FGD #35: The Panic Button (8pm MT; CBC)

The Calgary Flames (17-15-2) have lost 7 games in a row. Their opponents tonight, the Vancouver Canucks (), have lost 5 games in a row.

A streak will end tonight.

I really hope it’s Calgary’s streak, because I am sick of writing the same “The Flames played well but something went bad and they lost!” game story. You can try and dress it up, but fundamentally we’re experiencing the exact opposite of October and November, and it’s just more fun to analyze wins than it is to dissect a string of very similar losses.

Is it time for us to crack each others’ heads open and feast on the goo inside? We’re not sure. We hope not. But with a loss against a reeling Vancouver Canucks team that is having a similarly tough go of it, it may be time to hit the proverbial panic button.

(For a view from the other side, visit our chums over at Canucks Army)

THE FLAMES

They’ve lost seven in a row. Likely we’ll see some line-up changes, but they’re not skating today. Keep an eye on Daily Faceoff, who have last night’s lines (seen below) and will have updated rosters as the day goes forward.

Screen Shot 2014-12-20 at 12.12.24 AM

Screen Shot 2014-12-20 at 12.12.38 AM

The Flames were alright last night at home, but now they’re (a) tired, (b) dejected and (c) playing a Vancouver Canucks team that has had a lot of success against them at Rogers Arena AND will be desperate to get back in the win column. As for the Flames – they’re just desperate to get some points on the board and turn this streak around before it hits Oilers territory.

I’d expect Karri Ramo to start, but that’s just me guessing based upon back-to-back games.

THE CANUCKS

Vancouver has lost six straight and have arguably been worse than the Flames during that stretch. Here’s lines from their last game via Daily Faceoff.

Screen Shot 2014-12-20 at 10.19.02 AM

Screen Shot 2014-12-20 at 10.19.16 AM

I would expect Ryan Miller to start, but that’s a guess.

These teams last met in the season opener in Calgary, but this is their first meeting in Vancouver this season. On-paper, the Canucks are the superior team. However, the Flames have surprised them before.

SUM IT UP

Via Hockeyfights.com

Nobody wants to enter the Christmas Break on a losing streak, but it’s likely one of these teams will. Calgary has lost seven straight. Vancouver has lost five. This is the opposite of a clash of titans. This is two desperate teams clashing for points that they both really, really need.

I expect a tight-checking game, filled with two nervous groups desperately trying not to mess things up for themselves.

Care to see what those Canucks fans are saying about tonight, click here.



  • beloch

    Remember how many have been saying the Flames were due to regress? This isn’t regression. It’s an atomic face-plant of over-regression.

    Flames “Luck” stats:

    ——————–sh%—–sv%—–PDO
    First 26 games – 11.9 91.2 103.1
    Last 7 games — 4.91 86.2 91.1
    Overall avg —— 10.5 90.1 100.6

    Over the last 7 games, the Flames shooting percentage was less than half of what it’s been over the season and the goal-tending has been MacBackup bad. These 7 games alone have brought the Flames average PDO for the entire season down to nearly 100. The Flames are now well within the normal range for that stat. No team should expect a PDO of 91.1% to last. Hiller and Ramo are not 0.862 goalies. This team is not going to keep shooting less than 5%!

    To sum up, the Flames have been so horrifically unlucky over the last 7 games that it has effectively cancelled out the good luck they had over the previous 26. The Flames are not flying above 0.500 on borrowed karma any more. They have received full karmic retribution over the last 7 games and are no longer in debt to the hockey gods. It’s time to start winning again.

    • Hacker

      This is the problem with analytics. It’s all based on regression to the mean. So we are told not to look at small sample sizes during a season.

      BUT from a macro level, the length of time they have been tracking advanced stats is in itself a small sample size. Ironic. Is 1 season long enough to average out? Might it take 5 years to find the statistical average with no luck? This would essentially make the stat completely pointless, because teams change so much year over year.

      They should track PDO for 10+ years before starting to make statistical assumptions.

      Just look at the league rankings and the bottom teams are Edmonton, Carolina, Arizona, Buffalo. Are they teams unlucky or do they just suck? If you really think PDO is accurate, the Oilers are much better than their record, since they crush the league in terms of worst PDO.

      I’m really starting to think advanced stats are one big giant pile of BS. I think in 5 years they may look back and laugh at what everyone believed to be true.

      • This has been discovered already. The influence of luck tends to be overrun by the influence of skill after about 70 games.

        Seriously folks, there’s lots of existing research out there. Maybe do the reading before jumping to vapid conclusions.

        Which isn’t to say some of what we know isn’t wrong (there’s lots we don’t quite know yet).

        • Kevin R

          I think what has soured many, I find it has gotten my back up a bit as well, is when I read writers on other sites that are similar to Lamberts writings where the use of stats is out of context. They make opinions that are questionable and not practical & then they use stats to their convenience to back up their opinion to come across that their opinion is indeed fact. Then they get insulting those who question both the stats & the opinion.

          I got hooked on this site Kent from your writing approach where you have taken the results & then looked at the stats & try to formulate an opinion of what is going on which leads to lots of dialogue as to what might be needed to change.Use the stats right & it will only enhance the hockey analysis & the discussion as well. Just my opinion anyway.

        • RKD

          ‘Which isn’t to say some of what we know isn’t wrong (there’s lots we don’t quite know yet).’

          Well said Kent, and the “never ending” search for the stat pact to end all stat packs continues, forever!

          All stats are useful, but you have to work with what you got, e.g.. Corsi is not possession but shots directed subbing for it. As well you need to develop the best interpretation or rather application for them. I don’t think we should over analyze things to death no amount of stats can guarantee who will win tonight. And that is how I like it.

      • beloch

        This is the problem with analytics. It’s all based on regression to the mean. So we are told not to look at small sample sizes during a season.

        No, you should always be aware of the sample size any stat is based upon. This is a core message of statistical analysis.

        BUT from a macro level, the length of time they have been tracking advanced stats is in itself a small sample size. Ironic. Is 1 season long enough to average out? Might it take 5 years to find the statistical average with no luck? This would essentially make the stat completely pointless, because teams change so much year over year.

        They should track PDO for 10+ years before starting to make statistical assumptions.

        There’s no need to look at ten years worth of data to say what PDO should average out to across the league. It’s always very close to 100% because it’s an average of goalie save percentages plus the percentage of shots (the other team takes) that aren’t saves. Those would always add up to 100% if it weren’t for empty net goals, and you’d need a lot of those to throw the league average PDO significantly off of 100%.

        Just look at the league rankings and the bottom teams are Edmonton, Carolina, Arizona, Buffalo. Are they teams unlucky or do they just suck? If you really think PDO is accurate, the Oilers are much better than their record, since they crush the league in terms of worst PDO.

        Some of both. It’s a mistake to think that PDO is purely luck. Edmonton has two crappy goalies who have no business being NHL starters, and that’s one reason why their PDO is consistently low. Their sh% has also been lower than it has in previous years too, which might be partially luck and might be partially something else.

        I’m really starting to think advanced stats are one big giant pile of BS. I think in 5 years they may look back and laugh at what everyone believed to be true.

        If you don’t understand precisely what an advanced stat can tell you, it’s easy to come to spurious conclusions. The fault is not in statistics, but in your understanding of them.

  • MichaelD

    Remember when we used to talk hockey without resorting to analytics in every discussion? II don’t have an issue with the topic but to include it in every conversation is painful. Let’s start a discussion that doesn’t involve percentages of everything under the sun but intelligent hockey fans sharing their thoughts.

  • SydScout

    Hit the road to Glens Falls after a disastrous two week stretch from the big club. Let’s hope the drive is worth it.

    C’mon Emile, Sven, Joni – show the big boys how to do it. Expecting Ferland to be a man amongst boys.

    Pumped to see the Rocket Skates!