07-25-2018, 05:53 AM
|
#2221
|
Franchise Player
Join Date: Feb 2013
Location: Boca Raton, FL
|
Quote:
Originally Posted by Singularity
You guys are arguing about advanced stats, but a lot of the times we don't pass the eye test. We saw glimpses when the system worked great and everything looked awesome, but that happened rarely.
|
Don't apply it just to the Flames. Go look at some other teams and see how their eye test matches to their advanced stats. More often than not you'll see that there is a correlation. What we're arguing here is the chicken and the egg scenario where it's uncertain whether advanced stats lead to winning hockey, or winning hockey leads to strong advanced stats. I'm leaning more to the latter in that discussion.
__________________
Quote:
Originally Posted by ResAlien
If we can't fall in love with replaceable bottom 6 players then the terrorists have won.
|
|
|
|
07-25-2018, 08:02 AM
|
#2222
|
Owner
Join Date: Dec 2001
Location: Calgary
|
Quote:
Originally Posted by dustygoon
I tried to distill this down to these two parts. There is no doubt they had lots of high danger chances and low shooting percentage.
I like the thinking that “they let defenders get set”, but i don’t recall seeing this kind of pattern...is there analysis of this? Are you meaning they took too much time to get shot off? That is read/react in a high tempo game...not something that GG system would demand unless he told the non-Gaudreau players to dumb down the creativity (I think he limited Bennett to a simple game).
I wonder how many high danger chances came right after a low danger chance. Funnel to the net (low danger) and hope for rebounds, shots off end boards, tips (high danger).
|
Took too much time to get the shot off, yes, but also didn't move the puck across the zone into a pass getting the goalie and the defenders to move.
|
|
|
07-25-2018, 08:05 AM
|
#2223
|
Franchise Player
Join Date: Oct 2014
Location: Springbank
|
Quote:
Originally Posted by Bingo
Took too much time to get the shot off, yes, but also didn't move the puck across the zone into a pass getting the goalie and the defenders to move.
|
Too much time to create a shot was certainly the PP issue.
|
|
|
The Following User Says Thank You to GioforPM For This Useful Post:
|
|
07-25-2018, 10:19 AM
|
#2224
|
Franchise Player
Join Date: May 2004
Location: Helsinki, Finland
|
Quote:
Originally Posted by Cali Panthers Fan
Don't apply it just to the Flames. Go look at some other teams and see how their eye test matches to their advanced stats. More often than not you'll see that there is a correlation. What we're arguing here is the chicken and the egg scenario where it's uncertain whether advanced stats lead to winning hockey, or winning hockey leads to strong advanced stats. I'm leaning more to the latter in that discussion.
|
My pet peeve with corsi is the idea that it's so strongly connected to winning and can be used to predict winning.
No stat like corsi, which tracks a small subsection of what a team does, can be predictive of overall team success. This I think SHOULD be obvious. Corsi just isn't a predictive tool. It's an analytical tool. Stats like corsi are useful for analyzing what the team is doing wrong or what it's doing right, not how it's going to do in the future.
What ever a teams current success is, the secondary stats (including corsi) that they have were what got them there. The same corsi is not likely to produce different results for the same team. There's just no reason why it should. It's like saying that the same shooting percentage is going to produce more goals in the future. It's just not a proper way to use stats.
But, when wondering something like "why did the Flames have such a low shot percentage and missed the net a lot", secondary stats can help. As Bingo pointed out in this thread, using secondary stats like they should be used, we can rule out "shooting from too far away" as a problem. What ever the problem was, that wasn't it. It's not a super helpful piece of information on it's own, but it's clearly better to have it than not to have it.
|
|
|
The Following 5 Users Say Thank You to Itse For This Useful Post:
|
|
07-25-2018, 02:20 PM
|
#2225
|
Franchise Player
|
Quote:
Originally Posted by Cali Panthers Fan
Don't apply it just to the Flames. Go look at some other teams and see how their eye test matches to their advanced stats. More often than not you'll see that there is a correlation. What we're arguing here is the chicken and the egg scenario where it's uncertain whether advanced stats lead to winning hockey, or winning hockey leads to strong advanced stats. I'm leaning more to the latter in that discussion.
|
I don't think there is much uncertainty at all.
|
|
|
07-25-2018, 02:37 PM
|
#2226
|
Franchise Player
|
Quote:
Originally Posted by Itse
My pet peeve with corsi is the idea that it's so strongly connected to winning and can be used to predict winning.
No stat like corsi, which tracks a small subsection of what a team does, can be predictive of overall team success. This I think SHOULD be obvious. Corsi just isn't a predictive tool. It's an analytical tool. Stats like corsi are useful for analyzing what the team is doing wrong or what it's doing right, not how it's going to do in the future.
What ever a teams current success is, the secondary stats (including corsi) that they have were what got them there. The same corsi is not likely to produce different results for the same team. There's just no reason why it should. It's like saying that the same shooting percentage is going to produce more goals in the future. It's just not a proper way to use stats.
But, when wondering something like "why did the Flames have such a low shot percentage and missed the net a lot", secondary stats can help. As Bingo pointed out in this thread, using secondary stats like they should be used, we can rule out "shooting from too far away" as a problem. What ever the problem was, that wasn't it. It's not a super helpful piece of information on it's own, but it's clearly better to have it than not to have it.
|
The bolded is exactly what Corsi suggests. It suggests that bad luck is responsible for underperformance and regression to the mean should take care of under performance. It assumes that all shots have an equal chance of going in therefore a higher Corsi over time will = more goals and more wins.
|
|
|
07-25-2018, 03:02 PM
|
#2227
|
Franchise Player
|
7 years right around $4.5M with Hanifin would be a solid contract for the Flames in my eyes, it also wraps up the off-season almost perfectly:
CapFriendly.com Armchair-GM User-Generated Roster
FORWARDS (13)
Right wing: E. Lindholm ($4,850,000) - J. Neal ($5,750,000) - M. Frolík ($4,300,000) - A. Czarnik ($1,250,000) - T. Brouwer ($4,500,000)
Centre: S. Monahan ($6,375,000) - M. Backlund ($5,350,000) - M. Jankowski ($1,675,000) - D. Ryan ($3,125,000)
Left wing: J. Gaudreau ($6,750,000) - M. Tkachuk ($925,000) - S. Bennett ($1,950,000) - C. Lazar ($950,000)
DEFENSE (7)
Right: T. Brodie ($4,650,400) - T. Hamonic ($3,857,143) - M. Stone ($3,500,000) - D. Prout ($800,000)
Left: M. Giordano ($6,750,000) - N. Hanifin ($4,500,000) - B. Kulak ($900,000)
GOALTENDER (2)
M. Smith ($4,250,000) - J. Gillies ($750,000)
BUYOUTS
L. Bouma ($766,667) - R. Murphy ($137,500)
DETAILS
Roster Size: 22
NHL Salary Cap: $79,500,000
Cap Hit: $78,611,710
Cap Space: $888,290
Bonuses: $850,000
Cap Space After Bonuses (if all are met): $38,290
Kulak, Prout, Lazar, and Gillies can all be challenged by Foo, Mangiapane, Kylington, Andersson, Valimaki, and Rittich.
|
|
|
07-25-2018, 03:18 PM
|
#2228
|
Scoring Winger
|
Quote:
Originally Posted by ComixZone
7 years right around $4.5M with Hanifin would be a solid contract for the Flames in my eyes, it also wraps up the off-season almost perfectly:
CapFriendly.com Armchair-GM User-Generated Roster
FORWARDS (13)
Right wing: E. Lindholm ($4,850,000) - J. Neal ($5,750,000) - M. Frolík ($4,300,000) - A. Czarnik ($1,250,000) - T. Brouwer ($4,500,000)
Centre: S. Monahan ($6,375,000) - M. Backlund ($5,350,000) - M. Jankowski ($1,675,000) - D. Ryan ($3,125,000)
Left wing: J. Gaudreau ($6,750,000) - M. Tkachuk ($925,000) - S. Bennett ($1,950,000) - C. Lazar ($950,000)
DEFENSE (7)
Right: T. Brodie ($4,650,400) - T. Hamonic ($3,857,143) - M. Stone ($3,500,000) - D. Prout ($800,000)
Left: M. Giordano ($6,750,000) - N. Hanifin ($4,500,000) - B. Kulak ($900,000)
GOALTENDER (2)
M. Smith ($4,250,000) - J. Gillies ($750,000)
BUYOUTS
L. Bouma ($766,667) - R. Murphy ($137,500)
DETAILS
Roster Size: 22
NHL Salary Cap: $79,500,000
Cap Hit: $78,611,710
Cap Space: $888,290
Bonuses: $850,000
Cap Space After Bonuses (if all are met): $38,290
Kulak, Prout, Lazar, and Gillies can all be challenged by Foo, Mangiapane, Kylington, Andersson, Valimaki, and Rittich.
|
That's a little too tight to the cap, especially given that you're only rostering 22 players (not the max of 23). In reality, most teams like at least $1M in cap room to start the year so they can call up players to cover short-term injuries too.
It's appearing more and more likely that we have to buy-out Brouwer just to fill our 23 man roster and have a small amount of flexibility in-season. I don't see any other way around it, unless it involves a trade of Stone or Frolik.
|
|
|
The Following User Says Thank You to burnitdown For This Useful Post:
|
|
07-25-2018, 03:27 PM
|
#2229
|
Franchise Player
Join Date: Apr 2004
Location: I don't belong here
|
Quote:
Originally Posted by ComixZone
7 years right around $4.5M with Hanifin would be a solid contract for the Flames in my eyes, it also wraps up the off-season almost perfectly:
CapFriendly.com Armchair-GM User-Generated Roster
FORWARDS (13)
Right wing: E. Lindholm ($4,850,000) - J. Neal ($5,750,000) - M. Frolík ($4,300,000) - A. Czarnik ($1,250,000) - T. Brouwer ($4,500,000)
Centre: S. Monahan ($6,375,000) - M. Backlund ($5,350,000) - M. Jankowski ($1,675,000) - D. Ryan ($3,125,000)
Left wing: J. Gaudreau ($6,750,000) - M. Tkachuk ($925,000) - S. Bennett ($1,950,000) - C. Lazar ($950,000)
DEFENSE (7)
Right: T. Brodie ($4,650,400) - T. Hamonic ($3,857,143) - M. Stone ($3,500,000) - D. Prout ($800,000)
Left: M. Giordano ($6,750,000) - N. Hanifin ($4,500,000) - B. Kulak ($900,000)
GOALTENDER (2)
M. Smith ($4,250,000) - J. Gillies ($750,000)
BUYOUTS
L. Bouma ($766,667) - R. Murphy ($137,500)
DETAILS
Roster Size: 22
NHL Salary Cap: $79,500,000
Cap Hit: $78,611,710
Cap Space: $888,290
Bonuses: $850,000
Cap Space After Bonuses (if all are met): $38,290
Kulak, Prout, Lazar, and Gillies can all be challenged by Foo, Mangiapane, Kylington, Andersson, Valimaki, and Rittich.
|
That Cane prediction site thingy is predicting that Hanifin will sign for 2 years at $2.3M. However, according to his algorithm he'll be around $4.44M for a 5,6 or 7 year term. So your guess is potentially close. I'm not sure how he predicts a 5 year term is about $3500 more per season than a 6 year term and a 7 year term is the same as 6. I would assume it goes up a bit more than that from 5 to 6 and 6 to 7.
|
|
|
07-25-2018, 03:45 PM
|
#2230
|
Franchise Player
|
Quote:
Originally Posted by Buff
So your guess is potentially close. I'm not sure how he predicts a 5 year term is about $3500 more per season than a 6 year term and a 7 year term is the same as 6. I would assume it goes up a bit more than that from 5 to 6 and 6 to 7.
|
I imagine it's heavily based on comparable contracts for those years. Of course age, point production, UFA years, etc. would be needed to take into account but I wonder if a defenseman of his stats signing 5 years would be using comparables such as:
Brodie (5 years at 4.7M)
Tanev (5 years at 4.5M)
Savard (5 years at 4.3M)
While 6 years would have:
Gostisbehere (6 years at 4.5M)
Brodin (6 years at 4.2M)
Pesce (6 years at 4.0M)
And 7 years would have:
Klingberg (7 years at 4.3M)
Klefbom (7 years 4.2M)
Josi (7 years at 4M)
And that's why the numbers don't go up as much for buying out additional UFA years as logic would dictate.
|
|
|
07-26-2018, 09:42 AM
|
#2231
|
Franchise Player
Join Date: May 2004
Location: Helsinki, Finland
|
Quote:
Originally Posted by GGG
The bolded is exactly what Corsi suggests. It suggests that bad luck is responsible for underperformance and regression to the mean should take care of under performance. It assumes that all shots have an equal chance of going in therefore a higher Corsi over time will = more goals and more wins.
|
Corsi does no such thing. Corsi is just a number. It does not suggest anything nor does it assume anything. It's just a piece of data about one specific thing. Any assumptions, suggestions or interpretations people make about that data is their own idea of what the number means.
There is absolutely nothing baked in Corsi that suggests it in any way measures luck.
Heck, even PDO, which was originally developed to specifically measure "luck" doesn't actually do that. Nor does it regress to mean either. (As proven by statistical analysis.)
The problem with advanced stats isn't the stats. It's people who think they know what they mean and who think it's easy to draw the right conclusions from them. Making the right conclusions about statistical data is actually generally really hard. Most scientists have trouble at it, and they're specifically trained for it.
It's a big reason why I think adding stats might actually degrade decisions coaches and managers makes. They are not trained to process lots of statistical information, which means they are very, very likely to draw the wrong conclusions from it. (Luckily teams have started to hire people for that work... Unfortunately some of them have pretty dodgy qualifications for professional statistical analysis.)
|
|
|
The Following 12 Users Say Thank You to Itse For This Useful Post:
|
badger89,
bdubbs,
blender,
Finger Cookin,
Flames Draft Watcher,
FlamesAddiction,
Hey Connor, It's Mess,
jayswin,
Lanny_McDonald,
mikephoen,
Pellanor,
Textcritic
|
07-26-2018, 10:01 AM
|
#2232
|
Franchise Player
|
Quote:
Originally Posted by Itse
Corsi does no such thing. Corsi is just a number. It does not suggest anything nor does it assume anything. It's just a piece of data about one specific thing. Any assumptions, suggestions or interpretations people make about that data is their own idea of what the number means.
There is absolutely nothing baked in Corsi that suggests it in any way measures luck.
Heck, even PDO, which was originally developed to specifically measure "luck" doesn't actually do that. Nor does it regress to mean either. (As proven by statistical analysis.)
The problem with advanced stats isn't the stats. It's people who think they know what they mean and who think it's easy to draw the right conclusions from them. Making the right conclusions about statistical data is actually generally really hard. Most scientists have trouble at it, and they're specifically trained for it.
It's a big reason why I think adding stats might actually degrade decisions coaches and managers makes. They are not trained to process lots of statistical information, which means they are very, very likely to draw the wrong conclusions from it. (Luckily teams have started to hire people for that work... Unfortunately some of them have pretty dodgy qualifications for professional statistical analysis.)
|
I'm going to disagree that Corsi is just the number of shot attempts directed at the net. The intended use of what Corsi means is inherent to the number. When people say Corsi they don't just mean the difference in shot attempts, the underlying assumptions that give it value are included in any statement.
The validity of these assumptions is definitely questionable but the assumptions come with the number to form what Corsi is.
THis might just be semantics though.
|
|
|
07-26-2018, 03:44 PM
|
#2233
|
In the Sin Bin
|
Quote:
Originally Posted by Itse
Corsi does no such thing. Corsi is just a number. It does not suggest anything nor does it assume anything. It's just a piece of data about one specific thing. Any assumptions, suggestions or interpretations people make about that data is their own idea of what the number means.
|
I very much agree with most of what you've said.
However Corsi was presented as a "possession" stat. It does not directly measure possession. Therefore the stat has been miscast and misleading from the very beginning by the way it was presented initially. I think its application and interpretation has been questionable from the very beginning and I think thats why a lot of people are skeptical of advanced stats and their value. Surely someone can compile actual possession numbers based on time spent controlling the puck. That would actually measure possession. Corsi meanwhile should have been presented as merely a shot differential and nothing else. The advanced stats people hurt their case right from the start by using stats that measure something to stand for something else.
Last edited by Flames Draft Watcher; 07-26-2018 at 03:46 PM.
|
|
|
The Following User Says Thank You to Flames Draft Watcher For This Useful Post:
|
|
07-28-2018, 11:19 AM
|
#2234
|
Resident Videologist
Join Date: Mar 2002
Location: Calgary
|
Starting to look like a long term deal for Hanifin is going to be more and more expensive.
Montour: 24 years old, 1.5 years in the NHL
38 pts in 107 games (0.36 PPG), coming off a 32 pts in 80 game season.
Signed for 2 years, AAV $3.39M.
Skjei: 24 years old, 2+ years in the NHL
64 pts in 169 games (0.38 PPG), coming off a 25 pts in 82 game season.
Signed for 6 years, AAV $5.25M.
Hanifin: 21 years old, 3 years in the NHL
83 pts in 239 games (0.35 PPG), coming off a 32 pts in 79 game season.
Last edited by AC; 07-28-2018 at 11:23 AM.
|
|
|
07-28-2018, 02:26 PM
|
#2235
|
First Line Centre
|
Quote:
Originally Posted by AC
Starting to look like a long term deal for Hanifin is going to be more and more expensive.
Montour: 24 years old, 1.5 years in the NHL
38 pts in 107 games (0.36 PPG), coming off a 32 pts in 80 game season.
Signed for 2 years, AAV $3.39M.
Skjei: 24 years old, 2+ years in the NHL
64 pts in 169 games (0.38 PPG), coming off a 25 pts in 82 game season.
Signed for 6 years, AAV $5.25M.
Hanifin: 21 years old, 3 years in the NHL
83 pts in 239 games (0.35 PPG), coming off a 32 pts in 79 game season.
|
The Skjei contract looks a bit high, but does have 3 UFA years. So as a comparable, it's probably equivalent to 4.75M/RFA year and 5.75m/UFA year (assuming the usual 1M/yr overpay on UFA deals). The Montour deal, being a bridge deal, is 3.39M/RFA year. If you use those 2 as comparables, Hanifin should come in at 4.07M/RFA year (just averaging the 2 quoted comparables) and 5.07M/UFA year, which would work out to 4.4M/yr on a 6 year deal. Alternatively, if he wants to bridge instead, probably just a cc of Montour's deal.
|
|
|
07-28-2018, 02:55 PM
|
#2236
|
Powerplay Quarterback
|
Won't hanifin be using Brady Skjei and his contract as a comparable?
|
|
|
07-28-2018, 04:18 PM
|
#2237
|
First Line Centre
|
Quote:
Originally Posted by AC
Starting to look like a long term deal for Hanifin is going to be more and more expensive.
Montour: 24 years old, 1.5 years in the NHL
38 pts in 107 games (0.36 PPG), coming off a 32 pts in 80 game season.
Signed for 2 years, AAV $3.39M.
Skjei: 24 years old, 2+ years in the NHL
64 pts in 169 games (0.38 PPG), coming off a 25 pts in 82 game season.
Signed for 6 years, AAV $5.25M.
Hanifin: 21 years old, 3 years in the NHL
83 pts in 239 games (0.35 PPG), coming off a 32 pts in 79 game season.
|
Skjei has 3 UFA years vs 2 so maybe the way to use this as a comparable is to knock around 2M off the contract overall. 4.9M AAV or 29.4/6
Though his camp may value the difference between an RFA and UFA as less.
Both of these comparables had arbitration rights, so Treliving can wait Hanifin out more to bring the cost down.
Worth noting that the difference between Skjei and Montour is close to the buyout hit of Brouwer and a 4th liner on an ELC
Last edited by Imported_Aussie; 07-28-2018 at 04:20 PM.
|
|
|
07-29-2018, 10:28 AM
|
#2238
|
First Line Centre
|
Quote:
Originally Posted by Imported_Aussie
Skjei has 3 UFA years vs 2 so maybe the way to use this as a comparable is to knock around 2M off the contract overall. 4.9M AAV or 29.4/6
Though his camp may value the difference between an RFA and UFA as less.
Both of these comparables had arbitration rights, so Treliving can wait Hanifin out more to bring the cost down.
Worth noting that the difference between Skjei and Montour is close to the buyout hit of Brouwer and a 4th liner on an ELC
|
The Skjei contract is definitely the high water for comparables, likely because of arb rights and the 3 UFA years, plus he's on the Rangers' first pairing. Hopefully, BT can keep the cap hit under 4.5 per year for 6 years, although this will still likely necessitate a trade or buy-out for a bit of roster flexibility. Interesting that Pulock, who has similar numbers in a 68-game season and is not arbitration eligible, only got 2M per year from the Islanders for a 2-year bridge contract. This does give Treliving some leverage in the discussion of long-term vs. bridge contracts, because I'm sure Hanifin does not want the Pulock contract.
Last edited by Macindoc; 07-29-2018 at 12:47 PM.
|
|
|
07-29-2018, 10:38 AM
|
#2239
|
Pent-up
Join Date: Mar 2018
Location: Plutanamo Bay.
|
Quote:
Originally Posted by AC
Starting to look like a long term deal for Hanifin is going to be more and more expensive.
Montour: 24 years old, 1.5 years in the NHL
38 pts in 107 games (0.36 PPG), coming off a 32 pts in 80 game season.
Signed for 2 years, AAV $3.39M.
Skjei: 24 years old, 2+ years in the NHL
64 pts in 169 games (0.38 PPG), coming off a 25 pts in 82 game season.
Signed for 6 years, AAV $5.25M.
Hanifin: 21 years old, 3 years in the NHL
83 pts in 239 games (0.35 PPG), coming off a 32 pts in 79 game season.
|
I think Hanifin settles right in the middle there. Which I would be thrilled about if it was 6/7 years.
|
|
|
07-29-2018, 11:28 AM
|
#2240
|
Owner
Join Date: Dec 2001
Location: Calgary
|
Quote:
Originally Posted by Flames Draft Watcher
I very much agree with most of what you've said.
However Corsi was presented as a "possession" stat. It does not directly measure possession. Therefore the stat has been miscast and misleading from the very beginning by the way it was presented initially. I think its application and interpretation has been questionable from the very beginning and I think thats why a lot of people are skeptical of advanced stats and their value. Surely someone can compile actual possession numbers based on time spent controlling the puck. That would actually measure possession. Corsi meanwhile should have been presented as merely a shot differential and nothing else. The advanced stats people hurt their case right from the start by using stats that measure something to stand for something else.
|
I agree ... but only to a point
Its not a big leap to suggest you have to have the puck more in order to put shot attempts on net.
Having said that I agree with the rest. It's additional information, and that's all.
It silly to use it as the only way to measure hockey players, and we see that a lot. But it's also stupid to ignore it when trying to determine how players are playing beyond counting stats that can be somewhat limited in their accuracy in projecting consistency season to season.
I always go back to Mikael Backlund on this topic because his progression in Calgary has always fit the window of advanced stats' explosion. He was good at a lot of things, you could see it with the eye test, but he wasn't producing. Many thought he'd be in Europe after a bridge contract. Advanced stats came along and backed up everything I was seeing.
On the other side you have Kris Russell. He went the other way for me. Eye test was warrior and high on try. I tried to make the advanced stats back that up in some way, but the deeper I went the more I couldn't protect the guy from the fact that he gets owned when he's on the ice. When he gets in trouble he's the first to stand up and do his best to prevent a goal, which is admirable, but his work in backing in, gap control, and collapsing creates the danger that he so admirably tries to thwart.
So it's a tool and it's helped me test a lot of what I see. It's not useless at all, but it's not everything.
|
|
|
Posting Rules
|
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts
HTML code is Off
|
|
|
All times are GMT -6. The time now is 07:19 AM.
|
|