Quote:
Originally Posted by Macho0978
When you look at the breakdown of GA's by type for each team you can make a case the Flames goalies haven't been good this year.
GA at 5 on 5:
99 Total
88 SCGA
53 HDGA
So that works out too:
11 - GA that weren't a scoring chance
35 - Low or Medium scoring chance GA
53 - High Danger goals against
Only Seattle, Toronto and Nashville have given up less GA on non-scoring chances. I would think most of these would be bad goals.
23 teams have given up less goals that 35 on low and medium danger scoring chances. Flames are one of the worst in the league here. Medium danger save percent is bottom 5.
High Danger save percent is also bottom 10 in the league as well.
Too many goals this year on medium danger scoring chances against. Ones that you can say, can't blame him but it would be nice to get a save here once in a while.
I'd love to see the NHL break high danger chances into a few buckets. Not sure how they would do it, but goalies being set vs moving to make the save or contested chances for open chances. Not sure but it would be the next level for advanced stats.
How do people who say the Flames give up more terrible chances even though the stats say they don't vs what other teams do? Do you watch every single game or just the Flames. Our team does not create a ton of high danger chances so is it not possible that when teams play us, we think our d stinks vs other teams d but it could be our forwards making the d look good?
|
^ stats capture high or medium danger chances (as defined by the particular model). They absolutely do not capture ‘terrible chances’
Stats do not capture anything at all about the team defensive position or contribution.
They do not capture if the goalie is set or not
They do not capture pressure on the shooter and resulting shot placement
Etc.
Those are the kind of things that can make a terrible defensive breakdown become a goal
The Flames can hold a team to a low number of HDCA but sprinkle in a handful of those, and you can easily allow a goal and lose a game
The models are pretty good at comparing overall balance of play
Bingo appears to be assuming that these things are essentially equal for all teams. I personally don’t agree.
The key point is that the stats don’t refute it. They don’t measure it. It’s a limitation.
Remember the Chicago game? Just off the top of my head, the third and 4th goals
The goal where both Weegar and Zadorov covered Patrick Kane, who dished to Dickinson to go in all alone, in the slot?
The odd man rush where Hanifin inexplicably challenges the puck carrier at center ice, gets blown by, Huberdeau is the guy back and Coleman dove, Lafferty easily taps in a cross crease pass?
This team has made brutal costly mistakes all year on the defensive side of the puck
The advanced stats I am interested in have to do with zone entries, chances from rush vs cycle, odd man rushes, etc.
These danger models are an attempt to improve on raw save percentage, to group probabilities of goals by separating in to rough buckets. Good.
But they don’t really actually tell you if a goalie has been particularly good or bad. And absolutely do not capture terrible chances