Flames in March

 

 

It’s been a strange last leg for Calgary. The club has gone through consecutive five game winning and losing streaks in March, with the latter ironically occurring as the club has gotten healthier. Throw in Matt Stajan getting hot and Olli Jokinen falling off a cliff and you have perhaps the oddest month of the year for the Flames.

Beware the Ides of March

Frequent commenter the wolf asked me to look at the club’s underlying numbers over the last 15-games or so, specifically because the team’s results have sunk despite the return of various veterans. That seems counter-intuitive, but suggests maybe the club was better off playing the kids.

I’m going to do this a number of ways: first we’ll take a look at how the team has done overall this month by looking at possession numbers (shots, fenwick and corsi) and the the percentages (on-ice SH% and SV%). Then we’ll break things up into halves since the first part of the month was wholly different than the second part. The results are both interesting and instructive.

Before I reveal everything, let’s engage in a brief experiment…

Here are the possession numbers for each player this month. Instead of adding their percentages, I’ve included a "GRADE" column beside each of their names. Take a few moments and mentally grade each player based on his performance over the last little while. Use whatever scoring system you want (A to F, 1 to 5, etc) and you can even employ the possession rates to inform your marking (or not). Concentrate on just a few players who have stood out (good and bad) or look at the entire team if you want. 

Player shots% Fenwick% Corsi% GRADE
Iginla 48.0 50.4 50.1  
Tanguay 47.7 49.7 48.0  
Cammalleri 47.0 50.4 50.3  
Jokinen 50.8 51.5 51.0  
Glencross 51.3 50.0 49.6  
Moss 51.9 52.5 52.5  
Stajan 49.7 49.7 49.6  
Comeau 55.6 55.9 54.9  
Kostopolous 46.9 47.5 47.8  
Jackman 51.9 53.0 48.3  
Horak 46.8 51.4 48.6  
 Defense        
Bouwmeester 43.8 47.3 47.2  
Butler 50.0 50.6 46.9  
Giordano 52.2 52.8 51.1  
Hannan 50.0 49.6 47.7  
Sarich 48.8 51.3 52.8  
Babchuk 52.4 55.3 54.9  
Smith 47.5 48.2 48.1  
 Kids etc.        
Jones 50.0 51.9 58.5  
Wilson 60.0 58.6 63.2  
Nemisz 47.6 51.6 51.2  
Baertschi 47.4 42.3 39.2  
Desbiens 40.0 43.8 43.6  
Bouma 41.7 48.6 40.4  
Kolanos 59.3 61.5 63.8  

The first group is the regular forwards, the second the regular defenders and the third the call-ups and kids whose samples are so small, all of their results should be taken with a grain of salt.

Now, here they are with the percentages included:

Player shots% Fenwick% Corsi% Es SV% ES SH% PDO
Iginla 48.0 50.4 50.1 91.0 14.6 105.6
Tanguay 47.7 49.7 48.0 92.2 12.3 104.5
Cammalleri 47.0 50.4 50.3 94.3 12.8 107.1
Jokinen 50.8 51.5 51.0 86.4 6.0 92.4
Glencross 51.3 50.0 49.6 87.1 7.1 94.2
Moss 51.9 52.5 52.5 92.0 5.3 97.3
Stajan 49.7 49.7 49.6 95.0 10.1 105.1
Comeau 55.6 55.9 54.9 98.3 4.1 102.4
Kostopolous 46.9 47.5 47.8 93.6 4.3 97.9
Jackman 51.9 53.0 48.3 100.0 0.0 100.0
Horak 46.8 51.4 48.6 92.7 5.6 98.3
             
Bouwmeester 43.8 47.3 47.2 93.5 5.8 99.3
Butler 50.0 50.6 46.9 92.9 0.0 92.9
Giordano 52.2 52.8 51.1 90.9 10.4 101.3
Hannan 50.0 49.6 47.7 89.8 9.5 99.3
Sarich 48.8 51.3 52.8 98.8 6.3 105.1
Babchuk 52.4 55.3 54.9 89.8 11.1 100.9
Smith 47.5 48.2 48.1 89.4 9.1 98.5
             
Jones 50.0 51.9 58.5 87.5 0.0 87.5
Wilson 60.0 58.6 63.2 100.0 8.3 108.3
Nemisz 47.6 51.6 51.2 90.9 15.0 105.9
Baertschi 47.4 42.3 39.2 95.0 16.7 111.7
Desbiens 40.0 43.8 43.6 96.3 0.0 96.3
Bouma 41.7 48.6 40.4 100.0 0.0 100.0
Kolanos 59.3 61.5 63.8 90.9 0.0 90.9

PDO is, of course, even strength SV% ad SH% for the player added together. This invariably regresses to the league mean of 100.0 over time for each player.

Discussion

With your player grades in mind, take a look at each guys percentages and in particular his PDO. You’ll find almost invariably your best marks went to those with the highest percentages and your poor marks to those with the lowest.

Matt Stajan’s incredible revival and Jarome Iginla’s emergence as a Hart candidate are both narratives based on the more or less random fluctuation of percentages around a mean. On the other hand, Olli Jokinen has suffered a precipitous fall from grace even though his possession numbers have improved with the return of David Moss. Please keep in mind that Jokinen was leading the Flames in terms of PDO (104) heading into this month with a on-ice SH% north of 13%. When he was riding that wave, of course, Olli was celebrated as a guy who was "the most consistent player on the team" this year because he had "completely changed his game to be a capable two-way center". This was despite, of course, some fairly unimpressive possession rates at the time.

Blake Comeau may have been an exception to your grading rule because he ended up with a fairly decent PDO due to a really nice on-ice save percentage. However, the factor that more often determines perception for forwards is on-ice SH% and he suffered through month at 4.1% at even strength (league mean = 8%). That is despite the fact he had one of the very best possession rates on the team. This probably heightened the general dislike and frustration with Comeau in the Flames fanbase – no matter how much he had the puck in the offensive zone, his line almost never seemed to score.

The exercise isn’t to castigate anyone as blinkered or stupid, just to again highlight the role randomness and scoring outcomes has in shaping our view of players and teams as a whole. It’s always tempting to feel that the most recent streak is also the most indicative of a players abilities and worth, but the truth is natural variance plays a giant role in determining each guys percentages as it bounces around the mean.

We can investigate this further by looking at the Flames month of March split into the first and latter halves.

Here is the first 7 games.

Here is the the last 8 games.

For those too impatient to click and wait, here’s a summary of some key figures –

First half:

Player shots% Fenwick% Corsi% Es SV% ES SH% PDO
Iginla 45.3 45.3 46.2 89.5 19 108.5
Tanguay 4 42.9 43.1 91.4 14.5 105.9
Cammalleri 43.3 45.2 46.4 97.4 13.8 111.2
Stajan 49.1 45 44.2 94.5 9.4 103.9
Jokinen 55 53.4 52.6 88.9 10.6 99.5
Glencross 55.7 52.8 51.2 90.7 11.1 101.8
Comeau 54.2 50.8 49.4 100 3.8 103.8
Giordano 51.3 50.5 47.1 93.4 12.5 105.9
Team 48.9 48.1 48 93.2 10.3 103.5

Second half:

Player shots% Fenwick% Corsi% Es SV% ES SH% PDO
Iginla 49.2 53.5 51.9 92.5 12.3 104.8
Tanguay 50 54.6 50.7 92.5 12.3 104.8
Cammalleri 54.5 58.8 57.1 86.7 11.1 97.8
Stajan 49.5 53.2 52.2 96 12.2 108.2
Jokinen 49.6 50.3 48.6 84.1 1.6 85.7
Glencross 47.9 48.4 48.6 84 2.2 86.2
Comeau 55.4 57.1 55.1 97.8 5.4 103.2
Giordano 54.1 54.8 4.1 87.1 9.6

96.7

Team 49.6 52.4 51 91.7 6.8 98.5

You can see how various players percentages changed over the course of the month. Glencross and Jokinen in particular have seen their fortunes dip drastically, despite the fact they were two of the luckiest players on the squad up until recently. The top line was still technically lucky over the last 8 games or so, but nowhere near as fortunate as they had been to open March…particularly Jarome who rode an uncanny on-ice SH% of 19% through the first 7 games. He had previously been hovering around the league average of 8% by the way.

Giordano is an interesting case. Over the last two weeks his possession rate has been the best amongst regular blueliners and well above what he’s managed all year. This may be because he’s playing easier minutes now that Butler is back or seeing a lot more offensive zone ice with the team trailing late in games. Whatever it is, he has looked more like the player of old for whatever reason. Even a lousy on-ice SV% hasn’t been able to undermine his perceived (and apparently real) improvement.

The numbers to really pay attention to are the overall team results. Through their earlier winning streak, the Flames were below water in terms of shots and possession at even strength, but won thanks to a PDO of 103.2. As the team has become healthier they’ve been able to control shots and possession better, but the percentages have ironically collapsed (PDO = 98.5) resulting in that five game winless streak prior to last night.

Big Picture

Calgary’s ability to control the play has gradually improved as the month has gone on. This is due in part  to the return of useful players like Moss, Butler and Jones and in part it’s because their schedule has been littered with pretty lousy teams. Add in some playing to score effects since Calgary has trailed a bit more often in the last 8 games or so and you have better possession numbers.

However, the club’s shooting percentage cratered for a spell and Kipper went from excellent to merely human – and probably not for any significant reason besides randomness. This is why the team got better on paper, but suffered worse results.

Ironically, if you combine the first and second halves of March, the Flames stats are wholly average across the board: 50.2% corsi, 92.3 SV% and a 8.2% shooting percentage. The natural outcome of which is a mediocre .500 record when all is said and done.

  • loudogYYC

    I agree with Jeremy, he leaves it all on the ice when he plays. The fact that he’s a big body that can skate also makes him stand out a little too.

    Regardless of stats, Blaire Jones plays like he’s happy and grateful to play in Calgary. It’s all about perspective I guess.

  • loudogYYC

    I think everyone should take the time to thank Mr. Wilson for posting this data. It undoubtably took an enormous amount of time and energy and the detail put into this is evident.

    However, as much as I appreciate all that Mr. Wilson does for this site, I feel I should encourage him to seek a social life. Much like Mr. Steinberg…there must be a lady out there just waiting for him but he never stops working. Get out, have fun, forget about hockey for 5 minutes and enjoy life’s pleasures!

      • I believe you to be too modest sir. I’m willing to bet it took longer than that. Regardless, I highly doubt myself or any others could put this up without alot more lead time and preparation.

        It may seem like I am kissing ass, but I assure you I am merely stating my honest opinion. I don’t always agree with your posts, but I do recognize the magnitude of the work you and the other Flames Voices put into this site.

    • wawful

      Well if that isn’t the most presumptuous comment I’ve read in a while, I’m not sure what is.

      Back on topic: I like Jones as well. He seems reasonably capable, but for me the major bonus is his ability to piss off other players.

  • Vintage Flame

    I’ll throw my two cents in as well regarding Mr. Wilson. When visiting FlamesNation for the first time, I was amazed by the content from the writers, and in particular the insight Kent provides. It’s an honor and privilege to work with him now, though I will not be so bold as to say along side of him yet.

    I tip my cap to you good sir.

  • wawful

    “Ironically, if you combine the first and second halves of March, the Flames stats are wholly average across the board: 50.2% corsi, 92.3 SV% and a 8.2% shooting percentage. The natural outcome of which is a mediocre .500 record when all is said and done.”

    Currently, the average point % for teams in the NHL is 0.560, not 0.500, thanks to extra points. The Flames have gone 0.567 over the last 15 games and are 0.552 over the season.

    The Flames are performing like an average team and are getting the same points as an average team (League average is 85.0 points, same as what the Flames have, although that’s a tiny big generous since the Flames have played 1 more game than league average). The Flames are painfully, horribly, *average*.

    • True enough. I was using .500 as in win/loss record rather than point percentage. Flames are 7-3-4 in march.

      Either way…yeah, Flames are average. And keep in mind this is one of their *best* months this year in terms of shot ratios and such. Definitely the best since the calender changed. The only reason they maintained a middling record (rather than falling off a cliff) was 6 weeks of Kipper imitating Hasek.

  • wawful

    Thanks Kent, appreciated.

    So what you’re saying then is that there is no real difference between when the injury bug struck and since?

    Or am I out to lunch here?

    And if there is no difference, then what are the implications?

    • RexLibris

      I was under the impression that one could derive the conclusion that the young call-ups were perhaps overachieving while the veterans were underachieving.

      That may be a misinterpretation or oversimplification, but in the end it seems that water finds its level.

    • wawful

      I think the obvious implication is that Calgary has some overpaid and under-performing players. Not all of them are pending FA’s either. Calgary needs to find ways to get better value for their cap dollar.

      I have to admit, I’m really curious how things are going to play out in the Summer. e.g. Cammalleri. Cammy is a great example of a player who has, over the whole season in both Calgary and Montreal, performed well below his pay-scale. Over his entire stay in Montreal he averaged 0.70 ppg, but the first half of this season he was down to 0.58 ppg. In the second half of this season, with Calgary, he’s gone just 0.61 ppg. That’s not the resurgence Feaster was hoping for and a little bit disappointing compared to his first stint in Calgary over which he went 1.01 ppg.

      The kicker is that he’ll still be overpaid even if he does have a bounce-back season next year and returns to his former glory. His cap-hit is comparable to that of either of the Sedin Twins, Phil Kessel, or Zach Parise. At his salary he should be a *league* leader, not a 40-50 point player (He currently has 37 this season).

      Still, Cammalleri is a proven playoff monster and is a huge fan favorite. The Flames might just be able to trade him for a lower-profile 40 point player who is much cheaper. That would be both a sensible move and, paradoxically, a very unpopular one with the fan-base.

      I have a funny feeling that the Flames Ownership wants to hitch their wagons to Cammalleri as Iginla’s short-term successor. At least they haven’t hung a 100-foot poster of him on the Saddledome yet, like Montreal did at the Bell Center.

    • The team has improved it’s possession rate in the latter half of the month. Since about Janurary, the Flames have been one of the worst teams in the league at outshooting, but since the bodies started coming back the Flames are at least at even these days. Which is a truer measure of their abilities than percentages, since those tend to bounce around randomly.

      So the Flames have been *better* recently than they have been since, say January, but some bad luck makes them seem worse.

      • Greg

        OK, so that makes sense, the vets are performing better, but enough to justify their higher salaries?

        Interesting that the team seemed to have more ‘jump’ with the young guys here, but were actually just riding the wave statistically.

        As Wawful notes, it wil be an interesting post-season.

  • RexLibris

    I was going to say the same thing Ravage brought up here. That is one nice piece of work, Kent.

    I try to make it a rule to throw in the occasional table and graph as well. it acts as an effort multiplier, making any work all that much more impressive. šŸ˜‰

    When Baertschi was sent back I believe I asked someone on chat about whether they were worried about the team taking a step back once all the vets returned. It just seemed like the kids were putting out a better effort than the vets had during their time and once the roster returned to it’s previously scheduled programming the team found it’s level (sub-par mediocrity) again.

    As for the Flames and their seemingly perpetual mediocrity: so you mean that last year when I made my birthday wish, that those things can actually come true?!

  • This rule of thumb may help those trying to make sense of these numbers.

    shots%/possession(corsi and fenwick) = circumstances + talent and a small amount of luck.

    This is the best indication of a players ability to affect the play, but must always be considered in context (where does he start his shifts? who are his linemates? who does he usually play against). The best player in the league have high numbers here, regardless of circumstances.

    percentages (on-ice SH, SV%, PDO) = mostly luck + small amount of talent.

    In small samples, a players percentages are almost all bounces and variance. Over a longer period of time, some guys do tend to have higher on-ice SH%, while others have lower, but that’s only for the guys at the very edges of the extremes. On-ice SV% for individual players tends to be completely random and mostly depend on how the goalie behind hem performs over any given sample.

    I hope that helps.

  • KW,

    Please ignore Ravage’s entreaties and continue to work your a$$ off for my enjoyment. Sleep will be permitted only when the off season allows for a lull in our insatiable demand for hockey analysis. Furthermore… DANCE, MONKEY, DANCE!!!

    Oh, btw… good work. Everyone has been “noticing” Gio lately, but it sounds like it’s still going to average out to a relatively poor year. You mention that the weak schedule this month has helped out the team (and Moss/Butler/Jones), so I would assume Gio has had the benefit of some lighter work as well.

    Beers soon?

  • @KW

    I just think anyone with Giordano’s talent can play well for short stretches (especially against poor teams), and no doubt he looked great last night. But when you get to your end of season reviews, I fully expect to see that he slid backwards year-over-year and we will be having trouble justifying his new dollars. I’m interested to keep reading as you discover the hidden patterns in the numbers (or as I suspect, you will sacrifice a small animal and “read” it’s entrails).

    You know I’m Gio’s second biggest fan, so I’m hoping he can play to potential next season.

  • Jeff Lebowski

    Kent,
    A very thought provoking piece. This team seems to buck the trends in a few regards.

    What I have seen this season are many games where Calgary essentially gives time and space away to the opposition. Although I couldn’t quantify it I think this drives some of the soft possession and shot differentials.

    The defensive shells, the reliance on Kipper, the fact they win more with Kipper playing this way frustrates me to no end.

    The question is why do this? The only thing I can think of is this unrelenting desire to be positionally sound and to play low risk offensively. On the D side, you see a puck up for grabs and Calgary backs off, allowing posession and shots against. Perhaps this desire to keep defensive shape is to curtail getting out of position by being aggressive. They will take the trade off of being able to cover the ice instead of getting the puck first.

    Some guys just won’t give up those battles so you see a Gio improve his numbers because he and Hannan engage early. That’s their style. A guy like Babchuk has been drilled to be in position-he hesitates.

    In other rare stretches you see Calgary getting out of their end quick with great outlets. Brodie’s absence has really been felt.

    On offense I’ve seen good direct play but also many bad turnovers at the attacking blueline. Especially from Glencross and Comeau. They do too much in the wrong areas.

    In this last game we saw both: dumb over handling early then a return to simple cycling that suits these forwards. Essentially they hold the puck but don’t do anything productive with it.

    Calgary lacks aggressive play, gives up time and space then they try to do too much to compensate.

    To me, if they were just more aggressive to pucks in their own end it would go a long way. They’ve done it and need to get back to it. This is when they are at their best.

    • I agree on your point about the overhandling of the puck in the O-zone. It seems like this is all Iginla does – get the puck in the neutral zone, then stick handle the crap out of it as he crosses the blue line. 99% of the time, he loses the puck and it’s back up the ice they go.

      When Phaneuf and Langkow were still in town, the Flames seemed fairly aggressive with the puck. It usually resulted in them skating up, Phaneuf taking a high and wide shot, and old man Langkow having to chase the other team back into their own zone. 2 on 1, 3 on 1, with Langkow our only defender.

      Just can’t win with this team. They get aggressive and score goals one year, but can’t stop a puck to save their lives. Next year, they play hesitantly and don’t score as often as they need, but the defense stops most pucks coming at them. It’s a real rollercoaster being a Flames fan. And not a fun, Six Flags rollercoaster – more like a haunted, abandoned amusement park rollercoaster made of wood and built over 100 years ago.

  • Sven Baertschi with a goal and an assist through a period of game 3 between Portland and Kelowna. He has 1G and 6A through 2 and 1/3 games of the playoff series. Kid is incredible. He’ll be a Flame next year, at least to start.

  • Greg

    Nice write up Kent, makes sense to see the flames have actually been playing better but just not getting the bounces. Sounds right given the number of shots and chances they’ve been generating and coming away empty handed.

    Question for you though: I’ve never looked into advanced stats much before (I just drink the koolaid you pour :), but I poked around a bit on behind the net tonight and looked at Iggy’s PDO stats the last few years in particular. From what I can tell, that’s not a real significant change for him there (108 to 104), and actually even the 108 number doesn’t seem at all out of line with his historical stats even with his progressive decline in recent years. I’m confused by the statement that PDO invariably regresses to 1000 for each player. I looked at datsyuk, Lidstrom, Sedin… Guys like this are almost always over and not regressing back to 1000.

    It seems to be implied that any PDO over 1000 means a guy was lucky, but there are guys like these are legitimately always over 1000, and of course guys not like these ones who are always below it too. So does any PDO over 1000 actually indicate luck? Or can it actually indicate a skill level when it’s in line with a player’s averages, and only luck when there’s a big fluctuation?

    I guess my real question is (other than trying to understand PDO), if you graded iggy at an A or a B, and then saw his PDO was 108, were you misled by lady luck, or just correctly observing that’s he’s still pretty good at least scoring goals?

    • From what I can tell, that’s not a real significant change for him there (108 to 104), and actually even the 108 number doesn’t seem at all out of line with his historical stats even with his progressive decline in recent years.

      108 is a number no single player maintains through a full season. The Luckiest players each year *might* settle at 103 or 104.

      Last year, Jarome’s PDO was 100.3. The year before that, it was 100.6. The year before that, it was 100.2. etc.

      So Im not sure where you’re looking, but you’re being deceived – Iginla, just like most NHLers, usually settles right around league average.