MATH ARMY

by Lowetide

Last week David Johnson closed down Stats.HockeyAnalysis.com and thousands of hockey nerds like me wept. It was not good. Although I disagreed with Johnson’s conclusions many times, his site was beautiful. WOWY all night long, sort Fenwick like hell won’t have it, all kinds of answers in less than a minute. Johnson was hired by the Calgary Flames, no doubt to help them figure out how to outscore the Oilers during the McDavid era.

The entire episode brought me to a couple of questions that have been out there for a time and I’ve been thinking about these hot August nights.

  1. What kind of information should an organization be gathering?
  2. Who best to read the data?

Johnson knows the numbers and for all I know (I am serious here) is the absolute best at reading them. But how often is that true? You have to find the right numbers to crunch, make sure your math army is crunching them properly, and then have someone read the tea leaves properly. That’s a very difficult group of things for one person to do and I wonder how many teams are doing it right? Are we five years from hitting top dead center? 10 years?

HOW CAN WE TELL?

When Dellow was with the Oilers, I’d look at the moves made and imagine he  did or did not get heard on specific transactions. Andrej Sekera signed? I will never know, but that made sense and Dellow had a hall pass at the time. Winnipeg and not Edmonton trades up to get Logan Stanley? Oilers don’t take Tage Thompson or Riley Tufte? Maybe Dellow filed a report. But really, we’re pissing in the wind. David Johnson might be the smartest guy on the planet, might have the best metrics and may read them perfectly, but what if his recommendations sit on Craig Conroy’s desk unread? I’m not sure most NHL teams are helping themselves in this area and it’s going to be a bugger to figure it out.

WHAT WOULD YOU DO?

There’s a bunch of things I’d get a David Johnson to look at, beyond the stuff he’ll be able to comb out of the Flames data we don’t have access to because of the wall. Here are a few instances.

  • Oklahoma Barons general manager Bill Scott, January 2012: ‘Push Up’ is our organizational philosophy now. No one is getting ice time because they were picked in the first round or the second round. There’s no entitlement. They (Tyler Pitlick and Curtis Hamilton) both started with eight to 10 minutes a game and have been healthy scratches for games. Now they’re up to 10 or 12.” Source

I’d like Johnson to have a long look at the Oilers tendency to slow-play kids who are 20 and just out of junior. My brain tells me young athletes should play, and if the competition is too high then play them at a level where they can flourish. I don’t have a lot of information available but Johnson should have access now. How long would it take to build a 10-year history of kids at 20 in the AHL and how well they perform using the ‘Push Up’ method? If it works, wouldn’t we see more teams adopting it? I think you would see it. Or maybe the Oilers don’t do that anymore. I would love to ask Kyle Platzer about it.

DRILLING DOWN ON CHL DEFENSEMEN

CHL Forwards are easier to figure out than defenders, the stats available to us offer good insight, even boxcars tell a story. Defense? Hmm. Not so much. I’d ask a guy like Johnson to bin opposition for CHL teams and players like the big brains at Puck IQ. It can be done, the folks at Prospects-Stats.com are doing it. Here, let me run some Seattle Thunderbirds numbers to make my point.

  • Bear is clearly the class of the group, and plays the most while delivering the most while playing against the most difficult. Seems to me we can say the Ottenbreit and Tyszka are playing similar tough minutes and then the others are against the lesser. Now, let’s touch base with Caleb Jones and the Portland Winterhawks.

  • Jones has some company, in boxcars and qual comp (looks like two pairs sharing the load) but he played a lot more at 5×5 than his partners.
  • Bear delivered more offense, but the rest looks the same. Who had the better year?
  • A perhaps more important question: How good is Jokiharju if he posted similar numbers while being two  years younger?

I’d like to have many seasons to compare and the binning done by Woodmoneys to compare. If you had that, could you take advantage at the draft table? In trade? As long as you are asking the right questions, inputting the correct data and have the right savant, you would be wise to be aggressive. Right? How many teams are doing that right now? Before we leave the subject, I wanted to have a look at Edmonton’s defensive draft pick this year, Dmitri Samorukov from the Guelph Storm.

  • Guelph Storm have a fascinating defensive group. McFadden was 19 last season and he pulled the biggest load, followed by draft eligible Samorukov and Merkley, who was 16. Depth was a real issue and the defense got caved.
  • I like Samorukov’s numbers here but would love to see it against hundreds of other draft eligible seasons over the last decade. Do players who post this kind of player card at 17 develop into NHL  players 30 percent of the time? 40?

THE OILERS AND MATH

Although many were upset that the Oilers passed on Alex DeBrincat (I had him No. 15) at the 2016 draft to select Tyler Benson (I had him No. 34) (my final list for 2016 is here) the selection represented a massive improvement for the team’s second-round picks. Marco Roy, Mitch Moroz and David Musil were the three second rounders from recent drafts, not a one of them posting enough offense to suggest an NHL future. Benson? There is zero doubt he will bring enough offense if he can stay healthy, plus his resume suggests grit and two-way ability.

All of which is to say I think the Oilers have improved in this area. The forward selections this season (Kailer Yamamoto, Ostap Safin, Kirill Maksimov, Skyler Brind’Amour) are all home run picks if they cash. Edmonton is swinging so hard they’re coming out of their shoes with these selections. Music!

What is the next step? Well, and I’m not sure the organization still does it, but I’d move heaven and earth to find a way to close the experiment known as ‘Push Up’ as soon as humanly possible. If we read next fall about Tyler Benson’s healthy scratches and his 10-minutes a night playing time in Bakersfield, we’ll know the Oilers aren’t really paying attention to what successful teams are doing. At the very least have someone do the research on it. I would. They should. The Edmonton Oilers owe it to Tyler Benson. Seriously.

81 comments
0

You may also like

0 0 vote
Article Rating
81 Comments
Newest
Oldest Most Voted
Inline Feedbacks
View all comments
€√¥£€^$

VOR,

You are a very precise and interesting writer, I have always enjoyed what you write.

I look forward to reading your novel next year!

Woodguy v2.0

Bank Shot,

I generally think there isn’t one right method based on results. I generally think that the players that are going to make it will rise to the top in most situations.

I agree with that.

I’d still like to know what works best and think they should default to “play them”

Woodguy v2.0

VOR,

That’s just fucking awesome.

Renaissance man.

Gayfish

treevojo:
VOR,

I hope that I can give you credit for the line “fork or spoon”.

That was golden.

He definitely wrote “I’m the cook”.

treevojo

VOR,

I hope that I can give you credit for the line “fork or spoon”.

That was golden.

Gayfish

pocession charge:
VOR,

Very cool VOR.That’s a wide range of interests that you have.I’ll look forward to the new book, too. My apologies for not recognizing your name from porn….I usually turn it off before the credits roll

I usually turn it off before the opening credits roll.

Material Elvis

VOR,

Very cool VOR. That’s a wide range of interests that you have. I’ll look forward to the new book, too. My apologies for not recognizing your name from porn….I usually turn it off before the credits roll 🙂

Bank Shot

Woodguy v2.0:

So it doesn’t appear that ANA or TBY has a mandate to “play the kids and damn the consequences” as it doesn’t show up this year.

That’s the toughest part in trying to figure out the playing time/development model.

We look at models that give kids a bunch a playing time and they graduate to the NHL and we say “Ah ha!see, it works”, but what if those kids were just so good that they deserved to get the playing time?

Chicken and the egg in some respects.

I agree 100% that players should be playing and playing a lot at the 20-23 year old age.

I’m just not sure if you can turn a C prospect into a B and a B into an A with TOI.

I think the best argument is “we really don’t know if they are an A, B, or C with 100% confidence so play the shit out of them to find out”

Not an easy question to be comfortable that you have a definitive answer for.

I’d say the European leagues create more NHLer per capita in terms of registered players than the the USA or Canada.

The Euro leagues generally don’t hand anyone icetime. If you are good enough to play Swedish Elite league 8 minutes a night at 16, well you play 8 minutes a night at 16.

The Russian league is famous for not gifting young players huge icetime, and you’ll see the stats of of guys like Ovechkin and Malkin definitely aren’t eye popping for guys that blew the doors off the NHL.

I generally think there isn’t one right method based on results. I generally think that the players that are going to make it will rise to the top in most situations.

VOR

Woodguy,

A while ago you asked me what I had written. Under my own name a novel called Turning Samoan, hundreds of book reviews for major news chains (also some film, theatre and music reviews), several hundred essays, many interviews, in depth magazine articles, and a huge number of blog posts (I used to host and administer one of the most visited blogs in the natural sciences – it was dedicated to cloning – and as an aside did you know many of the horses running at tracks throughout North America are clones? I bet they don’t clone slow horses.)

Under several pen names I have written fifteen erotic novels. I have also worked extensively as a ghost writer, written plays, and several screenplays – none of which you will have heard of unless you watch a lot of porn – yes they have scripts. I have also written hundreds of short stories for various online erotica sites. I would probably proudly own all of this work but A) for some of it I am contractually bound, ghost work is like that, B) I wrote one of the classics of erotica (easily my most successful book) – and my fandom aren’t people you would want to know your real name C) for reasons I don’t fully grasp when people find out about my writing they take me less seriously as a scientist, a mathematician, a teacher, a coach, and a writer of business and marketing plans.

It doesn’t help that whether I am writing novels, erotica, essays, or screenplays I simply can’t help it, I am funny. People assume if you are funny you are an intellectual lightweight. So I try to keep the scientist/mathematician/entrepreneur separate from the writer.

In passing, I have a new novel out for summer 2018 called The President’s 51 which will resume my career under my own name.

Pink Socks

godot10: If Sam Bennett and Curtis Lazar have breakout years….

Calgary’s upside potential depends on the performance of those two guys.

Or maybe Jankowski or Klimchuck, I agree, they need something internally. Kind of like we are in dire need of JP or Strome or Slepy or Caggiula to break out. If I were a betting man I would put money on more upside for the Oilers’ group.

Pink Socks

frjohnk: They have some very nice prospects though.

I will laugh if the goaltending blows up in their face, they miss the playoffs and the Islanders win the lottery with Flames pick.

That would be a dream. Here’s hoping!

frjohnk

Pink Socks: The best part, is that at the trade deadline next year they have only their 2019 1st rounder left over as trade bait for some proven scoring talent.

They have some very nice prospects though.

I will laugh if the goaltending blows up in their face, they miss the playoffs and the Islanders win the lottery with Flames pick.

godot10

Pink Socks: I agree, they may challenge at the top of the Pacific, but their offensive numbers are average at best.The best part, is that at the trade deadline next year they have only their 2019 1st rounder left over as trade bait for some proven scoring talent.

If Sam Bennett and Curtis Lazar have breakout years….

Calgary’s upside potential depends on the performance of those two guys.

Pink Socks

Georges:
Calgary

Shot an unsustainable, league-high 9.16% at 5v5 post All-Star Game. Went 20-9-1 to salvage their season while going 7-7 against playoff teams.

In the playoffs, offense scored 2!! 5v5 goals in 4 games against ANA.

GM has decided to make a sideways (at best) move on goalies and to bring in a defenseman with bottom of the league on-ice numbers last season. Also signed Luke Gazdic.

/Calgary

I agree, they may challenge at the top of the Pacific, but their offensive numbers are average at best. The best part, is that at the trade deadline next year they have only their 2019 1st rounder left over as trade bait for some proven scoring talent.

Pink Socks

WG thank you for all the numbers, keep working the blog. 🙂

After watching the junior hockey over the past week, is mild altering drugs the only reason why the Jets would have drafted Logan Stanley, let alone moved UP to get him? His feet are in cement 100% of the time and eerily reminded me of a large, slow, mid first round 2007 Oilers draft pick.

Georges

Calgary

Shot an unsustainable, league-high 9.16% at 5v5 post All-Star Game. Went 20-9-1 to salvage their season while going 7-7 against playoff teams.

In the playoffs, offense scored 2!! 5v5 goals in 4 games against ANA.

GM has decided to make a sideways (at best) move on goalies and to bring in a defenseman with bottom of the league on-ice numbers last season. Also signed Luke Gazdic.

/Calgary

OriginalPouzar

Its gorgeous in YYC today and projected to be beautiful through next weekend!

Great summer weather wise.

58 days left until the season.

Woodguy v2.0

frjohnk: If they can be as healthy as last year, I believe they will move the needle.Everybody that plays significant minutes is eithera) young and should be better than last year ( ie Nurse, McDavid, Draisaitl, Drake, Sleppy)
b) just entering their prime ( ie Larsson, Nuge)
c) still in their prime/or just exiting ( ie Talbot, Lucic, Sekera, Russell)

Nobody on this roster is really close to the cliff.Internal growth from within should make this a better team.

With that said, if we have injury trouble like in 15-16, in which many of the top TOI roster players miss significant time, we will have trouble making the playoffs.

I wonder if Talbot can keep it going as well.

Even very good goalies are not that consistent from year.

Lots of the variance has to do with injuries (Price, Lundqvist, Schnieder etc) but even ones who have stayed healthy can be variable.

CAM.TALBOT 20142015 1.32
CAM.TALBOT 20152016 0.33
CAM.TALBOT 20162017 1.30

BRADEN.HOLTBY 20142015 0.66
BRADEN.HOLTBY 20152016 0.85
BRADEN.HOLTBY 20162017 1.34

COREY.CRAWFORD 20152016 1.14
COREY.CRAWFORD 20142015 1.03
COREY.CRAWFORD 20162017 0.63

HENRIK.LUNDQVIST 20142015 0.95
HENRIK.LUNDQVIST 20152016 1.27
HENRIK.LUNDQVIST 20162017 -0.20

JAROSLAV.HALAK 20142015 1.31
JAROSLAV.HALAK 20152016 1.34
JAROSLAV.HALAK 20162017 0.78

SERGEI.BOBROVSKY 20142015 0.31
SERGEI.BOBROVSKY 20152016 -0.79
SERGEI.BOBROVSKY 20162017 2.02

CAREY.PRICE 20142015 1.90
CAREY.PRICE 20152016 Injured-only 12 gp
CAREY.PRICE 20162017 0.77

CORY.SCHNEIDER 20142015 1.56
CORY.SCHNEIDER 20152016 1.46
CORY.SCHNEIDER 20162017 -0.27

Gayfish

Woodguy v2.0:
Sorry about dominating the comment section.

Cool rainy day and I’m trying to salvage some posts for my blog that I’ve already started but can’t finish with the Super WOWY gone.

Its not going well so I’m just playing around with data and sharing things.

Should get another day off when weather sucks on a holiday.

Woodguy v2.0

frjohnk: A healthy Benn and additions of Hanzel and Radulov should make them one of the powerhouse offensive teams.

Also, the Korpikoski experiment is over in Dallas as well.

They’ll need to outscore their goalies….. 🙂

Adding Methot will help and maybe Hamhuis has one more decent season left in him….

Korpse actually had a 55% GF last year in DAL btw.

Woodguy v2.0

Sorry about dominating the comment section.

Cool rainy day and I’m trying to salvage some posts for my blog that I’ve already started but can’t finish with the Super WOWY gone.

Its not going well so I’m just playing around with data and sharing things.

frjohnk

Woodguy v2.0: Nill better hope he can regain his form

A healthy Benn and additions of Hanzel and Radulov should make them one of the powerhouse offensive teams.

Also, the Korpikoski experiment is over in Dallas as well. But he was a +5 last year. Funny player, that Korpse.

frjohnk

OriginalPouzar: Can they possibly repeat that or potentially move the needle further this year?

If they can be as healthy as last year, I believe they will move the needle. Everybody that plays significant minutes is either
a) young and should be better than last year ( ie Nurse, McDavid, Draisaitl, Drake, Sleppy)
b) just entering their prime ( ie Larsson, Nuge)
c) still in their prime/or just exiting ( ie Talbot, Lucic, Sekera, Russell)

Nobody on this roster is really close to the cliff. Internal growth from within should make this a better team.

With that said, if we have injury trouble like in 15-16, in which many of the top TOI roster players miss significant time, we will have trouble making the playoffs.

Woodguy v2.0

Let’s see what DAL did with their goalering…..

Last year:
KARI.LEHTONEN 20142015 -0.80
KARI.LEHTONEN 20152016 -0.45
KARI.LEHTONEN 20162017 -0.64

ANTTI.NIEMI 20142015 -0.22
ANTTI.NIEMI 20152016 -0.61
ANTTI.NIEMI 20162017 -2.14

Niemi is bought out and they brought in Bishop:

BEN.BISHOP 20142015 0.31
BEN.BISHOP 20152016 0.57
BEN.BISHOP 20162017 -0.54

Bishop was hurt last year. Nill better hope he can regain his form…….

Woodguy v2.0

Woodguy v2.0,

Its not out of the realm that CAR improves by over 40 goals this year.

Speaking of tough divisions….

CAR can improve by 40 goals and still be 5th in their division.

Last year:
Washington Capitals 84
Columbus Blue Jackets 54
Pittsburgh Penguins 49
New York Rangers 37
New York Islanders 1
Carolina Hurricanes -18
Philadelphia Flyers -19
New Jersey Devils -61

Woodguy v2.0

Marc: Agreed.A taste of what is theoretically possible with this kind of data, again from the NBA: http://grantland.com/features/the-toronto-raptors-sportvu-cameras-nba-analytical-revolution/

Watch the little video and imagine being able to break down hockey plays like that.

I drool at the thought of data that complete.

I’ve read that piece before, but thanks for linking it.

Its great stuff.

Woodguy v2.0

John Chambers,

They are as much of a threat to win the Pacific as Edm or Anh.

Agreed, which is why in the post you quoted I wrote:

I have all 4 of the top teams very close (I have EDM below +30 as it stands) and LAK not far behind.

Marc

Woodguy v2.0: Puck and player movement is the next step in the NHL.

Companies like Sportslogiq have reportedly created programs that track puck and player movement via the tv broadcast.

Companies like TruPerformace have people manually tracking a large number of things.

From what I’ve been told and gleened there is roughly 3000 data points per game from these types of services.

The shot data we currently analyze is ~400 data points or so per game.

The challenge for NHL teams is to:

1) integrate this data
2) figure out which data is useful and which is not (i.e. which “skills” repeat from year to year, which don’t)

Steps one and two are massive and I’m not sure how many teams have the staff and knowledge to do it right.

I’d guess less than 10 based on conversations with people who work for teams.

Agreed. A taste of what is theoretically possible with this kind of data, again from the NBA: http://grantland.com/features/the-toronto-raptors-sportvu-cameras-nba-analytical-revolution/

Watch the little video and imagine being able to break down hockey plays like that.

frjohnk

John Chambers: They are as much of a threat to win the Pacific as Edm or Anh.

My thoughts as well.

Any 3 of those teams could be the Pacific Champion and I would not be surprised.

And while San Jose is getting older and lost Marleau, they still put up 99 points last year. I dont expect them to challenge for the Pacific title but could grab 3rd if one of the other 3 teams above falter.

Woodguy v2.0

CAR goalering situation:

Last year:
Cam Ward 1712 SA in 16/17
CAM.WARD 20142015 -0.63
CAM.WARD 20152016 -0.34
CAM.WARD 20162017 -0.79

Replaced by Darling:
SCOTT.DARLING 20142015 2.88
SCOTT.DARLING 20152016 0.13
SCOTT.DARLING 20162017 1.21

If we assume the same GSAA/100 year over year and Darling plays exactly the 61 gp that Ward did that’s 34 less goals against.

Ward is replacing Lack as the backup.
EDDIE.LACK201420151.01
EDDIE.LACK20152016-0.67
EDDIE.LACK20162017-0.36

If we give Ward Lack’s 20gp (492 SA) that’s 2 more goals against.

CAR was -18 goal diff last year, so that takes them to +14.

Add to that:
-Hanifin moving to 2nd pair with Faulk
-TVR anchoring the 3rd pair.
-Kruger taking some of the tough minutes from Staal as 3C
-Williams on RW.

Its not out of the realm that CAR improves by over 40 goals this year.

OriginalPouzar

Looking at the goal differential for the Oilers once again and it reminds me that they were full value for their place in the standings at the end of the season.

They were pretty much in the 7-12 range in the NHL all season long and their goal differential and team advanced stats basically had they in that range all year – from start to finish.

They were full value for their standings finish – not relying on luck and non-sustainable metrics.

Can they possibly repeat that or potentially move the needle further this year?

I miss Reggie!

Material Elvis

Woodguy v2.0: That was Cam Charron and “does your GM draft better than a potato”

Cam used a lot of Josh’s data for that riffing off something Rhys wrote.

Josh developed the first CHL database among other things and Rhys’ work on projecting CHL Dmen into the NHL is still the best imo.

Here’s the link to the first post on it that Rhys wrote: https://canucksarmy.com/2014/05/20/we-think-the-vancouver-canucks-may-have-a-scouting-problem/

Here’s Cam Charron’s post: https://canucksarmy.com/2014/06/26/sham-sharron-takes-over-all-30-draft-tables/

Cam Charron was hired by the Leafs in 2014.

A different Cam than the one that works for FLA.

Right! You have a great memory. Thanks WG!

John Chambers

Woodguy v2.0: I’m not sure about win, but they’re in the mix.

Goal dIff last year in the Pacific:

EDM +36
ANA +23
SJS +19
CGY +3
LAK -2
VAN -63
ARI -67

If they go to +13 off the hop with goalering and not considering upgraded 2nd and 3rd pair then I’d say they are right in the mix.

I have all 4 of the top teams very close (I have EDM below +30 as it stands) and LAK not far behind.

The Pacific is tough again.

Calgary was awful to start last season. They got dreadful goaltending from Elliott, and as we’ve witnessed, a team can crater under bad goaltending.

Unlike last season when their incoming goaltender had just backstopped a team to the Conference finals and now played behind Wideman and Engelland, Mike Smith is moving from a pitiful AZ blueline to a rock-solid defense corps.

If we isolate the back half of the season the Flames were a powerhouse. Now I hate them more than the devil himself, but I expect their GD to improve for a multitude of reasons, including familiarity with the coach’s system.

They are as much of a threat to win the Pacific as Edm or Anh.

Woodguy v2.0

pocession charge: Were those the guys who ‘drafted’ the top scorers available and cross referenced the players actually drafted?Like the most basic model possible but it proved to be a more prolific drafter than Vancouver’s scouting staff.Am I thinking of the right guys?

That was Cam Charron and “does your GM draft better than a potato”

Cam used a lot of Josh’s data for that riffing off something Rhys wrote.

Josh developed the first CHL database among other things and Rhys’ work on projecting CHL Dmen into the NHL is still the best imo.

Here’s the link to the first post on it that Rhys wrote: https://canucksarmy.com/2014/05/20/we-think-the-vancouver-canucks-may-have-a-scouting-problem/

Here’s Cam Charron’s post: https://canucksarmy.com/2014/06/26/sham-sharron-takes-over-all-30-draft-tables/

Cam Charron was hired by the Leafs in 2014.

A different Cam than the one that works for FLA.

Woodguy v2.0

rickithebear: What average WG

The defence allows a certain shot density that varies the working average a goalie performs around.
If the average is variant to the expected mean from the d. It has great value.
If it is the league average. It can have an error +/- 0 to 40%

Have a nice day WG and FrojohnK

The average is based on historical SV% of the same shot based on x,y location + shot type.

We are still missing pre-shot puck movement (especially important is “across the royal road”), screens, odd man rushes etc, but because those are baked into the historical numbers its probably not *too* far off.

Once we have that data and can create historical average SV% including those variables it will be better.

Its pretty good now imo and better than SV% as SV% doesn’t account for shot type and location at all.

Woodguy v2.0

Woogie63:
I have spent a career in one most data intense industries.I have worked for the largest company in the world in that industry.We bought all the best data and hired the best minds to help guide our decisions, fact based was our DNA.

8 years ago the granularity of that data took a quantum leap, our old computers could not process this level of data.What we have learned in the last 8 years makes us look dumb in the pervious years.We were dead wrong with the old data, but we all believed it, we all actioned it.

I wonder were the NHL’s data is on this same path.

Puck and player movement is the next step in the NHL.

Companies like Sportslogiq have reportedly created programs that track puck and player movement via the tv broadcast.

Companies like TruPerformace have people manually tracking a large number of things.

From what I’ve been told and gleened there is roughly 3000 data points per game from these types of services.

The shot data we currently analyze is ~400 data points or so per game.

The challenge for NHL teams is to:

1) integrate this data
2) figure out which data is useful and which is not (i.e. which “skills” repeat from year to year, which don’t)

Steps one and two are massive and I’m not sure how many teams have the staff and knowledge to do it right.

I’d guess less than 10 based on conversations with people who work for teams.

rickithebear

Woodguy v2.0:
Since we’re on the subject:

Top 15 Goals Saved Above Average/100 shots in the last season. All situations.

(I used to do it as GSAA/60 minutes, but it was pointed out that not all goalies see the same amount of rubber in 60 min so its best to do it per shot)

PlayerGSAA/100 SHOT
SERGEI.BOBROVSKY 2.02
PHILIPP.GRUBAUER 1.71
JOHN.GIBSON 1.54
AARON.DELL 1.52
MATTHEW.MURRAY 1.42
BRADEN.HOLTBY1.34
CAM.TALBOT 1.30
JIMMY.HOWARD 1.23
SCOTT.DARLING 1.21
ANTTI.RAANTA 0.93
JONATHAN.QUICK 0.90
ANDERS.NILSSON 0.86
JAROSLAV.HALAK 0.78
CAREY.PRICE 0.77
CRAIG.ANDERSON 0.67

What average WG

The defence allows a certain shot density that varies the working average a goalie performs around.
If the average is variant to the expected mean from the d. It has great value.
If it is the league average. It can have an error +/- 0 to 40%

Quick for example had 5 to 6 top 60 hd dmen for most of his career

Have a nice day WG and FrojohnK

Material Elvis

Woodguy v2.0:
Marc,

The NHL isn’t yet at a place (as far as we know) where major decisions are being made based on the analytics that amateur enthusiasts have developed over the last decade or so. As such, the limitations of those analytics – the unknown unknowns – haven’t been fully explored. Only when that has happened, when we have a clearer picture of what conventional hockey analytics are good for, and not so good for, will we be at the same place the NBA reached….ten years ago.

FLA hired 3 guys from Canucks Army blog specifically to develop drafting tools based on their data.

Josh Weissbock, Cam Lawerence were hired in 15/16 and they added Rhys Jessop last year.

Those 3 had some of the best info and data on drafting that I’ve read.

It will be interesting to watch how FLA drafts going forward.

Were those the guys who ‘drafted’ the top scorers available and cross referenced the players actually drafted? Like the most basic model possible but it proved to be a more prolific drafter than Vancouver’s scouting staff. Am I thinking of the right guys?

Woodguy v2.0

frjohnk: So one might expect the Flames goaltending to be around average.

That is better than I thought.

If that is the case, I would not be surprised if they win the Pacific.

I’m not sure about win, but they’re in the mix.

Goal dIff last year in the Pacific:

EDM +36
ANA +23
SJS +19
CGY +3
LAK -2
VAN -63
ARI -67

If they go to +13 off the hop with goalering and not considering upgraded 2nd and 3rd pair then I’d say they are right in the mix.

I have all 4 of the top teams very close (I have EDM below +30 as it stands) and LAK not far behind.

The Pacific is tough again.

Woodguy v2.0

pocession charge: Hey it worked for Sheldon Souray, right?Right?

Exactly.

Alienating your players always works out and doesn’t create any hard feelings.

Woodguy v2.0

Marc,

The NHL isn’t yet at a place (as far as we know) where major decisions are being made based on the analytics that amateur enthusiasts have developed over the last decade or so. As such, the limitations of those analytics – the unknown unknowns – haven’t been fully explored. Only when that has happened, when we have a clearer picture of what conventional hockey analytics are good for, and not so good for, will we be at the same place the NBA reached….ten years ago.

FLA hired 3 guys from Canucks Army blog specifically to develop drafting tools based on their data.

Josh Weissbock, Cam Lawerence were hired in 15/16 and they added Rhys Jessop last year.

Those 3 had some of the best info and data on drafting that I’ve read.

It will be interesting to watch how FLA drafts going forward.

rickithebear

Until i see a data base displaying

(Above 0%chance shots)/CAx,y

I know they do not get it!

Have a nice day LT.

frjohnk

Woodguy v2.0:
For fun, here’s the GSAA/100 data on some goalies/situations that are interesting.

If you want to see someone, just ask.

Smith (CGY)
MIKE.SMITH20142015-0.81
MIKE.SMITH201520160.03
MIKE.SMITH201620170.36

A decent 16/17, but at 35 I wouldn’t expect better.

He replaced Elliott:
BRIAN.ELLIOTT20142015-0.05
BRIAN.ELLIOTT201520162.05
BRIAN.ELLIOTT20162017-0.38

So if Smith is about the same as last year he should be good for 0.74 less goals per 100 shots.Add to that an upgrade in defence and CGY might be pretty stingy last year as they were already 8th in SA with a total of 2356.

Backup in CGY is Lack:
EDDIE.LACK201420151.01
EDDIE.LACK20152016-0.67
EDDIE.LACK20162017-0.36

Replaces Johnson:
CHAD.JOHNSON20142015-2.29
CHAD.JOHNSON201520160.30
CHAD.JOHNSON20162017-0.05

Lack didn’t like his goalie coach in CAR so maybe he bounces back a bit.Seems like a slight downgrade from Johnson, but it will be close.

Last year Elliotte had 1338 SA and Johnson had 969 (Gilles, Rittich and EN had the rest)

If we give Smith 60 starts and Lack 22 and give CGY 2356 SA (28.73/gm) and assume identical performance GSAA/100 we would see a ~11 goal improvement in terms of GA without accounting for an improved 2nd and 3rd pair.

Will be interesting to watch

So one might expect the Flames goaltending to be around average.

That is better than I thought.

If that is the case, I would not be surprised if they win the Pacific.

Woogie63

I have spent a career in one most data intense industries. I have worked for the largest company in the world in that industry. We bought all the best data and hired the best minds to help guide our decisions, fact based was our DNA.

8 years ago the granularity of that data took a quantum leap, our old computers could not process this level of data. What we have learned in the last 8 years makes us look dumb in the pervious years. We were dead wrong with the old data, but we all believed it, we all actioned it.

I wonder were the NHL’s data is on this same path.

Material Elvis

Woodguy v2.0:

Agreed on Halak as well.Snow seems to think punishing player is a good thing.

Hey it worked for Sheldon Souray, right? Right?

Woodguy v2.0

For fun, here’s the GSAA/100 data on some goalies/situations that are interesting.

If you want to see someone, just ask.

Smith (CGY)
MIKE.SMITH 20142015 -0.81
MIKE.SMITH 20152016 0.03
MIKE.SMITH 20162017 0.36

A decent 16/17, but at 35 I wouldn’t expect better.

He replaced Elliott:
BRIAN.ELLIOTT 20142015 -0.05
BRIAN.ELLIOTT 20152016 2.05
BRIAN.ELLIOTT 20162017 -0.38

So if Smith is about the same as last year he should be good for 0.74 less goals per 100 shots. Add to that an upgrade in defence and CGY might be pretty stingy last year as they were already 8th in SA with a total of 2356.

Backup in CGY is Lack:
EDDIE.LACK 20142015 1.01
EDDIE.LACK 20152016 -0.67
EDDIE.LACK 20162017 -0.36

Replaces Johnson:
CHAD.JOHNSON 20142015 -2.29
CHAD.JOHNSON 20152016 0.30
CHAD.JOHNSON 20162017 -0.05

Lack didn’t like his goalie coach in CAR so maybe he bounces back a bit. Seems like a slight downgrade from Johnson, but it will be close.

Last year Elliotte had 1338 SA and Johnson had 969 (Gilles, Rittich and EN had the rest)

If we give Smith 60 starts and Lack 22 and give CGY 2356 SA (28.73/gm) and assume identical performance GSAA/100 we would see a ~11 goal improvement in terms of GA without accounting for an improved 2nd and 3rd pair.

Those 10 goals would have brought them up to 9th in GA instead of 14th place last year.

Basically the GA would catch up to the SA.

Will be interesting to watch

Revolved

Marc,
LT,

Really interesting article and fitting story on the NBA. I would agree that top dead center is a long ways off for the NHL. Many great new innovations have been made in the last decade, but outside the NHL everything is still based on goals and assists. It will be difficult to get anywhere without more data and the NHL doesn’t seem very forthcoming.

Woogie63

Woodguy v2.0:
The two orgs that have promoted the most AHLers to the NHL in recent years are TBY and ANA.

When you look at TBY and their Syracuse/Norfolk teams that graduated a lot of players you see all /most of the top point players (and therefore top TOI)players being in their early 20’s.

You see the same on the in the ANA org with Norfolk/Syracuse (TBY and ANA swapped AHL affiliates after the 11/12 season)

I was listening to an interview with Trent Yawney who coached in the AHL for ANA and he was asked about not playing AHL vets ahead of the kids.

His answer was (paraphrased) “it wasn’t really a plan it more necessity because we didn’t have the money to sign AHL veterans”

Interesting.

Both of these team played a ton of kids and graduated many to the NHL.

Maybe they played their kids because the kids were the best on the roster for better or for worse?

Both teams would have 0,1 or maximinum 2 25yr+ in their top 10 in scoring for years.

Now?

Last year San Diego (ANA AHL affiliate) had 3 in the top 10 25+ with 4 players being 23.

Syracuse had 5 players in their top 10 being 25+ with another being 24 and 2 being 23.

So it doesn’t appear that ANA or TBY has a mandate to “play the kids and damn the consequences” as it doesn’t show up this year.

That’s the toughest part in trying to figure out the playing time/development model.

We look at models that give kids a bunch a playing time and they graduate to the NHL and we say “Ah ha!see, it works”, but what if those kids were just so good that they deserved to get the playing time?

Chicken and the egg in some respects.

I agree 100% that players should be playing and playing a lot at the 20-23 year old age.

I’m just not sure if you can turn a C prospect into a B and a B into an A with TOI.

I think the best argument is “we really don’t know if they are an A, B, or C with 100% confidence so play the shit out of them to find out”

Not an easy question to be comfortable that you have a definitive answer for.

The answer might be more in the “process”, than the game result at the AHL level. Young 20-22 year old men often need to learn new habits, how to take practice serious, how to train off the ice, how to eat the right foods, get the right sleep etc., these kids all know the right way to be a pro, but it is a major commitment.

The coaching and mentorship must be a huge part of taking a”B ish” prospect and moulding them into a serviceable pro athlete.