Random Thoughts - The End is Nigh and New Stats Genesis

Kent Wilson
April 24 2013 09:05AM

 

 

Sorry about the lack of post game article last night folks. The truth is, with the end of the season looming and the games mattering less and less, it's becoming harder and harder to muster the fortitude it takes to write about them. For those interested though, here's an overview:

Despite a lackluster line-up, the Flames kept pace for the first half of the game and were even leading 2-1 thanks to a beauty snipe from Brodie and break-away marker from Cervenka. The Predators took over the game after that, though and ended up outchancing the Flames at almost a 2-1 margin overall. The Flames nevertheless made it close via Cervenka's 2nd goal (a tipped point shot) and a last minute flurry with Kipper pulled.

Sven Baertschi had another two assists, moving his point streak to six games.

- Also, the scoring chances:

 (Flames in Blue)

 

# Player EV     PP     SH    
5 GIORDANO, MARK 18:09 1 4 01:10 1 0 02:16 0 0
7 BRODIE, TJ 17:45 8 5 00:50 0 1 01:02 0 0
10 CERVENKA, ROMAN 11:59 3 7 00:49 1 0 00:00 0 0
16 MCGRATTAN, BRIAN 09:51 1 2 00:00 0 0 00:00 0 0
19 JONES, BLAIR 17:05 5 1 00:00 0 0 02:23 0 2
24 HUDLER, JIRI 15:15 2 6 00:58 0 1 00:00 0 0
26 WIDEMAN, DENNIS 18:59 1 5 01:10 1 0 03:53 0 3
29 ALIU, AKIM 12:36 2 8 01:11 0 1 00:00 0 0
32 BYRON, PAUL 12:52 2 2 00:00 0 0 00:41 0 0
33 BABCHUK, ANTON 16:31 3 9 00:00 0 0 00:00 0 0
34 KIPRUSOFF, MIIKKA   12 18   1 1   0 3
38 STREET, BEN 14:17 6 3 00:13 0 0 01:31 0 0
42 CUNDARI, MARK 17:29 9 6 00:50 0 1 00:41 0 0
44 BUTLER, CHRIS 17:06 2 7 00:00 0 0 02:14 0 3
46 BANCKS, CARTER 11:31 1 2 00:00 0 0 01:02 0 0
47 BAERTSCHI, SVEN 13:24 3 7 00:49 1 0 01:29 0 1
51 HORAK, ROMAN 14:16 3 6 00:58 0 1 01:25 0 2
58 HANOWSKI, BEN 12:51 5 3 00:13 0 0 00:00 0 0
59 REINHART, MAXWELL 12:00 2 7 00:49 1 0 01:35 0 1

- TJ Brodie +8, -5 despite playing all night. Not sure how he does it. Also Ben Street , Blair Jones and Ben Hanowski's numbers are surprising. Akim Aliu, on the other hand, took two penalties and was outchanced +2 to -8 in just 12 minutes of ice time. Yeesh.

- Last night was a good one for scoreboard watchers. New Jersey, Buffalo and Philadelphia won, pushing them four and three points ahead of Calgary respectively, while Nasvhille and Carolina both won, putting them just one point back of the Flames.

Colorado, on the other hand, lost to St. Louis which was a kind of double-edged sword for Calgary - the two points guatantees the Blues a playoff spot and therefore another first rounder for the Flames in June. On the other hand, that means the Avs can't possibly catch the Flames even if they win out. As a result, the lowest Calgary can drop in the standings is now 28th.

The range of outcomes is rapidly starting to narrow at both ends. If PHX, NJD, or BUF get just one more point in their remaining contests, the Flames can't possibly catch any of them. Likely, the real "battle" for draft positioning is between Calgary (42 points), Nashville (41), Edmonton (41) and Carolina (41) with Tampa Bay (38) and Philadelphia (45) at the margins, which equates to a probable range of 27th-24th overall (draft picks 4-7). Edmonton has the most games left of those teams listed (3) so are technically in a better position to finish higher, but let's not underestimate their ability to fail their way to yet another high draft pick.

- On a more general note, Jesse Spector of the Sporting News had an interesting article on Jim Corsi yesterday, the Buffalo goalie coach who created the "corsi" stat. The article also discusses a bit how corsi came to be used by people like myself as a measure of skater performance.

- Related, Cam Charron gave a detailed explanation of the "PDO" stat back in January, which I think it perhaps the most important (and least intuitive) "advanced stat" in hockey today. If you're a gambler or a fantasy hockey player, PDO is all you'd really need to follow and understand to make money.

- Double related: thanks to the talk about corsi and PDO, some folks dug through some old blog archives where these things initially came to be and were discussed. For example, here's where corsi made its appearance. The author in question, Vic Ferrari, is kind of the Bill James of the advanced stats movement in hockey. Some of you will likely recognize some names in the comments, including my old online pseudonym. The first comment is also by "PDO", the guy who initially conceived of the "luck stat", which was later named after him as a result.

- Many people ask why most of hockey's new stats have such odd names. The nomenclature kind of arose out of on-going discussions between a number of folks in isolated corners on the internet more or less out of convenience. No one was thinking about the future "marketability" or accessiblity of the names at the time, because, frankly, none of us thought it would be anything more than just that - an interesting discussion.

At one point I may be tempted to go back and chart the history and growth of adavanced analysis in the blog world, as well as how I came to be involved and my own minor contributions.

- Tyler Dellow, one of the significant early pioneers and adopters of advanced stats, had a note on why the de-centralized, bottom-up model of stats discovery and experimentation is successful versus isolated "expert" models here:

I think it’s fair to say that many academics take a dim view of what goes on online and from non-academic experts. As a result, I think a lot of them aren’t involved in the internet discussion about these topics and don’t even follow it. They just do their thing their academic way.

His reference is a THoR player evaluation model presented at the Sloan Analytics conference that was hopelessly confounded by issues that have been well known in online stats circles for years: playing-to-score effects and rink biases. The paper was presented by a company that provides consulting services for NHL teams.

- Skipping back to the Blues for a second: the win means Jay Bouwmeester will, for the first time in his long career, make the post-season. I imagine he was tempted to skate off the bench and celebrate like he'd already won the cup when the buzzer went last night (assuming Jay Bouwmeester actually feels anything at all).

- A little teaser for something we'll be announcing soon:

The game we've come up with is a fun one and we have some big, big media names who are going to take part. And the prizes are not only awesome, but open to be won by anyone in the world.

Stay tuned...

- Finally, on a completely personal note, I was recently hired by Mosaic Studios here in Calgary to help them with new business development and digital strategizing (that's a word, right?). Mosaic is an established, full service marketing/communications shop that can do everything from websites and mail-outs to branding, campaigns, event management, etc. 

So if any of you out there are looking for marketing services for your business or company and you'd like to work with me outside of the narrow world of hockey fandom, drop me a line sometime via email (kent.wilson@gmail.com) or Linkedin.

Of course, this doesn't mean I'm leaving FN or the Nations. It just means I really hate having free time, apparently.

Around the Nation Recently

39d8109299a9795cb3b41a4e9b49d501
Former Nations Overlord. Current FN contributor and curmudgeon For questions, complaints, criticisms, etc contact Kent @ kent.wilson@gmail. Follow him on Twitter here.
Avatar
#1 SeanCharles
April 24 2013, 09:21AM
Trash it!
0
trashes
Props
0
props

Hoping for a STL vs LA first round matchup. STL turfed in 5.

Avatar
#2 SmellOfVictory
April 24 2013, 09:21AM
Trash it!
0
trashes
Props
0
props

I always assumed you had no concept of "free time" anyway. Good to know that your actions are in keeping with my expectations.

Avatar
#3 mattyc
April 24 2013, 09:36AM
Trash it!
0
trashes
Props
0
props

I won't go on a big rant about Tyler Dellow's article here, but 2 quick things:

1. Not all science/scientists/academia are equal

2. Anyone who's done any academia knows that you spend more time reading the literature related to your topic than actually doing your own research.

Avatar
#5 mattyc
April 24 2013, 10:03AM
Trash it!
0
trashes
Props
0
props

@Kent Wilson

Agreed. I just object to the blanket statement "Academics". Its a two-way street - I'm sure there's lots of 'hockey stats' that amateurs have come up with through flawed methodology as well. I'm also not really sure Shuckers' paper counts as academia since it wasn't peer-reviewed.

If his argument is "just because you have a PhD doesn't mean you know more than everybody else" he's right and I agree. I just don't see it as an amateur vs. academic thing.

Avatar
#7 kittensandcookies
April 24 2013, 10:36AM
Trash it!
0
trashes
Props
0
props

PDO Bear is not impressed.

Avatar
#8 ChinookArch
April 24 2013, 10:36AM
Trash it!
0
trashes
Props
0
props

I learned about PDO a few months ago and I've been tracking Team PDO in Cam Charron's (@ NUMBERS) articles. The water cooler guys at work now think I'm a hockey genious. Most of them expected the Rangers to be a cup contender at the beginning of the season, and I told them the Rangers couldn't hold up the second their goaltending came back from the stratosphere. PDO is the ultimate stat that predicts regression for players and teams.

Avatar
#9 ChinookArch
April 24 2013, 10:44AM
Trash it!
0
trashes
Props
0
props

On a personal note, thanks for overview of last night's game. I'm traveling where there is no NHL coverage and I can't stream the games, on the WIFI I have access to right now. FN is a far more reliable, realistic and trusted source of the Flames to me, than any of the Calgary news sites and papers.

BTW congratulations on the new job!

Avatar
#11 RossCreekNation
April 24 2013, 10:49AM
Trash it!
0
trashes
Props
0
props

What was your old name, Kent?

Avatar
#12 ChinookArch
April 24 2013, 10:51AM
Trash it!
0
trashes
Props
0
props

When I was looking to learn about PDO, I found an old article by Vic Ferrarri that discussed the Oliers 2006 players, and how they all cashed in with big contracts (I.e. Shaun Horcoff $5M/yr contract) after their unlikely appearance in the cup final. Every GM in the league should take 5 minutes to learn about PDO, and drive agents nuts.

EDIT: just noticed The PDO link in Kent's post mentioned the article. I guess that's where I found it in the first place.

Avatar
#14 RossCreekNation
April 24 2013, 11:05AM
Trash it!
0
trashes
Props
0
props

@Kent Wilson

Ha, I was looking for FiveholeFanatic or whatever it was you used to go by. There's a goline in some of those old blog posts. I think that's what most anti-stats guys don't get - this conversation has been ongoing for the better part of the past decade, not just the last couple yrs. Interesting stuff.

Avatar
#15 mattyc
April 24 2013, 11:55AM
Trash it!
0
trashes
Props
0
props

@Kent Wilson

haha really... interesting. Kind of a funny discipline, especially with the fact that every team probably has some proprietary stuff as well that they want to hide from everyone else.

Avatar
#16 beloch
April 24 2013, 12:03PM
Trash it!
0
trashes
Props
0
props
Kent Wilson wrote:

Right, but his point is there is a divide between publish academic material and the amateur work that is outside of that sphere. And that the former would gain by engaging by the latter, but they don't tend to.

What are academics actually doing in this field. i.e. Where should I look for papers on predicting professional sport outcomes? I can't imagine it's a field that gets a lot of funding, but hey, English departments still exist so anything is possible.

I'd imagine that anyone actually doing paid research in this field is probably digging into really complicated, obscure and esoteric aspects of it that wouldn't be very interesting (or useful) to most amateur's. Let's face it, they're probably not going to get many publications by writing about Corsi and PDO. Some fields of academia unfortunately tend to reward complexity and incomprehensibility (Related: http://snarxiv.org/).

Avatar
#17 clyde
April 24 2013, 12:09PM
Trash it!
0
trashes
Props
0
props
ChinookArch wrote:

I learned about PDO a few months ago and I've been tracking Team PDO in Cam Charron's (@ NUMBERS) articles. The water cooler guys at work now think I'm a hockey genious. Most of them expected the Rangers to be a cup contender at the beginning of the season, and I told them the Rangers couldn't hold up the second their goaltending came back from the stratosphere. PDO is the ultimate stat that predicts regression for players and teams.

Latest PDO rankings include: 1- Toronto 3- Columbus 7-Dallas 8- Tampa 10 - NYR 11-Buf 14-Phoenix 15- Det Not a lot of playoff teams in that group but: 21- Ottawa 22-Stl. 23- LA 24 - NYI

Avatar
#18 Monaertchi Gaudnett
April 24 2013, 12:17PM
Trash it!
0
trashes
Props
0
props

Where does one fine team PDO? I can't seem to find it at behind the net.

Edit: Found them! NHLNumbers.com (see link at top of page!)

Avatar
#19 Monaertchi Gaudnett
April 24 2013, 12:27PM
Trash it!
0
trashes
Props
0
props

Re: PDO

So EDM has a PDO of 0.998. That, combined with a really craptastic record, means "They are who we thought they were"? They are a crappy team and are exactly where they should be in the standings?

CGY PDO is 0.974 (29th) so their standing in the league is lower than it should be and they are bound to be better next season? They probably won't really be the same team next season, so any conclusions made would be false, but assuming the team stays the same.

Avatar
#21 vowswithin
April 24 2013, 01:12PM
Trash it!
0
trashes
Props
0
props
SeanCharles wrote:

Hoping for a STL vs LA first round matchup. STL turfed in 5.

This

Avatar
#22 Baalzamon
April 24 2013, 01:37PM
Trash it!
0
trashes
Props
0
props

Why are we all assuming Pittsburgh will go deep this year? Has no one noticed that Crosby and Malkin are both currently injured?

I want Iginla to win the cup as much as the next bloke, but it doesn't have to be this year, does it? Might have more luck next year when, you know, their top 2 centers aren't named Brandon Sutter and Jussi Jokinen.

Avatar
#23 jeremywilhelm
April 24 2013, 01:38PM
Trash it!
0
trashes
Props
0
props

Very positive, even though some thought Cundari was struggling last night, he ended up +9/-6 in the same amount of EV str minutes as TJ.

Avatar
#25 jeremywilhelm
April 24 2013, 01:40PM
Trash it!
0
trashes
Props
0
props

Surprisingly also, Blair Jones played the most EV minutes of any forward and outchanced the oppo +6/-1. Solid.

Avatar
#26 jeremywilhelm
April 24 2013, 01:41PM
Trash it!
0
trashes
Props
0
props

@Kent Wilson

Maybe coat tails, but he did end up with 1 more EV SC than Brodie, so he must have been driving possession at least slightly.

Avatar
#27 SeanCharles
April 24 2013, 01:42PM
Trash it!
0
trashes
Props
0
props
vowswithin wrote:

This

They went 0-3 vs Kings this year. 1-9-1 last two years.

All other potential matchups are close or in STL's favour.

Not saying these things matter much come playoff time. But its hard to deny that LA is probably the one team in the West, that they could meet in first round, they don't match up too well against.

Avatar
#28 Clyde
April 24 2013, 01:44PM
Trash it!
0
trashes
Props
0
props
beloch wrote:

What are academics actually doing in this field. i.e. Where should I look for papers on predicting professional sport outcomes? I can't imagine it's a field that gets a lot of funding, but hey, English departments still exist so anything is possible.

I'd imagine that anyone actually doing paid research in this field is probably digging into really complicated, obscure and esoteric aspects of it that wouldn't be very interesting (or useful) to most amateur's. Let's face it, they're probably not going to get many publications by writing about Corsi and PDO. Some fields of academia unfortunately tend to reward complexity and incomprehensibility (Related: http://snarxiv.org/).

One of my first courses in coaching masters was a statistics course. It is a very in depth area.

Avatar
#29 jeremywilhelm
April 24 2013, 01:46PM
Trash it!
0
trashes
Props
0
props

Nice to see that Sven's PDO is basically even on this scoring streak, so his point production isn't even being buoyed by some crazy on ice S%.

He might actually be really really good.

Avatar
#30 SeanCharles
April 24 2013, 01:54PM
Trash it!
0
trashes
Props
0
props

Question to anyone who would like to answer:

What prospects, outside of the ones we have yet to draft, will make the team next year?

My guess: Sven-top6, Cundari/Breen-top7/8 D, Horak-top9, Bouma-4th line.

I'm hoping for: Aliu-4th line, Reinhart after spending majority of season in AHL fine-tuning his game-top9.

The rest I dont think have much of a chance to make the team, as they all need to work on aspects of their game.

Avatar
#32 everton fc
April 24 2013, 02:40PM
Trash it!
0
trashes
Props
0
props

I bet they sign Cervenka.

Could Aliu actually put pucks in the net off the 4th line? I don't think he's much of a pugilist...

Begin's earned a contract. Good to have in the room.

Avatar
#34 jeremywilhelm
April 24 2013, 03:33PM
Trash it!
0
trashes
Props
0
props

@everton fc

Reinhart is getting 2.6 shots per game on the 4th line (small sample size, but thats pretty darn good). He is probably gonna be a 4th liner next year. If the Flames structure that 4th line properly, it might help them quite a bit.

Avatar
#35 jeremywilhelm
April 24 2013, 03:34PM
Trash it!
0
trashes
Props
0
props

@everton fc

I would sign Cerv to a 2 year 850K deal. His shooting percentage is very high right now, so that is something to be very wary about. I guess it would depend on what type of commitment he was willing to give this offseason to his first stride and fitness level.

Avatar
#36 PrairieStew
April 24 2013, 10:37PM
Trash it!
0
trashes
Props
0
props

@jeremywilhelm

I think the important thing to consider is that the guy has really only had 40 games to adjust to small rink and NHL game. Considering his ice time I'd have to say he's been pretty good - I think he is 3rd on the team in even strength goals

Avatar
#37 Section205
April 24 2013, 10:43PM
Trash it!
0
trashes
Props
0
props

Kent,

Thanks for all the valuable insight over the season. So much more quality than Calgary Sun or Herald... yeah I guess I like to read those too, but I always feel dumb afterwards.

Really appreciate the links and background to the people and ideas behind some of the advanced stat metrics. Very good reads.

I eventually found myself reading an Ellen Etchingham post on Regression. That is one talented writer, describing Halak's "Mona Lisa" moment of peak performance. Her conclusion was worthy of a standing ovation!

She is now instantly a new favorite and I will read more of her work.

http://blogs.thescore.com/nhl/2012/03/16/messier%E2%80%99s-skates-regression-luck-and-transience/

Avatar
#38 exsanguinator
April 24 2013, 11:23PM
Trash it!
0
trashes
Props
0
props

I saw an article on NHL.com that mentioned Corsi as a stat that measured possession and I was oddly overjoyed since I'm not really that much of an advanced stats guy.

Comments are closed for this article.