Past Studies Done Here (Links Only)

[Czech Your Math]

This is meant to be a resource where posters may:

- post links to and the results of studies which utilize the mathematical analysis of hockey data and statistics

- access the variety of studies which may be presented here

If you wish to give substantial feedback on a particular study, please do so in the thread for that study or on the general discussion thread located here:

http://hfboards.mandatory.com/showthread.php?t=1232177

If you do not value these studies in general and/or any specific study, then this is not the place to show your lack of appreciation. Non-constructive criticism may be deleted by the moderators.

It might be best for the authors of studies to use their first (or one of their first) posts to link to some or all of the studies which they wish to share with others. Then each author can update that post upon completion of any further studies. This would allow readers to access the maximum amount of specific studies in the shortest time and with the least amount of effort. Also, I was asked to start this thread, at least in part as a matter of circumstance, and am not necessarily assuming the responsibility of updating one central post containing the combined works of various authors.

Click to read more...

 

[Czech Your Math]

This is a study of a fixed group of higher scoring players from 1946 to ~2007, and how they performed from season to season. It needs some improving, including more complete data, especially since the lockout, but overall I believe the methodology and results hold a lot of promise:

Improving Adjusted Scoring and Comparing Scoring of Top Tier Players Across Eras

The results may be used to A) assist in comparing offensive production across different seasons and eras, and B) see how "simple" adjusted goal/point data tends to help/hurt various seasons and eras.

This is sort of a subsequent, companion study, but totally separate methodology. It uses linear regression to study some factors which seem to most affect how the very top group of players' scoring fluctuates:

Using Regression to Adjust "Adjusted Points" for Top Tier Players 1968-Present

Other studies:

Adjusted Playoff Scoring

Best "Half Seasons" Since 1994

An Estimate of How the Available Hockey Population Pool Has Changed Over Time (Focuses on Goalies and Top Line Scorers)

 

[overpass]

Here are a few statistical pieces I've posted in the past in the history section.

Historical adjusted plus-minus

With or without you: Mario Lemieux

Historical special teams usage

Forwards playing the point on the power play

Historical strength of opposition for goaltenders

Best penalty killing defencemen post-expansion

Best power play defencemen post-expansion

Sakic vs Forsberg (even strength)

The Triple Crown Line

Home/road plus-minus splits

 

[Hockey Outsider]

Comments/disclaimers

1. A lot of these studies are several years old. I'm posting some of them for reference purposes. As of now (July 2012), I don't necessarily agree with the methodology or results in some/many of these posts.
2. The data used in many studies are now at least a few years out of date. I'm not sure when (if ever) I'll get around to updating them.
3. I shudder to think about how many hours these all took... at least it's been spread over 7.5 years so it's not too bad on a yearly basis...

Analysis of Award Voting

• Third team all-stars
• Hart trophy analysis
• Hart trophy raw data
• Norris trophy: analysis
• Norris trophy: raw data
• Gretzky & Lemieux's impact on stats & awards

Offensive statistics

• Adjusted scoring statistics
• Top twenty scorer post expansion: a list of the players who ranked in the top five, ten, fifteen and twenty the most often from 1968 to 2009

Goalie statistics

• Adjusted Save Percentage
• Adjusted Playoff Save Percentage
• Wins Added
• Wins Added 2009-2012

"Applied" case studies

• Gretzky vs Lemieux – why it helps to retire young
• 1960s Chicago Blackhawks -- a team that struggled due to a weak supporting cast while Hull, Pilote, Hall and Mikita took too much blame
• Marcel Dionne – detailed analysis -- a few bright spots, but generally as bad as you would expect
• Tony Esposito – it’s not his fault his teammates couldn’t score
• Bill Durnan -- why does he have such a poor reputation in the playoffs?
• Henri Richard – it looks like Harvey & the first-line scorers, rather than Richard, improved the most in the playoffs
• Joe Sakic – a look at his impact on Colorado's record

Miscellaneous

• Top ten in scoring & Selke voting
• Durability (ie fewest games missed ten & fifteen years into their careers)
• Debunking a myth: “Howe earned a lot of phantom assists, while Richard didn’t”
• Penalty killing & the Norris trophy: Karlsson bucks the trend in a major way

 

[Canadiens1958]

Geocities Links

[*] A lot of these studies are several years old. I'm posting some of them for reference purposes. As of now (July 2012), I don't necessarily agree with the methodology or results in some/many of these posts.
[*] The data used in many studies are now at least a few years out of date. I'm not sure when (if ever) I'll get around to updating them.
[*] I shudder to think about how many hours these all took... at least it's been spread over 7.5 years so it's not too bad on a yearly basis...[/list]

Sadly the Geocities links do not work generating a response similar to the following:

http://www.geocities.com/thehockeyoutsider/Hart_shares.pdf

 

[Canadiens1958]

Counterpoint to Henri Richard Case Study

Comments/disclaimers


"Applied" case studies

• Gretzky vs Lemieux – why it helps to retire young
• 1960s Chicago Blackhawks -- a team that struggled due to a weak supporting cast while Hull, Pilote, Hall and Mikita took too much blame
• Marcel Dionne – detailed analysis -- a few bright spots, but generally as bad as you would expect
• Tony Esposito – it’s not his fault his teammates couldn’t score
• Bill Durnan -- why does he have such a poor reputation in the playoffs?
• Henri Richard – it looks like Harvey & the first-line scorers, rather than Richard, improved the most in the playoffs
• Joe Sakic – a look at his impact on Colorado's record

Counterpoint to the Henri Richard case study with link:

http://hfboards.mandatory.com/showthread.php?t=514771&page=10

see post #241

 

[Say Hey Kid]

seventieslord's "The Big Defensemen Study"

This is seventieslord's "The Big Defensemen Study"

http://hfboards.mandatory.com/showpost.php?p=47027609&postcount=583

http://hfboards.mandatory.com/showpost.php?p=47027793&postcount=584

http://hfboards.mandatory.com/showpost.php?p=47027837&postcount=585

http://hfboards.mandatory.com/showpost.php?p=47028099&postcount=586

 

[pnep]

Talent level by years/era

"HHOF Monitor" (1917-18/2005-06)

Players, Best and Worst "Total Regular Season W-L Dif."

 

[Bear of Bad News]

Some of the goalie metrics that I'd published here a few years back (this link is to the 2008-09, and that thread has links back further):

http://hfboards.mandatory.com/showthread.php?t=634696

This will all be on the goaltender site by the end of the summer (it's in the database, now I just need to get the database on the site - more tedious than it may sound).

 

[Hockey Outsider]

C1958: I think Geocities no longer exists. In late August I'll upload all of the PDF files to another site.

I'll take a look at your Henri Richard post then as well. I will say that I have a much more favourable opinion of him now than I did several years ago (when the argument in his favour was essentially "he won a lot of Stanley Cups"). His defensive abilities are much better documented now, and his offense (once his relatively low PP ice time is taken into account) is impressive for the era.

Yurog: correct me if I've misunderstood, but are you trying to predict the results of Vezina voting based on goalie stats? That sounds very interesting; look forward to seeing the results. I once tried to do the same with Norris trophy voting but had no luck (presumably because defense is not really captured by any mainstream statistic, but obviously influences Norris voting).

 

[Bear of Bad News]

I've moved a lot of the "whole threads" over to this forum, since they'll get more visibility here in a (currently) smaller forum than being buried a few years old in the other fora.

At some point, I may cull out single posts from larger threads and create new root threads here - if you have ideas for which posts deserve that treatment, please let me know.

 

[Sensfanman]

Here's an extension I've published based on a paper I published in Informs-Interfaces which I can't post here:

http://www.sloansportsconference.co...Novati_Split-personalities-of-NHL-players.pdf

 

[Mathletic]

Not sure if this thread is reserved for people who want to present their own research but I posted various links over the years. Thought I'd share in case people missed them.

Strategies for Pulling the Goalie in Hockey
David Beaudoin and Tim B. Swartz

http://hfboards.mandatory.com/showthread.php?t=677905&highlight=

http://people.stat.sfu.ca/~tim/papers/goalie.pdf




NHL Draft Order Based on Mathematical
Elimination

Adam M. Gold

http://hfboards.mandatory.com/showthread.php?t=736108&highlight=nhl+draft+order+based+mathematical

http://www.degruyter.com/view/j/jqas

Abstract

The NHL determines draft order from a lottery that favors teams that are lowest in the standings.
Losing can help a franchise acquire a coveted prospect, which encourages fans to cheer
against their favorite teams. Draft order based on mathematical elimination would force the teams
that performed poorest into a highly competitive atmosphere. The teams that are eliminated earliest
would instead have more games to earn the top picks. If substandard teams are to survive in
mediocre markets, the injustice of incentives for losing must be eradicated.

...

5 Conclusion
The National Hockey League (NHL) formula that considers the reverse stand-
ings to determine draft order triggers logical reasoning that can destroy emo-
tional attachments and fanaticism, without which hockey teams cannot thrive.
Losing can help a franchise acquire a higher draft pick, which encourages fans
to cheer against their favorite team. Franchises that endure poor seasonal
performance should not accept considerable rejection and departure from sup-
porters. Although the teams with the most losses receive the highest draft
picks, the promise of future success by losing in the present creates a false
sense of security. This current formula yields the distressing paradox where
success and failure become synonymous. The NHL should use my formula to
create competitive draft orders and inspire fans with passion and optimism.





Referee Analytics: An Analysis of Penalty Rates by National Hockey League Officials

http://hfboards.mandatory.com/showthread.php?t=1155785&highlight=

http://www.sloansportsconference.co...02/53-Schuckers_Brozowski_MIT_Sloan-Final.pdf




A Closer Look at the Relative Age Effect in the National Hockey League

http://hfboards.mandatory.com/showthread.php?t=830330&highlight=

http://www.degruyter.com/view/j/jqas

A Closer Look at the Relative Age Effect in the National Hockey League

Abstract

At young ages, a few extra months of development can make a big difference in size,
strength, and athletic ability. A child who turns 5 years old in January will be nearly 20% older by
the time a child born in December has their 5th birthday. In many sports, including hockey,
children born in the early months of the calendar year get noticed by their coaches because of the
superiority they demonstrate due to their age advantage. As a result, boys born early in the year
are more likely to reach the professional ranks of the National Hockey League (NHL). The
phenomenon just described has been labeled the relative age effect (RAE). Previous work studying
the RAE in the NHL has focused on individual NHL seasons, often encompassing many of the
same players across multiple seasons. We investigate the RAE using complete data on every
player who has ever played in the NHL. We focus the majority of our analysis on Canadian born
players and examine the RAE across hockey position and hall-of-fame status. For the first time,
we provide strong evidence of when the RAE began to manifest itself in Canada. Our change point
analysis indicates that the RAE began for players born since 1951. Finally, we make a case for
what initiated this change in the way young hockey players develop, particularly in Canada, which
produced over 90% of NHL players at that time.

 

[Black Gold Hockey]

I recently looked at the possibility of using a player's change in performance in his CHL playoffs to predict his change in performance come NHL playoff time.

http://blackgoldhockey.com/does-junior-playoff-performance-translate-to-nhl-playoff-performance/

 

[Cunneen]

I recently looked at the possibility of using a player's change in performance in his CHL playoffs to predict his change in performance come NHL playoff time.

http://blackgoldhockey.com/does-junior-playoff-performance-translate-to-nhl-playoff-performance/

http://nhlnumbers.com/2012/10/3/clutch-nhl-playoff-big-game-performers

Take a read at this article, it pertains to playoff "clutchness"

 

[Black Gold Hockey]

http://nhlnumbers.com/2012/10/3/clutch-nhl-playoff-big-game-performers

Take a read at this article, it pertains to playoff "clutchness"

Very interesting. He always does such great work.

 

[Cunneen]

Very interesting. He always does such great work.

I'd say that right now he is the top mind in hockey analytics. Dude is a genius. I'm actually doing some work with him right now which is awesome. Hopefully I can eventually work with him on some of his projects.

 

[Jyrki]

http://nhlnumbers.com/2012/10/3/clutch-nhl-playoff-big-game-performers

Take a read at this article, it pertains to playoff "clutchness"

Cool stuff. I recall doing something along these lines to measure whether if Canucks players were "chokers" by different measures, I might consider unearthing the spreadsheet... if I can find it.

 

[Sturminator]

Here's a link to the VsX study we worked up mostly in the ATD section.

http://hfboards.mandatory.com/showthread.php?t=1361409

 

[Wesleyy]

Accounting for possession in hits and blocked shots
Accounting for shot quality (high percentage vs low percentage shots) in corsi
Analysis of different factors affecting home ice advantage

 

[Black Gold Hockey]

Measuring Teams’ Draft Success Using Analytics:
http://blackgoldhockey.com/measuring-teams-draft-success-using-analytics/

And a part 2, breaking down the draft, position by position:
http://blackgoldhockey.com/measuring-teams-draft-success-using-analytics-pt-2/

 

[Hippasus]

Greatest goal scorers relative to the average first-liner in the season in which they scored a lot of goals. By "a lot", I mean, primarily, one of the top three placings for a given regular season.

http://hfboards.mandatory.com/showpost.php?p=118059189&postcount=357

 

[ImporterExporter]

We're we planning on seeing the above links fixed? They all seem to be broken due to the transfer last year.

 

[oilerbear]

Accounting for possession in hits and blocked shots
Accounting for shot quality (high percentage vs low percentage shots) in corsi
Analysis of different factors affecting home ice advantage

You do know I have created 20+ base theories presented on Lowetide.ca and HF boards.

Theory #2 Expected ga was defined in 06-07 thru my High danger shot, Corsi theory. It is 1/2 of defining elite def dmen and defining true goalie performance. Alan Ryder 2006 shot quality came up with the same numbers.

I had a conversation about my theories relative to shot density with a Mike Vallely former goalie coach of Dallas stars ( now of elite goalies) on the bridge in Disneyworlds Animal kingdom raft ride
It is and has always been a density based system.
Where the high danger area is defined by a ratio between HD and led shot success.
Avg HD .825 save%
Avg LD .965 save%
Hd 17.5%:LD3.5%
A 5 to 1 ratio.

Theory #3 From this you can define dmen performance relative to CA.
CA is a product of the failing of teammates offence.
Cumulative shot success/ Ca the best dmen have the lowest ratio.

Theory # 1 Performance mean and ranges based on comp, teammates and ZS.
Days after behind the net posted his database.
I knew I could generate a 3 d matrix of groups based on forwards,
Upper, lower 1st. 2nd, 3rd, 4th line teammates and competition
As well as ZS modified by excluding the biggest factor affecting
Forwards CF and CA. The coach decision to line change a player with or without Pocession.
This must be excluded cause this affect is not dependent on the player.

Theory #4 table hockey goalie.
For the longest time goalies were stand up skater goalies.
The some moved to butterfly.
When you play table hockey you realize the goalie only stops a shot by moving with the puck.
The first goalie to have a repetitive jump in sVe% was John Van biesbrook. His former teammate even called him a table hockey goalie last year.
So modern goalies move in unison with the puck.

Theory #5 Avg game shot density curve.
Allows a game by game performance standard for a goalie based on density.
An avg shot game 30/Gm
With avg HD 10.5 and Ld 19.5 shot total defines the shot density
5 to 1 success
(10.5 x 5) + ( 19.5 x 1) = 72
Then divide this number by .125ga intervals to generate a curve.
You measure a goalies performance relative to densityfaced/ga and compared to curve.
Last year Elliot faced a 142 density game were he gave up one weak goal and 4 Ga.
142/4 = 35.5
When compared to the average curve.
It interpolates as a 2.03 GA performance.

Theory #6
Dpairs establish the mean expected save % a goalie performs around.
Worst HD pairs:
14.5 HD shots; .825 save%
15.5 ld shots ; .965 save%
.897 save% mean

Avg d pair
10.5 HD shots; 19.5 ld shots
.916 save% mean

Best D pair
7 HD shots
23 ld shots
.935 save% mean

Theory #7 belichek and me.
You want players with the highest baseline of performance.
They are players teammates can trust and know we’re they will be and what they can do.

Theory #8 False eye affect.
Peo0le let one pay define a players skill.
Rather than looking at a per shift performance rate.
Poster child for this was Souray.
When he got beat it was dramatic.
Many fools let that one play to define him.
Yet he had one of the 5 lowest mistakes leading to goal rates for dmen in the league.
Bill Belichek and I would take him.

Theory #9 We’re is Waldo.
Offensive dmen abandon the free path to their own HD
Area when they cheat for offence.
Yielding high Expected ga rates and low DPair save% means.
When a lot of goals are scored that dmen is not seen in the screen.

Theory#10 offensive dmen are OFFENSIVE!
95% of dmen generate even offence at a 4th line to #17 fwd pace.
Clearly not justifying abandoning their HD area.
You want elite transition passing from your d at even play.

Their is a secondary affect in defensive play.
The first was density. Then the second is a binary quality definition.
Which had from day one allowed me to define Kris Russel as the best binary shot Quality Dmen in the game.

I have many more theories being used by teams. Who I will not work for cause their pay is about 1/4 of mine and the life cycle is 1-5 years.

Ia, currently developing a server to populate my theories and work based on empirical ( real analysis) rather than theoretical and regression based analysis which is situationally flawed.

 

[ted2019]

Does anyone here have some of the broken links saved somewhere? I'm looking for Goaltending advanced studies.