answer those questions than questions about whether humans are descended from tree-dwelling creatures (also true but easier to grasp) or whether the moon goes around the Earth (also true but intuitive). Shtulman’s research indicates that as we become scientifically literate, we repress our naive beliefs but never eliminate them entirely. They lurk in our brains, chirping at us as we try to make sense of the world.

    Most of us do that by relying on personal experience and anecdotes, on stories rather than statistics. We might get a prostate-specific antigen test, even though it’s no longer generally recommended, because it caught a close friend’s cancer—and we pay less attention to statistical evidence, painstakingly compiled through multiple studies, showing that the test rarely saves lives but triggers many unnecessary surgeries. Or we hear about a cluster of cancer cases in a town with a hazardous waste dump, and we assume pollution caused the cancers. Yet just because two things happened together doesn’t mean one caused the other, and just because events are clustered doesn’t mean they’re not still random.

    We have trouble digesting randomness; our brains crave pattern and meaning. Science warns us, however, that we can deceive ourselves. To be confident there’s a causal connection between the dump and the cancers, you need statistical analysis showing that there are many more cancers than would be expected randomly, evidence that the victims were exposed to chemicals from the dump, and evidence that the chemicals really can cause cancer.”

    http://ngm.nationalgeographic.com/2015/03/science-doubters/achenbach-text

    • Burnward

      So what’s your point? Seems to me we live in an environment where multi variables impact behaviour at any random given time. So in the case of the cancer, it could be the dump but that doesn’t explain why every male didn’t get cancer living by the dump. There was probably a genet variable as well.

      So back to hockey. If this multi variable concept is consistent, one must assume that other players Corsi has a drag on another players Corsi who happens to be on the ice at that random moment in time. How is that factored in?

      So Justin, if you say that the stats & Corsi is the best correlated way to predict wins, may I ask how accurate are those predictions based on these Stats? Can you predict 75% of the games based on the teams relative corsi? 50%? 25%? What can you openly say you can predict more efficiently & accurately using the stat advantage as compared to just a Joe schmuck hockey fan going on gut instinct & how many times they went to the bathroom the night before? More specifically how much more can you predict wins over and above what random guessing can do?

      • MonsterPod

        The other thing is that this uses the rearview mirror to look through the windshield.

        Teams can grow, improve, make trades and changes. So Corsi stats from October may not predict future wins in March, especially if a team like the Flames works hard to improve in the dot and/or makes a trade to bring in a faceoff specialist.

        I would think this is especially true of a young team like the Flames.

        Also, is a team’s Corsi like a goaltender’s save percentage? Capable of big undulations at the beginning of the season, but small ones at the end due to sample size, even if there are marked improvements?

        • Parallex

          & really isn’t that the real hard part is that each NHL team is constantly changing to improve & therefore stats will always be a snap shot on a small sample size because of this. It can get silly if you take this to any extreme. I like stats, I would much rather like to see how these stats can be utilized into models that can improve a hockey team.

          • mattyc

            I like stats, I would much rather like to see how these stats can be utilized into models that can improve a hockey team.

            acquire better possession players, find guys who are undervalued because they’re not getting the bounces, opt for the Paul Byrons over the Brandon Bolligs of the world; the Backlunds and Grabovskis over the Colbornes and Bozaks.

      • Burnward

        between 75-80% correlation. that’s honour roll level and nothing else (except for other shot-based metrics, of course) comes close. iirc goals are between 55-60%. that is huge.

  • Especially relative to the other major sports, advanced statistics and their implications are extremely young in regards to hockey. It’s important to think critically about what these stats can tell us and what their limitations are because otherwise, we just spin our wheels.

    I’m pretty excited about the next couple years, because we’ll have more (and better) data available and be able to manipulate things in a much more detailed and interesting manner.

    • Burnward

      Absolutely. It’s going to be tough to match what baseball has done due to the moving parts and more unpredictable nature of this game. But it is fun watching these things evolve.