|
Franchise Player
|
I didn't mind Bingo's graph. Took a little bit figuring out what was happening, but it is a decent graph that informs. What I hate is some graphs that really don't tell much AND are complicated - those make me feel like I just wasted my time and effort.
I always harp on counting stats, as I think they are used beyond their usefulness in order to prove something or other. Yes, they DO often paint the correct picture and align with what you see, but they also often don't.
For instance, Hartley's system was not 'stat friendly'. Why? The whole system was designed on trying to increase uncontested high danger chances for, while limiting them against. The counting stats might argue that Hartley's system was horrid at limiting high danger chances against simply due to some of the areas of the ice where shots originated, but most were highly contested (Flames were fairly good at clogging-up those areas and getting sticks/bodies in the way). In order to do this, they allowed perimeter shots. System worked like a charm for 1.5 seasons while Hiller and Ramo were 'average' goalies able to stop these low percentage shots. Yet the analytics community started heralding this as 'unsustainable'. I disagreed (and even an article by Bingo disagreed, IIRC).
Then came Hartley's last season. What happened? I think quite a number of low percentage shots started going through. Hiller in particular looked simply awful. Ramo was bad to start, and got demoted. However, when he got recalled, Ramo was excellent and stopping those shots, and the Flames started experiencing success again. Once he was injured, that was it - Ortio and Hiller weren't NHL-level goalies any longer, never mind simply 'average'.
I criticize Gulutzan's system quite a lot as it seemed to be a waste of effort in a lot of areas. Where?
Too much effort in trying to decrease the low percentage shots, and too much effort in trying to increase their shot count, without taking into consideration contested vs uncontested shots. For me, a contested shot is one where the shooter has to worry about an opposing player's stick or body in the way attempting to block the shot, and/or the goalie being set in position.
What happens when we see a goal against occur when a Flames' goalie is set in position and square to the shooter? Suddenly it is a soft goal usually, right? These happen from time to time for every goalie in the NHL, and of course you want to limit those, but the most important thing is to limit the high danger areas from having any shots on goal, and if they happen, to contest them as much as possible. Gulutzan's system was poor defensively more often than not in this manner. I argue that the last season of Hartley's tenure was also poor in this area, and it MAY have been a case of seeing too many soft goals go in and the Flames trying to play too aggressively on defence but not having the skill to do so.
What I really picked up on late last season was something that seemed to plague the Flames under Brent Sutter and often under Glen Gulutzan - teams would make the opposing goalies look 'spectacular'. Tonnes of shots against, goalies finishing the game with .955+ percentages... DelouxeMoustache's thread in the off-season about the Royal Road really got me to put it into perspective.
The Toronto game was a great example. Andersen played out of his mind - that's a legitimate strong goalie almost stealing a game. He had to move laterally OFTEN during the game. Against Buffalo? That was a stellar performance by the opposing goalie, with an actual danger of him stealing the game.
Too many nights under Gulutzan and B. Sutter did we see this team flounder in the offensive zone. Sure, the shots against might have significantly been in Calgary's favour, but it was a 'spectacular' goaltending display by the opposing goalie to 'steal' a game. Whey were so many games 'stolen'? Were they really stolen, or were the goalies just 'busy but not tested'?
There are games this season in which I was not impressed. The Flames did little to really make the goalie work. There were games where I felt that the opposing goalie was going to get injured with how often he had to stretch in order to stop the Flames.
What I think is that Peter's system is really a night and day difference in how they are approaching scoring. I do think that some of the Flames have been a bit guilty in over-passing and trying too hard to be more dynamic offensively and create opportunities in which the opposing goalie has to really move, but without question the Flames are MUCH better at doing so this season than last. I don't think that this has been driven exclusively from using their speed and the puck moving ability of the defencemen to generate odd-man rushes (which has certainly helped the goal scoring department this year, rather than waiting to enter the zone as a 5-man group). I just see many more plays where the puck is going against the grain of the defence and goalie, allowing for more tap-ins. I also see the Flames crashing the net harder.
Now how do you count these stats? This is where the analytics falls short - from switching from an objective count to a subjective one. There has been some attempt from what I can tell, but measuring anything subjectively is difficult and usually results in data-sets that are unreliable. I think the analytics community just needs to get together a bit more and think of a way to maybe assign values to stuff like 'tap-ins', 'number of shots generated from crashing the net', 'screen vs no screen', 'deflection', etc.
Simply assigning a value based MOSTLY on where the shot was generated from without taking much note of where the goalie was positioned, where the pass originated from, and if the shot was contested or not is labelling systems like Gulutzan's "sustainable but unlucky" while labelling systems like Hartley's as "unsustainable and lucky".
I actually really look forward to the continued evolution of analytics in hockey. I think that CORSI is something to note, but it often (and sustainably, in my opinion) does not show which team was the better team.
I think a great example was late last season against the Rangers where the Flames (IIRC) got blown-out. Why? They were annihilating the Rangers (who just somewhat finished a sell-off at the deadline, IIRC) in CORSI and possession (I mean actually having possession of the puck, not just what CORSI was indirectly showing). However, it seemed that every time that the Rangers got the puck, it turned out to be a high-danger shot against that went uncontested, and often turned into a goal. The Flames? Never looked dangerous. Tonnes of possession, tonnes of shots (IIRC), but never 'looked' dangerous and were being robbed yet again by spectacular goaltending.
A game like that should be some sort of a model for figuring out how to quantify that analytically. I do think that analytics will for sure get there, but I just don't think it is here yet. The current numbers are interesting for sure, and they will ALWAYS show 'something', but you need to watch the game to figure out for yourself what that something was. Did the Flames REALLY dominate? Did they really get a tonne of chances? Were they just throwing pucks on net hoping for a soft goal? Did the opposing goalie have the game of the season? Did the Flames shooters try to pick corners because shots were contested every time, or did they suddenly forget where exactly the net is?
I do, however, HAVE to say that Peter's system does not seem to be what I call "playing for CORSI" (hello Gulutzan). It does actually generate legitimate dangerous opportunities offensively. Defensively they need to tighten-up more - no surprises there.
I am seeing a guy like Brodie start to excel. Giordano is looking much more dangerous. Hanifin looks great in this system, as does Andersson and Valimaki. I also think Hamonic is doing great for a more defensive-defencemen, but due to his skating and ability to make a strong pass. Stone is struggling more due to his skating, IMO, and is getting passed by the rookies (which is an excellent problem to have, of course).
IF the Flames find a way to tighten-up more defensively, while skating hard and generating like they do offensively, I think they can go very far, and the analytics will correspond well to it. I feel they did not correspond well to what I saw under Gulutzan for long stretches, nor did I feel they corresponded well under Hartley.
Too many times under Hartely I felt: "Sure, they got out-shot, but it wasn't like the Flames' goalie had to steal that game. The other team didn't look or feel as dangerous as the Flames did, and the right team won." Too many times under Gulutzan I felt: "We significantly out-shot the opposing team, but I feel we weren't very dangerous tonight, and the other team was way more dangerous." - and I felt that way too often in both wins and losses.
This season, I felt the Flames were lucky to have won in NY, lucky to not have been blown out in Montreal, a bit unlucky to have been blown out THAT much by Pittsburgh (but should still have lost the game handily), unlucky not to have blown Toronto out, unlucky not to have blown Buffalo out...
I think the metrics are better aligning with what we see happening on the ice, and to me, that's sustainable improvement over last season.
|
