Visualizing Corsi and the Flames

Visualized Corsi provides us with an interesting method to evaluate players’ overall performance, relative to the league and each other. If you need a primer, go here, otherwise FlamesNation gets into the Corsi Visualization craze past the jump.

All of these guys have appeared in at least 10 games. Line mates without percentages have played less than 60:00 at even strength with the player in focus, and thus not counted. The darker the red, the better the results. The darker the blue, the worse the results. Players are sorted by even strength TOI/G. Due to injury, Giordano’s sample size is limited but he looks damn good comparatively. All of the data came from Extra Skater and Hockey Analysis. Thanks to the folks at Arctic Ice Hockey for getting everything started.

The first graph shows the standard visualized Corsi chart – under 50% is blue, over 55% is red and everything in between is orange.

You will immediately notice there’s something missing – the colour red. Well, the Flames have no one who exceeds the 55% mark. Glencross and Wideman are the closest, at 54.4% – but that’s not good enough. Also, notice what happens to Mike Cammalleri – he fades into a sea of blue here, but as we move downwards the picture will improve.

As a fan, you’d hope that the coach plays his best players the most. But as we see here, that’s not exactly happening. The defensive minutes are correctly distributed (although this chart suggests Butler and Smid still play way too much), but the forward minutes are not. Ideally, the top left should be a haven of orange (and, usually, red) with the blue being confined to the bottom right. Backlund has been getting more ice time lately, though, so perhaps the problem has been identified.

The second graph shows Corsi relative to the team’s total Corsi percentage – 43.9%. Red is above the total, blue is below the total.

As the Flames’ overall Corsi is terrible, there’s a lot more red here than there should be. Remember how I mentioned Cammalleri above? Well, here he’s all red instead of all blue. There are a lot of positives, as this shows us just how good the top four defensemen are. Although, Stempniak and Wideman’s 42.1% seems out of place, as does Brodie and Glencross’ 38.9%. I don’t know how the percentages are so low, but it seems odd.

Monahan, David Jones, Bouma, and Colborne look really bad here.

This last graph shows Corsi percentage relative to the rest of the team – each colour covers 10 percentages.

This graph sort of gives us a little more context on where each pair is relative to another. Mike Cammalleri and TJ Galliardi have been the best forwards by this measure, while Mark Giordano is the best defenseman. The worst forwards are Monahan, Colborne, Bouma and McGrattan. The bottom three defensemen continue their reign of suck. Shane O’Brien has 5 of the 10 worst Corsi percentages under his name – that’s impressive.


I question how much better this team would be with 3 bottom-tier defensemen that didn’t bleed shot attempts like they were coming out of a sliced subclavian. The three of them have 0 combined positive Corsi percentages where they were driving the bus. Same with the bottom-tier forwards – Colborne, Bouma, McGrattan, Monahan and David Jones. Certainly circumstance can explain some results (Bouma, Galliardi) but in other cases the numbers are too low to fault anything except for the player.

The problem with this team isn’t that they can’t hold third period leads or that they aren’t effective on special teams. The appearance of both of those problems stems from the fact that they can’t control shot attempts in any meaningful way when Backlund, Glencross and two top-4 defensemen aren’t on the ice. This team simply isn’t very good at controlling the puck; in fact, they’re historically bad now. Add that to worse than replacement level goaltending on 60% of the team’s shots against and you have quite the disaster on your hands.

  • mk

    Ugh – now the team sucks by all measures. You guys really shouldn’t post this stuff – other teams might read it and not want our players.

    There should be 2 sites: 1 for Flames fans for real info and then another one with “massaged numbers” for the general public. “Look at all these great players that we have that just stuck in a bad situation! You want to trade for them by using your great prospects…”

    Don’t take that the wrong way – I really value the information presented here, you do a great job.

  • ChinookArchYYC

    JA, you’ve done a nice job in illustrating and explaining these results. Thank you.

    A few of observations. Backlund has done incredibly well considering he had to work with Dancing Bear and Tim Jackman for about 10 games. In fairness to Hartley he’s been forced to play Smid and (especially) Butler because he has no one else to eat up minutes, with injuries to Wideman and Giordano. Finally, OMG is O’Brien bad. Last night I cringed every time he had the puck. He’s a turnover waiting to happen. So ya, the numbers only verify what is plainly obvious to the eye. SOB is terrible, and is a considerable downgrade from Cory Sarich (who is also very bad.)

    On the bright side, I really like the top 4 on this team. Their mobile with some offensive threat and each of them is defensively sound. Definitely something to build upon./p>

  • Skuehler

    I think we all see where this is pointing…we need to get rid of Backlund. He just doesn’t fit with what we’re trying to accomplish here. We need more truculence.

  • Burnward

    Honest question. How seriously do you guys take CORSI? Is it a science to you, or just a nice piece of information to fold into your evaluation of players?

    • McRib

      I think everyone at FN takes it fairly seriously, but even if you disagree with advanced stats at least FN is trying to present a topic to readers that is unique from the same crap we get from most news media day-in-day-out. I personally am warming to it and think it tells a unique story in certain cases like Backlunds for example, where a player is over achieving with the situation he has been dealt or in situations where two players are both producing offensively you tend to see who is driving the play between the two, sometimes it will surprise you as its not always the big name player. Advanced Stats really took off two years ago when all devotees of the method called that Minnesota’s season would collapse because their numbers didn’t add up and they did weeks later. Similar to what the AS followers predicted from Toronto this season after their strong start and they seem to be right, it does have legitimacy even if it is not the end all and be all for some.

      • Burnward

        Cool. Always interested to see how people are viewing these new stats as they emerge. I posted this on the other forum…but here’s my thoughts. Sorry for the re-post.

        I appreciate CORSI et al…but to me, the real value of it is to evaluate players that are 3rd-4th liners.

        Top six guys you can measure by watching play a lot and looking at their “real” stats.

        It’s the bottom six guys that I find this most useful for.

        I have yet to see this stat change my mind about a player, but I have taken notice of lesser players because of it.

        Oh, and it still has the Mike Weaver effect going for it. So there’s that.

    • SmellOfVictory

      Very seriously. It provides a grounding effect for our own perceptions of players. It’s a lot rarer for a player to look good via Corsi but actually be terrible (it can happen, though) than it is for a player to look good to the eye but actually be terrible.