Five things: A preview of sorts

1. Tonight, tonight

While one would normally hesitate to call any game in mid-February “a big one,” tonight’s game at Los Angeles is about as big as a mid-February game is likely to get.

The Flames are holding on somewhat comfortably to that playoff spot of theirs, and prior to last night’s games were tied for second in the division, which no one thought possible for Feb. 11 way back in September. They keep winning, but San Jose (albeit with two extra games played after Wednesday’s tilt with Washington) and Vancouver (still holding a game in hand following the Chicago matchup) are hovering close; that seems like it’s shaping up to be a three-horse race the rest of the way, and that’s if everything goes ideally for the Flames.

The only other threat in the Pacific is these Kings they play tonight, and they too hold a game in hand on Calgary but sit five points back. The narrative is that the Kings are underperforming and the Flames are a surprise, and this is one of the rare instances when the narrative is actually completely on-point.

This is a real stat on the Kings that I cannot believe is true: Since Dec. 6, they are 9-11-7. Nine wins from any 27 games seems impossible for any team with a roster of the Kings’ quality. Also since Dec. 6, Calgary is 13-13-1, and that run included literally the entire eight-game losing streak.

The Kings may have won their last two games, but Calgary’s taken down four of the last five, and nine of the last 12. Given the talent collected, you’d never guess this would be the case.

Hockey’s weird sometimes.

2. Stark reality

(Author’s note: If you think wins are the only arbiter of whether a team plays well and regularly say fancy stats “don’t tell the whole story” to refute them altogether, please stop reading here. Enjoy your day.)

But let’s consider something beyond the wins and losses over that period: LA has only won nine games out of 27, but do you know their goal differential at 5-on-5 during that time? It’s only minus-2. In the same timeframe, the Flames’ is minus-3. But LA is two games below .500 in regulation, and Calgary is four games below it.

Which should tell you a lot right there: The Flames have nine regulation wins from 22 games in that stretch, but they’re 13-3-1 overall because they keep winning in overtime, when it’s basically a coinflip. The Flames have had it come up heads four times out of five, while the Kings were aggrieved to see it land on tails seven times out of seven. So that’s one thing that colors opinions of both teams unfairly.

We also have to consider the Flames’ past meetings with the Kings this season if we want to look at this matchup rationally, and interestingly all three came in the 27-game window.

As you know, the Flames won them all. The Flames are therefore considered the favorites in tonight’s game, because they both “own” the Kings, and are higher in the standings. Stands to reason. But three games is of course a small sample size, and moreover all three were decided by razor-thin margins. Two were in overtime, the third was a 2-1 game. 

Altogether, the teams played nearly 145 minutes at 5-on-5, and goals-for was even at 3-3 (which I guess you’d expect in three one-goal games). Possession, not so much: the Flames were outattempted 168-105, or for those who don’t want to do the division at home, 38.5 percent corsi for the Flames. Do you want to guess how likely it is that you get outshot that badly and end up not only with an even goal differential, but also three wins?

Oh and you don’t need me to tell you now that the Flames got super lucky in those games too: the Kings shot just 4.47 percent (3 of 67) on Jonas Hiller and Joni Ortio, while Jonathan Quick did a very creditable job of holding down the fort in his own end with a .933 save percentage (42 of 45). Again, these are small sample sizes, but the 97.8 PDO for LA speaks to not getting any damn bounces to go their way.

This game should be of deep concern to Calgary. They’re playing a team that has flat-out dominated them.

3. Really, they aren’t good

As the Flames’ improbable run of success continues, data continues to emerge that this is a truly awful team getting by on amazing shooting luck.

They keep winning games, but as Travis Yost points out, they really don’t deserve to. Their corsi event differential is is minus-498, third-last in the league but well beyond Colorado (minus-600) and this historically bad Buffalo team (minus-1,214!!!!!!!!!). But in the Behind The Net era — 2007-08 to present — minus-498 with 28 damn games to go 21st in negative corsi differential, and they’re charging hard up that list. If they continue at the current pace, they’ll subtract another 258 from that number, bringing them down to minus-756. That would make them a worse possession team than everyone but these teams, plus the 2013 lockout-shortened Leafs, who went minus-518 in just 48 games:

Screen Shot 2015-02-11 at 3.52.59 PM

Please understand this has nothing to do with trying to be a downer or not-liking the Flames. The Hockey Media did not randomly decide to Hate The Flames this season, just like it didn’t decide to Hate The Avalanche or Hate The Leafs the two years prior. This is maybe the 10th- or 11th-worst team of the past eight years, just going on this metric. Which, no, it’s Not Everything. But it sure is telling; that’s a lot of last-place teams on that list, and somehow Flames fans think their team is immune to the same problems those teams suffered? That doesn’t make a lot of sense.

This is just objective fact: The Flames get badly outshot just about every night, and we know from years of recording this data that this kind of thing generally leads to losing. We also know that every year, one or maybe two teams can defy those odds by scoring or saving more pucks than they should, and typically they get clobbered in the first round, rebuilding or not.

Not that any of you care. You’ve already decided that this season is an unqualified Success, and that maybe this club has figured out a way to elevate its shooting percentage through Hard Work artificially.

But a look at how Calgary shoots compared with what its opponents do shows that they give up roughly league average in terms of chances right around the net and around the faceoff dots, but take less than that themselves. Most of their shots come from the high slot and down from the middle-left point.

Screen Shot 2015-02-11 at 4.08.13 PM

And so “shot quality” might not actually be the answer here, because have a look at the team’s shooting percentages relative to league average in those same areas:

Screen Shot 2015-02-11 at 4.11.07 PM

This basically says they’re scoring more goals on fewer shots from the area around the crease. As well as right up the middle of the ice. 

Now let’s compare those to last year, first with shot location, then with shooting percentage:

Screen Shot 2015-02-11 at 4.13.22 PM

Screen Shot 2015-02-11 at 4.14.01 PM

Unfortunately there’s no way to look at these heat maps based on date right now, so we can’t exactly go and examine actual data from when the Flames “started to take off” toward the end of last season (though I’d argue that not coincidentally it was around the time they had the bright idea to put Mark Giordano and TJ Brodie together). But given all this information we have on hand, I don’t see where got appreciably better at improving shot quality for or against; they’re more or less the same, but pucks are going in for them this year.

It happens.

4. A slumbering titan

Meanwhile, the Kings are secretly one of the best teams in the league even if their record does not in any way reflect it.

The Kings are, in fact, the best possession team in the league, but have struggled with a low shooting percentage once again, and also suffered 12 overtime losses that probably cost them six points or so in the standings (they’re 2-12 beyond regulation). Add six points there alone and they’re ahead of the Flames, who have benefited heavily from three-point games this season (10-3).

This is weirdly almost always a team that doesn’t shoot particularly effectively, and the shot charts suggest that they get around the net a fair bit more effectively than the Flames have this year. 

Another thing to consider: From Oct. 21 to Nov. 21, the Kings were laboring under the unfair penalty of having to carry Slava Voynov’s cap hit while not actually being able to play him (and he 100000000 percent should be banned from the league, so that’s not my argument). For that month, their cap was stretched thin and they regularly dressed just 17 skaters. Not surprisingly, they went 7-4-3 without him, which doesn’t sound bad at all, but you have to keep in mind they also started the year 6-1-1, so that’s a bit of a comedown, and it seems to have taken a bit of a toll.

People don’t like to hear the world “luck” thrown around in these regards because it sounds like mysticism to some extent, but what we really mean when we say “luck” is that teams are defying the odds; a team shouldn’t have the 14th-largest goal total in the league but the 28th-most shots (as the Flames do). Likewise, they shouldn’t have the eight-highest shot total but the 14th-most shots. It stands to reason that this isn’t sustainable if you actually think about it.

And that doesn’t even get into the quality of player involved. Let’s say Dean Lombardi calls Brad Treliving tomorrow and offers a straight-up roster swap: All 23 Kings for all 23 Flames. You of course take the Kings. Why? Because they are better at everything than the Flames. Except putting the puck in the net, this year only. And winning, this year only.

What I’m trying to do here is illustrate why there’s such a huge fallacy behind looking just at wins and losses or overall goal differential. This is about The Process that goes into winning hockey games, which the Kings have demonstrated for years on end works better than relying on smoke-and-mirrors shooting and save percentages, but which isn’t resulting in wins this particular season. Likewise, Calgary is Colorado and Toronto and Minnesota and Dallas of the last several years: They’re winning, but creating a kind of mythos about how it happens that obfuscates things for observers (media and partisan alike).

Put another way: Do you think it’s any coincidence that of the top-16 teams in the league in score-adjusted corsi prior to last night’s three games, the overlap here is so stark?

Screen Shot 2015-02-11 at 4.33.52 PM

Screen Shot 2015-02-11 at 4.34.19 PM

That’s four outliers out of the 16 “best” teams in the league. Montreal has the best goalie in the league on an insane hot streak to cancel out the team’s rotten possession. The Rangers, likewise, rode the best goalie of his era to huge success until his injury. Vancouver and Calgary are both tough to explain if you look at the quality of the rosters, except to say that the Canucks’ penalty kill has been amazing (fourth in the league? how?).

I literally do not understand a mindset where the takeaway is “The Flames are so good and unique you can’t use possession stats to evaluate them.” The same is true for determining the Kings are bad.

You play the games on the ice, obviously, and the Flames have done well in this regard. But it’s about controlling as many factors as you possibly can, and the best way to do that, proven time and again, is to be one of the 75-85 percent of teams in the top 16 in the league in terms of 5-on-5 possession stats every year. Those are the teams that make the playoffs, and a few lucky bastards who operate on the margins, or even well below them, end up sneaking in.

But for Flames fans, if you say “luck” is involved — good or bad, even if it clearly is — they excoriate you. Even if much smarter people than them have done the math and figure out luck accounts for 40 percent of results. The Flames get killed in the other 60 percent, so right now it’s only the 40 percent that counts.

5. Does any result matter?

But as I found a few weeks ago and have been shamelessly harping on since, it might turn out that none of this actually matters.

Anyone can win on any given night, the Sabres won 14 games this year, etc. etc. etc. I understand that. But as long as the Kings float a win or three back of the Flames, a team whose numbers scream that they’re going to go into freefall at any moment, there’s going to be that doubt. The doubt is simple: The Flames aren’t good, and the Kings — in theory at least — very much are. The longer they keep peeking in over the edge of the horizon, or hovering just beyond it, the more likely it is that they catch up. A win tonight could essentially put a knife in the whole “Kings are gonna catch up to Calgary” belief that is very rightly still swirling around the league.

The Wild, I guess, are a different story entirely. Boy, that playoff race is tightening.

  • Ramskull

    Ahem I should have finished typing before posting:

    The 2008 Blackhawks were not a dominant possession team with a 49.7% CF sitting 16th in the league.

    That year the 2008 Calgary Flames, were fifth in CF% at 52.5% and got knocked out in the first round, by the way.

    By your model, the Blackhawks should have “expected to be” right where they were the next year…..Then they come in as the 3rd best CF% team with 55.2% in 2008-2009

    Your model falls apart because you’ve got them standing pat instead of improving, as if they’re a complete veteran team.

    Oh, and the #2 team in CF% that year? The 09 Flames, who were again knocked out in the first round.

    It’s funny, if the Kings don’t win a championship last year, and lose a ridiculously close series to the Sharks (which was statistically probable when they went down 3-0 in the series), the Kings are a team that lost in the first round.

    And you wouldn’t be talking about them being the sustainable model of success while they’re a half dozen points out of a playoff spot.

      • RedMan

        Yes, because Kipper just decided to fall asleep for two entire playoff series.

        Maybe, just maybe, there’s more at play than evsv%

        But no, I don’t have any idea what I’m talking about. Which goalie has the highest evsv%? that goalie should win the next championship.

        • RedMan

          exactly, there is more at play. like the inability of the team to outplay their awful goalie.

          sv% is literally the amount of pucks he was stopping. he was doing a very very poor job of that. it is his fault. probably can blame him for the collapse, too, since he cost the team at least 5 or 6 points.

  • RedMan

    Is there anyone at the rink today that can give us an insider peak at what the team is working on for practice???

    I REALLY hope the team is doing CORSI and PDO drills, otherwise… you know…. sh*t

  • RedMan

    From what I have read, the correlation coefficient between many of these advanced stats and wins is ~0.25-0.3.

    That coefficient demonstrates that these stats are not in any way effective predictors of wins. Therefore, I’m confused why you’re making an effort to explain actual data (wins / losses) using parameters that are not correlative to your endpoint.

        • RedMan

          http://www.pensionplanpuppets.com/2013/7/10/4508094/what-statistics-are-meaningful-in-a-given-season-corsi-fenwick-PDO-hits-fights-blocked-shots

          rvalue points = >.50
          rvalue wins = >.50
          reliability = >.35

          that’s one season. three years increases those confidences to around 75-85%. assuming you’ve used the term “correlation coefficient” correctly. your milage may vary but personally i’ve never seen a post where a shot-attempt based metric has had an rvalue that was less than .40.

          • RedMan

            .35 would never be acceptable in real world use.

            I’m an analyst by trade. If your correlation coefficient drops to .35, you stop using it or you find out whats going on. You definitely don’t say “meh, just luck” when you’re already past 50% of your reporting year.

          • username

            read the post. the rvalue to wins is about .5-.6 over one season and gets higher the longer it goes up to about 3.5 yrs. your appeal to authority doesn’t mean more than the numbers. also i find it funny that you mention the real world because I’m assuming in the real world you have better tools to get better coefficients. guess what? we don’t and yet we’re still getting these kind of high-confidence results. why does everyone else always ignore the roughly 25+ teams that put up results that this suggests they do/should?

            my advice, and the last thing I will say on this matter this season, is to stop with the cognitive dissonance, accept the intuitive analysis given/told/whatever by these stats, realise their significance and still enjoy the game.

          • mattyc

            http://www.sportingcharts.com/articles/nhl/projecting-2014-15-nhl-standings-using-fenwick-corsi-possession-statistics.aspx

            That article examines these stats from 2009.

            5-on-5 Corsi’s R^2= 0.26
            Fenwick’s R^2 = 0.33

            Good luck trying to convince me that these stats are anywhere near reliable as predictors of team points.

            I don’t have an issue with these “advanced” statistics by the way. What I have an issue with is garbage articles like this that use them as the be all and end all to hockey discussion. Did you read this trash? He completely dismisses all of the results that don’t fall in line with his “analysis” as luck. We’re no longer at data point 10/82. We’re well past halfway through the season. Like I said in my original post, this is the anti-thesis of statistical analysis.

          • Southern_Point

            You didn’t read Justin’s post and you aren’t paying attention.

            The Shot based metrics have an R2 between .5-.6 when trying to predict WINS, not points. They are also fairly repeatable year over year. Goal based metrics are the better predictor of points but Goal metrics are more subject to stastical variance.

            Thus the goal should be to aim for wins through possession dominance (The Blackhawks) and not points based on shooting percentages (The Flames).

          • Ramskull

            Convenient how now WINS are all of the sudden the measure for success, but according to this article we are to dismiss the wins of the Flames as a measure of success, because their number for that very same statistic that depends on wins for reliability, is too low.

            Again, I do not have a problem with these “advanced” statistics. I think it’s awesome that we can paint a better picture of what’s going on. What I’ve had a problem with since this “Flames hate train” left the station is the outright dismissal of the Flames as anything but a “lucky” team (for 50+ games, including an 8 game streak where they out-corsi’d most of their opponents and still lost).

            There have been other teams that have managed to be successful with bad corsi numbers.

            Pittsburgh for example (thanks CP.com):

            Year corsi%5v5 pts
            06/07 46.2% 105
            07/08 45.5% 102
            08/09 48.1% 99

            Maybe you should take away their cup? Clearly they were just lucky.

            It’s very possible to have a successful system that isn’t driven by possession. This is particularly true if you have a small, inexperienced team with young talent (i.e. Flames). For example, Shooting is a skill that is not relative to your opposition. If you can shoot the puck it doesn’t matter if you’re shooting in an NHL game or alone in a backyard rink, a good shot is a good shot. Possession on the other hand is relative to your opposition. If you’re inexperienced and small relative to your opposition, you are likely going to have a hard time maintaining possession. So why dismiss the Flames because they acknowledged this deficiency in size and experience and built their system around shot blocking, conditioning, strong break out rushes and high percentage shots? To me that sounds a lot less like “Luck” and a lot more like awesome coaching. It’s very obvious in the Flames game. They usually lose the puck when battling in the corners or trying to cycle but when they’re passing on the rush , it almost seems like they can’t be stopped.

            Hockey is not Soccer. Possession is easily thwarted by keeping shots to the outside and reliable goaltending. No luck needed. In a sport like soccer, you have 10 opponents and a much greater distance to cover so controlling position is needed to move up the field. But even in that sport, the analysts acknowledge that it’s possible to build a system that doesn’t focus on possession and not call it “luck”. Why can’t Lambert?

          • mattyc

            Thanks for the Pittsburgh Stanley Cup winning year stats,Polak!

            Now we have an example of how the Flames can do it..and not just Leafs and Roackies as examples of regression…their were obviously other reasons than just bad advanced stats for failures of those two teams…

          • mattyc

            If your correlation coefficient drops to .35, you stop using it or you find out whats going on.

            I work w/ environmental data, and there’s tons of stuff that runs off/gets published with less than 0.3 R2. I think it depends a lot on your system and what the alternative is. Can you imagine if a civil engineer said “nah the r2 is only 0.28, I’m just gonna eyeball it.”

            Obviously you wouldn’t treat that number as the gospel, but barring a compelling alternative, it’s the best you’ve got.

          • mattyc

            Ah yes.

            Well, this simply confounds the error.

            Random time series can and do correlate spuriously, to the point where correlation coefficients less than 0.8 are routinely ignored or treated with great skepticism. This is why such data, when rigorously examined, are differenced.

            Nonetheless, all you’re doing in your example is amplifying the Type I error.

            Also, the reliability measure in that link is horribly circular reasoning. I’m somewhat shocked it is used as justification in any way.

  • I can sense a touch of increasing frustration from these articles.

    1. I really don’t believe there are many fans out there that believe this is a cup-contending team, or that at least to some extent the Flames have been lucky. This has been acknowledged at lenght.

    2. In my opinion, the whole point of fandom is to ride the highs and lows of your favourite team. So the Flames CORSI sucks, and their PDO is high, yadda yadda. We all have enough crap in our lives to start being all gloomy and waiting for the roof to fall in for our suddenly hot team. Like others have said, who the bleep cares? Just enjoy the ride!

    3. Even still, I don’t take issue with your opinion Lambert. I don’t have a problem with you saying we should stop enjoying our ‘luck’ and start worrying about doomsday, even if it goes against what it means to be a fan. I do take issue with your condescending tone. That because you believe in advanced stats, anyone that doesn’t or disagrees to the level of its importance, is somehow inferior to you.

    4. Again, like others have eloquently said, maybe its not as simple as saying CORSI defines how good a team is. I’m sure the ’04 Flames had horrible CORSI. That won’t take away from the fact that they were a goal away (or maybe not – argghh) from winning the freaking Stanley Cup! I don’t know what sports team you’re passionate about Lambert, but surely you can put yourself in the shoes of a fan and stop making it appear you’re on a high horse.

  • Southern_Point

    One more thing:

    Montreal 2015
    Colorado 2014.
    Calgary 2015

    I think we would agree these three teams are not “possession” teams. Ryan Lambert feels none of these teams would be in the playoffs “sustainably”

    Now, OP makes the argument that Carey Price is the only reason Montreal is good.

    Is he great? Sure.

    But maybe, just maybe, there’s another common thread being ignored? A thread that might even be giving Carey Price additional opportunity for success?

    Bob Hartley
    Michel Therrien (who is a Bob Hartley Protege as his former assistant)
    Patrick Roy (who is a Bob Hartley Protege as his goaltender and a guy who used to pick his brain all the time and it’s documented)

    …Maybe these three teams were on to something…systemically?

    Maybe less shots does not mean poor offense?

    One thing is clear – talented teams win championships.

    One thing that isn’t clear is how many other coaches in the NHL are playing Bob Hartley style hockey, other than Bob Hartley and his two proteges.

  • KiLLKiND

    1) Obviously the Kings strategy of dominating possession wise leads to more long term success and they are a better team.

    2)It is because possession is a piece of the puzzle needed to win. It may not be the piece used in every game to win but always can help. but when you watch the Flames play as of late do you actually feel like this team is going to fall off a cliff? I went to war on ice set the date range from January 15th till now and yes their 5-on-5 Corsi isn’t over 50 it is 47.7 and for all situations their corsi is 51.2!!!

    3) Why Calgary isn’t the same team nor should be judged by the same standards as the team they were expected to be. Calgary started off looking all for the world that every game was luck. This continued all up until slightly past the 8 game losing streak. Then they started to actually dominate games. Yes 5-on-5 isn’t a strength but in all situations which really should be the measuring stick especially when you consider how PP’s generate high scoring chances.

  • KiLLKiND

    The LA Kings have played 60 more games over the past 3 years than non-playoff teams, playing into June. Top flight play over this many games is not sustainable. The Kings pace themselves over the regular season.

    As for tonight’s game, Kings returned from an East Coast road trip…domestic challenges prior to tonight’s game? Flames have been relaxing since their game against the Sharks and will come ouy flying – hopefully!

    4-1 versus Sharks 3-0 versus Kings! Wins-Loses > Corsi.. Good Karma!

  • Southern_Point

    My prediction:

    Flames just make the playoffs (8th) and drop the series 4 games to 0. Lambert states: I told you so!

    Lambert’s prediction next season: The Flames will fall like the Leafs and the Rockies because luck is not sustainable!

    Flames are in a rebuilding mode, but there is a lot of good things happening – like disciplined hockey, good team play, surviving beginning of season injuries to 3 starting centres, the biggest losses being Backlund and Stajan, growth of the young prospects with Bennett, Poirier and Ortio on the way…..

    • mattyc

      you were being banned because you said 3 disparaging words and personally attacked writers, not because you are acting like a child. people doing either of the first two are gone, no questions asked.

      “Comments are moderated.”

  • Missy_GFG

    Ok Ok Ok we get it! The flames are supposed to lose and we are crazy fans for getting our hopes up blah blah blah.

    But here are my thoughts:

    -We all know that we are sitting better in the standings than expected/deserved. The standings are a shock to all of us!

    -We all know that the bottom may fall out at any time. We know this not only because of the fancy stats that keep being presented to us but because we are flames fans. I’ve been on the rollercoaster for too long. I know when it is too good to be true. But with that I’ve also learned to enjoy when we’re playing well! Which as a fan a win is what I want to see! Good stats or bad stats I loved watching us go to the Stanley cup final!

    -I don’t think these stats are a bunch of garbage. Looks to me like there is lots of consistency in winning teams showing these stats do matter. I think there is real value here. I just don’t think as a fan it’s for me to spend too much time reading in to. I do hope however, that flames management looks into all aspects of our team including these stats. And to continue to rebuild.

    I’ve read about the stats – got it. But I don’t need to read it everyday. I’d much rather read the “holy crap we just beat San Jose” board. Because I do know this could change at any time. I’ve learned to enjoy the highs.

  • smith

    My favorite part of Lambert’s articles are how he keep’s saying we will fall to the mean. Of course we will. What does he think the mean is? Everybody will given long enough.

    The problem is that no one knows what falling to the mean will look like. Do the flames massively increase their shot quantity? Do they get blown away for a few games? Also since nobody knows when this will occur it is completely pointless to keep harping on.

  • King Quong

    FYI – Someone on Calpuck just pointed out that the Flames outstanding record against their own division (17-4-2) is successful DESPITE having a 44% corsi in those games. Man the Flames must just be EXTRA lucky in those games right?

    What I personally see, is more evidence that the Flames have found a system that can be successful without high posession numbers.

    • Greg

      A) he probably doesn’t get paid to write them.

      B) we don’t all dislike them. It’s good to have a contrarian voice in a group. Personally, I enjoy the contrarian analysis, and look forward to my weekly dose of reading comments from people repeatedly saying “I’m never reading this again” after reading it yet again.