Calgarypuck Forums - The Unofficial Calgary Flames Fan Community - View Single Post - The trials and tribulations of Glen Gulutzan

Calgary4LIfe · 04-01-2018, 11:38 PM

It isn't like the Flames are JUST shooting from a distance. It is that a lot of the shots come from situations in which the goalie is already set in position. That is the biggest problem for the Flames. The top line is not showing any 'issues' scoring with career years - they are probably getting more shots off than they have in recent years, so the system must be ok, right?

Well, I would say that Gaudreau, Monahan and Ferland have some of the best offensive skill on the team (Ferland continues to have an incredibly underrated shot on these boards, for instance). They can pick corners. Add Tkachuk to that list maybe.

Backlund, Frolik, Bennett (yet), Jankowski, Lazar, Brouwer, Stajan, Hathaway, Shore, Stewart, Glass, Mangiapane, etc.,, etc., etc., are either at a lower skill level, or haven't developed enough in the NHL to have the necessary quick release and/or accuracy to finish in those situations. They require more of a mix of cross-ice one timers/snapshots, shots through better screens, or just plain ol' jam the net.

Flames don't attack the net enough, and they don't do nearly enough quick cross-ice passes. I think it is a huge reason why the PP sucks too. It doesn't help matters that the system relies on a 5 man zone entry, which corresponds to an opposing 5 man set defensive zone presence.

Yes, there are probably more shots than average coming from distance, but the issue here is that the Flames don't generate enough QUALITY scoring chances by my eyes. There is no advanced metric that measures quality - they measure quantity from certain areas of the ice, not quality.

Hartley's transition system was based on quick transitions both ways - offence and defence. It was trying to create odd-man rushes one way, while trying to limit them the other way. I find it perplexing that the 5 man zone entries is causing so many odd-man rushes against. Why use a system that appears to be harder to score with if it isn't really limiting the chances against?

There are three things teams are in the NHL:

1) Strong offensively
2) Strong defensively
3) Strong 2-way (this is what an elite team is)

Which bracket do the Flames fall into? The only 'elite' thing I can say they do is get shots on net. They rank high in shots on goal. They rank low in actual goals scored. They rank poorly in goals against and shots against too.

How people can say that Hartley employed a 'river hockey' system is astounding to me, when I think a strong case can be made that Gulutzan's system is worse in all facets - except getting shots on net.

I used to really believe in the advanced metrics until I started noticing how consistently Hartley's system was 'beating the odds', and diving into some reasoning, I get it. I now see why Gulutzan's "advanced metrics darling" system is also consistently beating the odds the other way - I really get it.

The one rationale I come up with as to why the confidence intervals from all the data over the last however many seasons' worth is so low is because quality matters. There is only a loose correlation that people take as some kind of law when it comes to shots on goal and win percentage, and I argue it is becoming increasingly more skewed as teams are 'playing for CORSI' (as the Dallas Eakins' led Edmonton Oilers confessed to!).

It is the amount of QUALITY chances. Sure, I would bet that there is a higher correlation with high danger chances for and against as a means of a predictive stat, but even then it is faulty as it in no way measures how 'good' those high danger chances are on most nights, and what kind of players are doing the majority of the shooting.

The metrics were pointing to the Flames 'about to break out', but the eye test did little to support it. The quality didn't go up, the Flames did a better job (IIRC) of limiting high danger shots against, but continued to allow a few too many 'high quality danger chances' against (like breakaways, easy cross-crease tap-ins where a defender was nowhere to be seen, etc).

It seems like I am attacking advanced metrics. This is not the case really. I just think that the Flames are being coached FOR CORSI, rather than come up with a system that works and allow CORSI to line up with the results.

Is it bad luck when you are always missing corners and five-holes on point blank shots? Or is the system faulty for only allowing you to HAVE to shoot for the corners and five holes?

Just something to think about. I think the wide acceptance of advanced metrics has been accepted too soon and the data is not yet at a mature enough time to be anything more than what it should be used as - additional information. Additional metrics that finds a way to OBJECTIVELY measure quality has to be developed. Until then, the 'predictive' element of CORSI and most of the other advanced stats (some of which are pure garbage, TBQH - like team PDO) are going to have very low confidence intervals. There is simply too many exceptions that point to CORSI being a poor predictor. Heck, when you see one single team being a consistent exception over the course of 82 (and more) games, and being that both ways, I think you can disregard any predictions based on CORSI. When teams make big changes - coaches, systems and players - it makes the predictive nature of CORSI beyond a season even less predictive.

So why then, do I see a coach that at least SEEMS to be trying to coach FOR CORSI? I hope this is the last I see of a coach doing something like this for the Flames.

04-01-2018, 11:38 PM	#4520
Calgary4LIfe Franchise Player Join Date: Jun 2010 Exp:	It isn't like the Flames are JUST shooting from a distance. It is that a lot of the shots come from situations in which the goalie is already set in position. That is the biggest problem for the Flames. The top line is not showing any 'issues' scoring with career years - they are probably getting more shots off than they have in recent years, so the system must be ok, right? Well, I would say that Gaudreau, Monahan and Ferland have some of the best offensive skill on the team (Ferland continues to have an incredibly underrated shot on these boards, for instance). They can pick corners. Add Tkachuk to that list maybe. Backlund, Frolik, Bennett (yet), Jankowski, Lazar, Brouwer, Stajan, Hathaway, Shore, Stewart, Glass, Mangiapane, etc.,, etc., etc., are either at a lower skill level, or haven't developed enough in the NHL to have the necessary quick release and/or accuracy to finish in those situations. They require more of a mix of cross-ice one timers/snapshots, shots through better screens, or just plain ol' jam the net. Flames don't attack the net enough, and they don't do nearly enough quick cross-ice passes. I think it is a huge reason why the PP sucks too. It doesn't help matters that the system relies on a 5 man zone entry, which corresponds to an opposing 5 man set defensive zone presence. Yes, there are probably more shots than average coming from distance, but the issue here is that the Flames don't generate enough QUALITY scoring chances by my eyes. There is no advanced metric that measures quality - they measure quantity from certain areas of the ice, not quality. Hartley's transition system was based on quick transitions both ways - offence and defence. It was trying to create odd-man rushes one way, while trying to limit them the other way. I find it perplexing that the 5 man zone entries is causing so many odd-man rushes against. Why use a system that appears to be harder to score with if it isn't really limiting the chances against? There are three things teams are in the NHL: 1) Strong offensively 2) Strong defensively 3) Strong 2-way (this is what an elite team is) Which bracket do the Flames fall into? The only 'elite' thing I can say they do is get shots on net. They rank high in shots on goal. They rank low in actual goals scored. They rank poorly in goals against and shots against too. How people can say that Hartley employed a 'river hockey' system is astounding to me, when I think a strong case can be made that Gulutzan's system is worse in all facets - except getting shots on net. I used to really believe in the advanced metrics until I started noticing how consistently Hartley's system was 'beating the odds', and diving into some reasoning, I get it. I now see why Gulutzan's "advanced metrics darling" system is also consistently beating the odds the other way - I really get it. The one rationale I come up with as to why the confidence intervals from all the data over the last however many seasons' worth is so low is because quality matters. There is only a loose correlation that people take as some kind of law when it comes to shots on goal and win percentage, and I argue it is becoming increasingly more skewed as teams are 'playing for CORSI' (as the Dallas Eakins' led Edmonton Oilers confessed to!). It is the amount of QUALITY chances. Sure, I would bet that there is a higher correlation with high danger chances for and against as a means of a predictive stat, but even then it is faulty as it in no way measures how 'good' those high danger chances are on most nights, and what kind of players are doing the majority of the shooting. The metrics were pointing to the Flames 'about to break out', but the eye test did little to support it. The quality didn't go up, the Flames did a better job (IIRC) of limiting high danger shots against, but continued to allow a few too many 'high quality danger chances' against (like breakaways, easy cross-crease tap-ins where a defender was nowhere to be seen, etc). It seems like I am attacking advanced metrics. This is not the case really. I just think that the Flames are being coached FOR CORSI, rather than come up with a system that works and allow CORSI to line up with the results. Is it bad luck when you are always missing corners and five-holes on point blank shots? Or is the system faulty for only allowing you to HAVE to shoot for the corners and five holes? Just something to think about. I think the wide acceptance of advanced metrics has been accepted too soon and the data is not yet at a mature enough time to be anything more than what it should be used as - additional information. Additional metrics that finds a way to OBJECTIVELY measure quality has to be developed. Until then, the 'predictive' element of CORSI and most of the other advanced stats (some of which are pure garbage, TBQH - like team PDO) are going to have very low confidence intervals. There is simply too many exceptions that point to CORSI being a poor predictor. Heck, when you see one single team being a consistent exception over the course of 82 (and more) games, and being that both ways, I think you can disregard any predictions based on CORSI. When teams make big changes - coaches, systems and players - it makes the predictive nature of CORSI beyond a season even less predictive. So why then, do I see a coach that at least SEEMS to be trying to coach FOR CORSI? I hope this is the last I see of a coach doing something like this for the Flames.