Calgary’s individual Corsi differentials are a mixed bag

If you’ve been reading this site for very long, you’ve probably figured out that we value shot metrics like Corsi in evaluating players. That’s because goals are rare in hockey and to determine how much a player actually drives play – and helps their team win – we have to rely on more common events like shots, missed shots and blocked shots (otherwise known as Corsi events).

One simple way of looking at a player’s performance is through their Corsi differential: subtracting their on-ice Corsi events against from their Corsi events for at five-on-five. We’ve collected Corsi differentials for most of the Flames regulars for each of the past three seasons, along with their offensive zone start percentages to give some context to their numbers.

Alex Chiasson

Two years ago in Ottawa, Chiasson was given 53% offensive zone starts and performed rather well, with a +24 Corsi differential (+0.32 per game). Last year in Ottawa, his zone starts were cut back dramatically (to 38%) and his differential crashed to -181 (or -2.35 per game). This season he’s been primarily riding shotgun to some combination of Sean Monahan, Sam Bennett and Johnny Gaudreau, so he’s gotten 62% offensive zone starts. To nobody’s great shock, his differentials are positive at +8 (or +0.21 per game).

If nothing else, Chiasson seems like a carryable third body on a line.

Dennis Wideman

Wideman has generally received the same zone starts over the last three seasons, spending about half his time starting in the offensive zone. Despite his unchanged circumstances, his numbers have quietly gone from horrendous (-412, or -5.15 per game, in 2014-15) to merely below average (-39, or -1.22 per game, in 2016-17). It’s worth noting that he’s spent a large portion of this season with T.J. Brodie, which undoubtedly has helped things.

Deryk Engelland

Like Wideman, Engelland’s numbers have quietly improved year over year while his circumstances (zone starts between 39% and 46% in the offensive zone) haven’t really been altered much. He’s gone from -364 (or -4.79 per game) in 2014-15 to -76 (or -1.95 per game) this season. Unlike Wideman, Engelland doesn’t seem to be getting a ton of help with partners this season.

Dougie Hamilton

During his last season in Boston, Hamilton was a monster: 47% offensive zone starts and +221 Corsi differential, good for +3.07 per game. Granted, he played a lot with Zdeno Chara. When he moved to Calgary, he seemed to take some time adjusting. In similar circumstances, he was -49 last season (-0.60 per game). This season, playing primarily with Mark Giordano, he’s already at +108 (+2.84 per game). Last season seems to have been a blip.

Johnny Gaudreau

Gaudreau has been given a steady diet of offensive zone starts during his NHL career: 56%, 55% and 66% (so far this season). He has, however, been consistently a poor possession player, with a -142 differential in his rookie year (1.78 per game) and improving somewhat to -19 last season (-0.24 per game). So far this season, he’s at -23 (-0.79 per game). He’s been keyed-in on by opposition lines since last season, so it’s a bit concerning to see his numbers degrade a bit.

Jyrki Jokipakka

Jokipakka had a ton of offensive zone starts during his time in Dallas – 57% in 2014-15 and 56% in 2015-16 – but was used a lot more in the defensive zone in Calgary, with zone starts fluctuating around 48%. He went from being about an even player in differentials (between -6 and +7, or about -0.12 to +0.18 per game) to being deep in the hole. He was -34 in 18 games after last year’s trade deadline (-1.89 per game) and is -79 this season (or -2.93 per game). In short: he’s bleeding shot attempts, and the bleeding seems to be getting worse.

Kris Versteeg

If you want to see a player that seems like his numbers are being crated by his linemates, it’s Versteeg. In Chicago in 2014-15, he had 67% offensive zone starts and a +126 differential (+2.07 per game). In Carolina in 2015-16, he had 59% offensive zone starts and an even better +211 differential (+3.35 per game). During his short stint with the Kings last season (just 14 games), his numbers actually got a bit better with identical zone starts: he amassed a +3.79 differential per game and a +53 total differential. That’s crazy!

This season? With 54% offensive zone starts, he has a -53 differential (or -1.96 per game). He’s played primarily with Sean Monahan and Troy Brouwer. Given his past results, it’s hard not to blame his linemates.

Lance Bouma

Bouma’s a bottom-six forward and has historically been buried in the defensive zone. Given those circumstances, his possession numbers are deep in the red. In 2014-15 he had 35% offensive zone starts and a -288 differential (-3.69 per game). Last season he had 47% offensive starts and a -68 differential (-1.55 per game). This season he’s gotten just 24% offensive zone starts, the worst of any regular, and his numbers are very close to last season’s with a -34 differential (1.62 per game). Bouma’s a good example of a checker being given rough deployments in order to give scorers the high ground. Given his circumstances, his numbers aren’t awful.

Mark Giordano

We’ve often praised Giordano. We will again here, because he’s converted some consistently middling circumstances into some nice Corsi differentials. In 2014-15 he had 43% offensive zone starts and a -52 differential (-0.85 per game). He was given slightly better starts in 2015-16 (47%) and bumped it up to a +17 differential (+0.21 per game). So far this season he’s had nearly identical starts (48%) and a +74 Corsi differential (+1.90 per game). Granted, he’s been given good defensive partners over the past couple seasons, but he’s facing the best opposition lines and starting a decent amount of his shifts in the defensive end and still producing good possession numbers.

Matt Stajan

Over the past two and a half seasons, Stajan has started two-thirds of his shifts in the defensive zone fairly consistently. In 2014-15, his differential was -129 (-2.19 per game). In 2015-16, his differential was -196 (-2.45 per game). This season? Playing in the same circumstances with many of the same people as years previous, he’s +19 (+0.49) through half the year. That’s a big, impressive turnaround.

Matthew Tkachuk

Tkachuk has nothing to compare to as a rookie, but he’s gotten 35% offensive zone starts and has produced a fantastic +119 Corsi differential (+3.37 per game). That’s the best differential on the team among all of the regular skaters and very impressive for a rookie.

Michael Frolik

In Winnipeg during his last season there (2014-15), Frolik was given mostly (54%) offensive zone starts. He killed it with a +203 differential (+2.48 per game). He’s gradually been transitioned into a second Mikael Backlund in Calgary: last season he had 48% offensive starts and a +45 differential (+0.70 per game) and so far this season he has 38% offensive starts and a +86 differential (+2.21 per game). Last season seemed to be an adjustment year for him, but he’s been superb at puck possession this year.

Micheal Ferland

Ferland has quietly been progressing really nicely into a rock-solid possession player for the Flames’ bottom six. As a rookie in 2014-15, he had 45% offensive starts and a -43 differential (-1.65 per game). Last season he was a full-time NHLer with 43% offensive starts and had a modest -33 differential (-0.46 per game). So far this season, he’s had 41% offensive starts and a +13 differential (+0.33 per game). He’s not a high impact possession player (yet), but he’s been starting in the defensive end and moving the puck in the right direction.

Mikael Backlund

Backlund’s long been considered Calgary’s possession beast. In 2014-15, on a Flames team that was awful at puck possession, he had 38% offensive zone starts and a -104 differential (-2.00 per game). He improved quite a bit with slighty better zone starts (49%) last season, posting a +80 differential (+0.98 per game). This season? He’s been buried in the defensive zone (37% offensive starts) but has a great differential of +67 (or +1.72 per game). He’s progressed nicely, which is probably a product of finding him linemates. In 2014-15 he was dragging people around, last season he and Frolik were dragging a third guy around, and this season he, Frolik and Tkachuk don’t have to drag around anybody.

Sam Bennett

Throughout his one and a half seasons as an NHL regular, Bennett has been used consistently and has performed fairly consistently. He’s had 54% offensive zone starts in both seasons. Last season he had a -48 differential (-0.62 per game) and this season he has a -39 differential (-1.00 per game). He’s not bleeding shot attempts, but he’s also not really driving play yet. That said: he’s 20, and at least he’s been consistent thus far.

Sean Monahan

We’ve written a bunch about Monahan’s possession game, and he’s frequently one of the worst possession players as reflected in game recaps. In 2014-15, he had 47% offensive starts and a -200 differential (-2.47 per game). Hartley gave him a bit more high ground last season (54%), which improved his numbers slightly to a -78 differential (-0.96 per game). This season Gulutzan’s given him even more high ground (60%) and his differentials are very similar: -43 (or -1.10 per game). He’s struggled this season, as based on his history you would’ve expected his numbers to creep up slightly given his increased offensive zone starts.

I wouldn’t panic yet, but the lack of progress in his numbers isn’t a good sign.

T.J. Brodie

For how good a puck-mover Brodie is, his puck possession numbers have been fairly poor over the past few seasons. In 2014-15, buried in the defensive end with 40% offensive zone starts (and playing on a bad possession team), Brodie had a -254 differential (or -3.14 per game). Over the past two seasons, he’s had 48% offensive starts. Last season he was at a -34 differential (-0.49 per game) and this season it’s a -52 differential (-1.33 per game). Some of that has to be a difference in partners – he played with Giordano last season and mostly Wideman this season – and he’s facing tough competition, but his numbers aren’t terribly encouraging.

Troy Brouwer

Brouwer’s played for three teams in three years, so comparisons are somewhat challenging. In 2014-15 in Washington he played mostly offensive zone shifts (55% offensive starts) and had a modest differential of +25 (+0.30 per game). Last season in St. Louis, with a lot of defensive starts (40% offensive starts) he had a modest negative differential of -21 (-0.26 per game). Not bad. So far with Calgary he’s had 53% offensive starts…and a woeful -94 differential (-2.61 per game). Is it the system? Is it his linemates? Is it a general drift in his play? Or many of these factors in concert? Either way, considering his track record to this point, Brouwer’s results have been a disappointment.

  • OKG

    Re:Brouwer it’s the fact that we’re not a stacked veteran team thst can carry him.

    Anyways Engelland has had a lot of help to keep his corsi above water. Look at his Kulak WOWY

  • Brodano12

    Monahan was above 50% before he and Brouwer were put together. Brouwer was also doing better before being put with Monahan. It’s not the fault of one of them, it’s just that those two don’t play well together.

    Monahan’s underlying numbers improved every year since his rookie year, including this year up until GG decided to staple Brouwer to his line. I can see a Gaudreau-Monahan-Versteeg/Ferland line doing quite well possession wise.

  • Just.Visiting

    Please, please, please try Tkachuk with Monahan and Gaudreau on at least the PP to see if they have chemistry and if one of JG or MT can play comfortably on the right side. Then take it from there as to whether you break up the three M line and shift Ferland on to it instead of MT.

    If it works, you have two very strong lines, a dependable fourth line of Stajan-Bouma-Chiasson/Hathaway and then just need to figure out how to get the Sam line moving.

    • Parallex

      This is Gulutzan we’re talking about the only guy allowed to play his offside is apparently Frolik.

      Regardless I was more thinking that they ought to put Gaudreau with Backlund and Frolik, Tkachuk with Monahan and Versteeg, and have the remaining top 9 line be Ferland/Bennett/Brouwer. Put the two big production + poor possession guys (Gaudreau & Monahan) with guys that have historically (or at least presently in the case of Tkachuk) had good possession.

    • KiLLKiND

      You are part correct, in that corsi is flawed and should be outdated especially when compared to baseball who realized years ago advanced stats can be really helpful. It is, however, a good way for most consumers who don’t have better or more advanced measures and is very easy to understand. One of the flaws is that it doesn’t differentiate between high danger scoring chances. Another flaw is that corsi events do not evaluate the impact of the player on the corsi attempt simply whether he was on the ice for it.

      Corsi does have shot quantity going for it, and over time it shows through enough corsi events which players drive play and which do not. Corsi is also a very simple tracking measure for most stations to be able to implement during the game and a simple enough statistic for possession that can help provide some insight into how certain players, teams, or even lines operate. I am a fan of corsi to measure players and teams to some extent and having more information is never a bad thing. If that information is flawed and you don’t either accept the flaw and take the information with a grain of salt, then it can still be useful just not as useful as some believe it to be, it is also not bad enough of a statistic that it will be removed, especially when you realize corsi is just +/- on steroids. Through enough time corsi does reveal telling information about players, teams lines, and can be used by coaches to know the enemy better and to better implement their own team. I also would hope that the Flames have more information than just corsi to make their decisions.

      Is corsi flawed? Obviously. Does Corsi have usefulness? Yes. Will it be completely outdated and irrelevant within 3 years? I highly doubt it, I think some new statistics mainly those focused on high danger scoring chances and chances that the player has an impact on will start being used. Is that enough to get rid of a very new statistic that tv stations just started mentioning? Absolutely not. Think about how long it took for corsi to get where it, now if you think a bunch of new completely superior possession measuring statistic will override corsi in 3 years, I either want your insider information or your overly optimistic view on life.

      • The Fall

        Thanks for taking the time to respond like this. I agree — with one exception. People do not need a “more advanced measure”. We need to be discussing more effective measures, not simply more advanced measures. The most effective stat is shots. And when shots are expressed as a percentage for (SF%) it most accurately predicts success.

        SF% has a direct linear relationship with goals for (GF%). Corsi on the other hand is all over the map when predicting goals. Corsi introduces noise; which is ironic because most people assume it provides a deeper and more accurate picture.

        The Corsi should be one of the last and finest details of a player or team evaluation. Shots on goal shot be much more heavily weighted over shot attempts.

        Logically this holds up that shots are more accurately predictive than shot attempts… Like goals are better than goal attempts, hits are better than hit attempts, face-off wins are better than face-off win attempts…

  • dontcryWOLF88

    No statistical metric is perfect. That doesnt mean the information isnt useful.

    I agree with you to the extent that people can become blinded into thinking it is the ONLY important way to measure a players usefulness.

  • Dan the flames fan

    I struggle to understand that Brodie tends to be given players to “help bolster”; so is it any wonder why his possession rates suffer. If I have any issue with TJ, it’s that he doesn’t shoot near enough. He is smart and creative, but doesn’t allow himself the confidence to be assertive.

  • Eggs Bennett

    Just a suggestion for future posts to put things into context: if you’re going to present ozs in %, please also show corsi in %. This adds a bit more context so you can compare player to player without the bias of differing TOI

  • jupiter

    You roll off a players corsi numbers for the last 3 years and then finish with the statement “Last year seems to have been a Blip”

    Why is last year a blip? How do you know which information to ignore?