HFBoards

Go Back   HFBoards > NHL Western Conference > Pacific Division > Edmonton Oilers
Notices

The positively spooky even strength SV% stat ...

Closed Thread
 
Thread Tools
Old
04-14-2004, 11:53 AM
  #1
igor*
 
Join Date: Mar 2002
Posts: 3,276
vCash: 500
The positively spooky even strength SV% stat ...

The NHL does log some really relevant information. Then they usually they keep it to themselves. But they do give us ES-save% for the last four years. 3 of those years in easily downloadable format.

And I'm bored with work, so I thought I'd skive for a bit and start a conversation about this stuff ... I think quite a few people here are into this kind of thing.

Anyhoo, the ES-save% stat is gold, very clean. Its all probabilities of course, and a sample of a 1000 or so EV shots against per year ... that leaves a bit of room for luck to play its cards, but in a predictable pattern. There really is an inevitability to it all IMHO. The annual change in EVsave% for the average NHL starting goalie is highly predictable ... and the distribution is near exactly what I calculated it should be (using very simple math based on pretty fundamental ideas and observations).

Some things that are surprises, or at least were to me:

* The average ES-save% in the league is now .923. This is climbing, and mostly because the younger goalies are better, and they are gradually getting more ice-time.

* A lot of the goalies whose perceived value has dropped a bit ... Khabibulin, Biron, Lalime, Joseph etc. These guys haven't gotten any worse, they are still good goalies, in fact they've had more consistent ES-save% numbers that should have been expected. Its just that the world has moved beneath their feet ... the competition is now better ... and while they are not any worse at all, they are relatively worse.

* I didn't watch many Kings games. But there was plenty of dissing of Cechmanek on here. And he finished the season with an impressive .929 save%. (Career average is .930). Possibly that was the reason that the Palffy-less Kings stuck around the playoff race for so long.

* Giguere finished the year with a .921 EVsave% ... close to last year's .923, but not as good as the .929 the year before that. (We really should have seen this guy coming).

* Salo finished the year with an EVsave% of .913. His average for the previous 3 years was .912.
- The chances of him finishing between .908 and .916 was a bit over 50%.
- The odds of him finishing above .923 for this year, and for next, were/are about 32:1. (3%)
I could narrow down these numbers even further with more data from previous years (except for 99-00, the year of the skate-in-the-crease-rule).

* A pet peeve --> media often use the word 'inconsistent' to describe goalies, and usually wrongly. Unless it is PoliteSpeak for 'crap' ... ir doesn't make sense. Fernandez is in fact the most inconsistent NHL goalie over the past few years ... and he's pretty damn good.

* The Jussi/Ty duo looks like a good business move. Neither has played enough games to really give a strong idea of how good they will be (either by eye or by the numbers). And I think most of us agree that it could go really wrong or really right for either guy ... but by the numbers; there is over a 75% chance that they will give the Oilers .920 Evsave% goaltending or better. And roughly a 40% chance of .930 or better. Fingers crossed.

At the beginning of every year oddsmakers run save% wagers for many NHL goalies. And at Skinny margins for a futures wager (15%). Obviously they post odds only for goalies whose save% projects to be far different from common perception. And they'll always put in the goalies that are moving teams.

If it was an ES-save% wager ... it would be like catching fish in a barrel. And it might still be, but the PP-save% number is a little tough to figure, it is a bit messy. The average change in PPsave%, for an NHL starting goalie, from year to year should be .0246 but it isn't quite ... and the distribution isn't quite right either ... I think the 5on3 PPs on bugger it up a bit. Possibly the play of the skating PKers too.

Thoughts?

EDIT: In case anyone figures out the logic/math here (and I'm pretty sure oilswell will, if he hasn't already). Please keep it to yourself or PM me ... this is NOT a math forum after all, we don't want to bore others to tears with it.


Last edited by igor*: 04-14-2004 at 12:02 PM.
igor* is offline  
Old
04-14-2004, 01:37 PM
  #2
jadeddog
Registered User
 
jadeddog's Avatar
 
Join Date: Sep 2003
Location: Regina, Saskatchewan
Posts: 11,810
vCash: 500
actually igor, i would like to see the math behind this.... im a math minor in university so im obviously interested in stats and math, so if you could PM the basic premise behind it all i would love to go over it, thx

jadeddog is offline  
Old
04-14-2004, 01:51 PM
  #3
barto
Registered User
 
Join Date: May 2002
Location: Edmonton
Posts: 2,465
vCash: 500
Igor, I don't think anyone would mind AT ALL if you were to post some of the background on here. After all, those who aren't interested just won't read it, while there might be a surprising number of us who WILL read it and it might spark some interesting discussions. Hell, I might read it all just to keep my mind off of what the Oilers COULD have been doing in these playoffs...!

Bart

barto is offline  
Old
04-14-2004, 01:52 PM
  #4
H-Bear
Registered User
 
Join Date: Nov 2002
Location: North Bay ON
Country: Canada
Posts: 1,503
vCash: 500
Excellent post igor. You clearly have too much time on your hands . A few comments:
1.
Quote:
* Salo finished the year with an EVsave% of .913. His average for the previous 3 years was .912.
- The chances of him finishing between .908 and .916 was a bit over 50%.
- The odds of him finishing above .923 for this year, and for next, were/are about 32:1. (3%)
I could narrow down these numbers even further with more data from previous years (except for 99-00, the year of the skate-in-the-crease-rule).
This stinks of z-curves

2. Conklin, Salo, and to a certain extent, Markkanen played behind one of the best even strength teams in the NHL. Could I not jump to a reasonable conclusion that there stats are somewhat skewed in their favour by having this advantage.

3. Stats analysis, while interesting, always lacks one dimension; and that is quality. Quite frankly, who cares if Goalie X can save 49 of 50 shots? I mean, if he is facing 15 quality scoring chances a night, the hats off to him. However, if he is only making a handful of difficult saves, then this save percentage stat is rather "blah" to me.

H-Bear is offline  
Old
04-14-2004, 01:55 PM
  #5
rabi_sultan
Registered User
 
Join Date: Sep 2002
Location: London, England
Country: United Kingdom
Posts: 3,782
vCash: 500
Send a message via ICQ to rabi_sultan Send a message via AIM to rabi_sultan Send a message via MSN to rabi_sultan Send a message via Yahoo to rabi_sultan
igor you should write a book, i kid you not, "The numbers of hockey"

rabi_sultan is offline  
Old
04-14-2004, 02:29 PM
  #6
igor*
 
Join Date: Mar 2002
Posts: 3,276
vCash: 500
Quote:
Originally Posted by barto
Igor, I don't think anyone would mind AT ALL if you were to post some of the background on here. After all, those who aren't interested just won't read it, while there might be a surprising number of us who WILL read it and it might spark some interesting discussions. Hell, I might read it all just to keep my mind off of what the Oilers COULD have been doing in these playoffs...!

Bart
No real motivation for me, except to check to be sure I didn't make a mistake ... in which case I'd PM it to oilswell. But I don't think I did, so I won't.

Its really basic stuff though, and anyone who has watched gametapes and tallied chances I'm sure thinks in the right lines ... but probably doesn't have the mathematical bent. And vice versa.

This thing was inspired by the comments of a statistician consultant a few months ago (a world renowned, and extortionately expensive, American chap) who made some comments to a friend of mine. Came up in shop talk, and this guy made a baseball analogy to explain why some data [hugely valuable data if it was valid] was not likely relevant. (And he can't imagine how wasted a baseball analogy was on my friend :lol ). The baseball analogy struck me as wrong and/or meaningless ... but I know this guy is smarter than me ... so it stuck in my head.

I've googled the guy, and he's never worked in anything as frivolous as baseball, and I doubt many teams would pay his rate anyways.

But its interesting stuff this, probability and statistics. I mean the formulas are all there in any text or online ... only takes a minute to run them on a any of a bunch of different kinds of software. Methinks its more a matter of knowing what fits where than it is being good at math. Being able to see the forest for the trees, in essence "what's the best analogy?" I've often said that a statician on his own is pretty much useless ... but a statistician working as Roger Nielsen's assistant ... useful.

I originally wrote this post completely as an non-hockey analogy with an offhand BTW: The EVsave% numbers for the last three years. . Draw your own conclusions bit at the end. And that would have been funnier

But I'll wait to see the lines for next years SV%s and see how much the boys are hedging. (If there is a next year with the CBA ending).

igor* is offline  
Old
04-14-2004, 04:11 PM
  #7
igor*
 
Join Date: Mar 2002
Posts: 3,276
vCash: 500
Quote:
Originally Posted by H-Bear
... This stinks of z-curves
I don't even remember what z-curves are

Quote:
... Conklin, Salo, and to a certain extent, Markkanen played behind one of the best even strength teams in the NHL. Could I not jump to a reasonable conclusion that there stats are somewhat skewed in their favour by having this advantage.
This strikes me as a bizarre argument. Having said that ... I think you are in the majority here. Which baffles me. [/quote]
BTW: Last year Markannen (vast majority with the Rangers) EVsave% .937
His cohort Dunham EVsave% .905

And ...
Conklin EVsave% .930

Quote:
Quite frankly, who cares if Goalie X can save 49 of 50 shots? I mean, if he is facing 15 quality scoring chances a night, the hats off to him. However, if he is only making a handful of difficult saves, then this save percentage stat is rather "blah" to me.
I really like you as a poster H-Bear. But you would be wise to heed this advice ... NEVER, EVER wager. Seriously.

igor* is offline  
Old
04-14-2004, 05:19 PM
  #8
H-Bear
Registered User
 
Join Date: Nov 2002
Location: North Bay ON
Country: Canada
Posts: 1,503
vCash: 500
NOTE: Sweet bajeebers; I made a lot of typos in that first reply ; and some of them my own pet peeves!! *Smacks self*
Quote:
Originally Posted by igor
I don't even remember what z-curves are
Stupid statistical tool used to determine (or guess) at how many subjects will fall into a certain range.
Quote:
This strikes me as a bizarre argument. Having said that ... I think you are in the majority here. Which baffles me.
BTW: Last year Markannen (vast majority with the Rangers) EVsave% .937
His cohort Dunham EVsave% .905

And ...
Conklin EVsave% .930
Wasn't really an argument; more like wondering out loud ... and I don't always make sense, so if it was a bass aakwards (not a typo ) argument to you, then that's fine.
Quote:
I really like you as a poster H-Bear. But you would be wise to heed this advice ... NEVER, EVER wager. Seriously.
Not too sure what you mean here, but on that note, I am indeed NOT a betting man. I occasionally buy scratch tickets and play poker, but that's as far as I go.

H-Bear is offline  
Old
04-14-2004, 06:38 PM
  #9
Lowetide
Registered User
 
Lowetide's Avatar
 
Join Date: Feb 2002
Country: Canada
Posts: 13,281
vCash: 500
I wish they'd had these stats when Gilles Meloche was saving 50 a night for the California Golden Seals.

One of the things that it can't reflect is the quality of the shots. igor, what is your opinion on that? Would they even out (iyo)over a period of time, making that argument a non-starter? Or would certain goalies on poorer teams tend to have poorer SP 5x5? And can you measure it? (Something by 9pm, thanks).

Lowetide is offline  
Old
04-14-2004, 07:49 PM
  #10
H-Bear
Registered User
 
Join Date: Nov 2002
Location: North Bay ON
Country: Canada
Posts: 1,503
vCash: 500
Quote:
Originally Posted by lowetide
I wish they'd had these stats when Gilles Meloche was saving 50 a night for the California Golden Seals.

One of the things that it can't reflect is the quality of the shots. igor, what is your opinion on that? Would they even out (iyo)over a period of time, making that argument a non-starter? Or would certain goalies on poorer teams tend to have poorer SP 5x5? And can you measure it? (Something by 9pm, thanks).
That indeed would be interesting; perhaps a co-relation between SV% and scoring chances (not a official stat I don't think).

H-Bear is offline  
Old
04-14-2004, 08:11 PM
  #11
mudcrutch79
Registered User
 
Join Date: Jul 2003
Location: The Big Smoke
Posts: 3,903
vCash: 500
I can't express enough how much I hate the NHL, and their serious lack of data that they have online. As I look at this stuff, it's difficult to say that there is enough of it there to draw any serious conclusions. Personally, I'd really like to see data for the goalies who've had these runs, and then completely fallen apart. I look at a guy like Jim Carey, couple of good seasons, and then his game fell apart. Why can't the NHL make this stuff available?

mudcrutch79 is offline  
Old
04-15-2004, 07:03 AM
  #12
igor*
 
Join Date: Mar 2002
Posts: 3,276
vCash: 500
Quote:
Originally Posted by mudcrutch79
I can't express enough how much I hate the NHL, and their serious lack of data that they have online. ...
The NHL is funny that way. They spend millions of dollars gathering all this data ... then guard most of it like it was the secret to cold fusion.

When they do publish the good stuff its usually in a format that it would take ages to scrape all of the data off of (like the Stats Machine) ... and/or impossible (for me anyways) to write a macro to scrape it all off automatically ... and/or the accuracy is intentionally compromised (Shift Charts). Or, not really a problem but still indicative of how goofy they are ... they'll shift the columns around in tables on occasion, just to try and beggar up someone who automates the download of the stuff.

It strikes me as madness, is any other sport this maniacal about keeping fans away from good stats?
Quote:
... Personally, I'd really like to see data for the goalies who've had these runs, and then completely fallen apart. I look at a guy like Jim Carey, couple of good seasons, and then his game fell apart. ...
I think that is more of a study in media stupidity than stats. Carey had a save% of .906 the year he won the Vezina. But a really low GAA, this seems to be the rationale for voting him the Vezina. (Washington must have been a heckuva good defensive team).

Hilarious to read the articles from back then. Apparently he was destined to win many more Vezina's in his career [facts be damned! :lol ]. And apparently, even though he wasn't putting up the gaudy save%'s that Hasek was ... Carey always made the 'clutch' saves. Perhaps, as the kids like to say, "He lost his clutch"? OR ... perhaps he just wasn't that good in the first place.

The outliers are interesting cases though. There is a pretty clear chance of any NHL player having a very good or very bad year in 2004/2005 ... just based on luck alone. One season just isn't long enough to narrow down the band of probability. If I were a gm I'd want to figure it out though ... you don't want to overpay for a guy who had a flukey year. Or sell off a player who just had a string of bad luck.

Proposed mudcrutch BillyBeaneStyle project :

Shooting percentage, by game situation (ES or PP or SH or EN) would be a cool thing to track. I'm sure you'd see some obvious trends [the average change in shooting% from year to year for a sample of players, and the distribution of this change] that should be enough to tell you how independent the stat is, and how accurately you are going to be able to project future results for a larger sample. You'll need a lot of years of data though ... because forwards only average about 100 or so ES shots per year I'd think.

Even with just imperfect data (overall Shooting%) and limited years ... at the very least you'd be able to spot some obvious patterns. And recognize some things that should be intuitively obvious anyways (like that Chimera's big Shooting% drop from last year to this one ... not much of a surprise, the guy didn't exactly have a long track record as a sniper in the AHL). And that if anyone is willing to bet you that Torres will score 20 or more next year ... take the wager without hesitation.

igor* is offline  
Old
04-15-2004, 07:37 AM
  #13
choppystride
Registered User
 
Join Date: Jul 2003
Posts: 471
vCash: 500
Quote:
Originally Posted by lowetide
One of the things that it can't reflect is the quality of the shots. igor, what is your opinion on that? Would they even out (iyo)over a period of time, making that argument a non-starter? Or would certain goalies on poorer teams tend to have poorer SP 5x5? And can you measure it? (Something by 9pm, thanks).
Not sure if the NHL stats are detailed enough to do so but perhaps each shot can be weighted by the shooter's rank as a goal scorer (or some similar metric).

choppystride is offline  
Old
04-15-2004, 07:51 AM
  #14
igor*
 
Join Date: Mar 2002
Posts: 3,276
vCash: 500
Quote:
Originally Posted by lowetide
I wish they'd had these stats when Gilles Meloche was saving 50 a night for the California Golden Seals.
...
Ya, they never used to keep very thorough stats on goalies. And when they did, nobody cared about SV% back in the day.

Quote:
One of the things that it can't reflect is the quality of the shots. igor, what is your opinion on that? Would they even out (iyo)over a period of time, making that argument a non-starter? Or would certain goalies on poorer teams tend to have poorer SP 5x5? And can you measure it? (Something by 9pm, thanks).
I think you already know my take on this. And if you pretended that goalie saves were not part of a hockey game, but rather part of an elaborate game of chance. You could test for 'fairness' if someone complained that the game was rigged.

Or, perhaps more intuitively:
How does the style of play and type and quality of chance change in regular season overtime iyo?

How much does the Save% for NHL goalies change in OT? (Just take a guess before you look it up).

How much does the rate of shots (shots per minute) change in OT? (just take a guess).

igor* is offline  
Old
04-15-2004, 08:08 AM
  #15
Blatherblah
Registered User
 
Join Date: Feb 2004
Location: New York
Country: Canada
Posts: 1,496
vCash: 500
Quote:
Originally Posted by H-Bear
*Smacks self*Stupid statistical tool used to determine (or guess) at how many subjects will fall into a certain range.
It's not stupid... It's just the standard normal distribution.

You can see the beautiful derivation of the normal here.


I should add that people often use mean=0, std deviation = 1, because it gives easier numbers to manipulate (simplifies that exponent!)

The normal conjures up memories of deriving the Student-t distribution a few years back.

Whoops: The site doesn't have the derivation, but a whole bunch of properties of the normal.


Last edited by Blatherblah: 04-15-2004 at 08:15 AM.
Blatherblah is offline  
Old
04-15-2004, 08:11 AM
  #16
H-Bear
Registered User
 
Join Date: Nov 2002
Location: North Bay ON
Country: Canada
Posts: 1,503
vCash: 500
Quote:
Originally Posted by Ikcotyck
It's not stupid... It's just the standard normal distribution.

You can see the beautiful derivation of the normal here.


I should add that people often use mean=0, std deviation = 1, because it gives easier numbers to manipulate (simplifies that exponent!)
Yeah, I've done 3 university courses in that crap (well, the business application anyhow) ... and I was kinda hoping to never see it again .

H-Bear is offline  
Old
04-15-2004, 10:53 AM
  #17
igor*
 
Join Date: Mar 2002
Posts: 3,276
vCash: 500
Quote:
Originally Posted by Ikcotyck
It's not stupid... It's just the standard normal distribution.

You can see the beautiful derivation of the normal http://mathworld.wolfram.com/NormalDistribution.html

I should add that people often use mean=0, std deviation = 1, because it gives easier numbers to manipulate (simplifies that exponent!)

The normal conjures up memories of deriving the Student-t distribution a few years back.
My memories of stats class are a lot fuzzier than yours. And most of this stats stuff is obviously a lot fresher in the minds of guys like yourself and H-Bear.

There is no shortage of papers written on the NHL by Mathematics PhDs. Most of it horribly misguided IMHO, but some of it really cool.

This is completely off the subject of goalies now, but:

There is a paper called "Simeon Poisson and the NHL" or some such, oilswell found it, and its wonderful stuff. By memory ... the guy who wrote it did a tonne of number crunching. And he was clearly a hockey fan who pointed out a lot of common sense items that weren't accounted for, but what he did worked anyways . He used Poisson's equation to predict the number of points a team should expect to garner in a season based on team goal differential.

Essentially: If your team has a goal differential of +20, and they are playing a team with a goal differential of -10 on home ice ... what are their chances of winning the game? Or tying it? How many points at the end of the year?

He got extraordinarily good results when compared to actual. Knocks the pants off of Bill James' "pythagorean theory" and all of the stuff that the guys at hockeyresearch.com have been farting around with for years.

It works very well, you can use his math and reasoning to and get within 3ish percent of the oddsmakers numbers for any game event (e.g. Oilers to win or tie, Oilers to win by 2 goals or more, etc.). And, given enough seasons, you'll match the point totals of the teams very closely (esp if you filter out OT results and empty-netters).

The two strange things about it IMO ...
1. It doesn't make a shred of sense.
2. It works well.

To my mind it would be analogous to this:
You and H-Bear play a crazy game for money as follows, we'll call it ikco-cars :-) ---> You each stand on a different street corner downtown and count cars. You're counting pink cars and H-Bear is counting orange cars. There is a time limit of, say, 20 minutes per game ... and whoever counts the most cars in that time wins one dollar (for the purposes of this example, you guys are cheapskates :-D ). Lowetide (who has been counting cars by colour downtown for years and has kept meticulous records ) knows that on average 2.4 pink cars will pass you every 20 minutes, and likewise 2.7 orange cars for H-Bear on his street corner. Oilswell points out that the results of ikco-cars is remarkably similar to the pattern of scores you'd see if two nearly-equal and average hockey teams played each other (H-Bear representing the home team) ... the rest of us are puzzled.

Not a whole heckuva lot like a hockey game, is it? Heck, the cars aren't even on the same street and have no influence on each other! But somehow it works well, and it is an easy and quick way to check perception against reality for a lot of things (if perception and reality are far enough apart). But it's still a poor analogy for a hockey game ... and I'm afraid you're going to have to completely rethink ikco-cars for it to get closer than it is .

My 2 cents.


Last edited by igor*: 04-15-2004 at 11:10 AM.
igor* is offline  
Old
04-15-2004, 08:32 PM
  #18
kraigus
Registered User
 
Join Date: Nov 2002
Location: Waterloo, ON
Country: Canada
Posts: 1,173
vCash: 500
Send a message via ICQ to kraigus Send a message via MSN to kraigus
Quote:
Originally Posted by igor
Quote:
Originally Posted by H-bear
Conklin, Salo, and to a certain extent, Markkanen played behind one of the best even strength teams in the NHL. Could I not jump to a reasonable conclusion that there stats are somewhat skewed in their favour by having this advantage.
This strikes me as a bizarre argument. Having said that ... I think you are in the majority here. Which baffles me.
I realise the argument's moved a bit beyond this, but this particular bit intrigued me enough to click reply.

It strikes *me* as a bizarre argument, because it strikes me as being a chicken/egg issue: are their save percentages good because the team plays very well 5 on 5 (or 4 on 4), or is the team very good 5 on 5 because they're good goalies in ES situations?

It's too bad (for us) that more pros good at a) crunching numbers, and b) tossing irrelevant numbers don't do hockey stats. I think somebody posted here last fall about a UBC project that's working on computerized scouting techniques. Well, there's a second component to that project here in Waterloo - I know the grad student working on it. This is not a diss against him or the rest of the team - he's smart and a decent enough programmer (and he bought me coffee for helping him with some computer problems, he's a nice guy too) - but... he's never seen a hockey game in his life. Not even half a period. I have to wonder how valuable their program is gonna be. I tried to invite him over to watch a game or two, but he said he was too busy working.

I told him if he can make it work, they should sell it to the Oilers, but if it sucks they should sell it to the Flames. He didn't get it, unfortunately.

I'm rambling now, so I should probably stop.

kraigus is offline  
Old
04-15-2004, 09:41 PM
  #19
Lowetide
Registered User
 
Lowetide's Avatar
 
Join Date: Feb 2002
Country: Canada
Posts: 13,281
vCash: 500
Quote:
Originally Posted by kraigus
I tried to invite him over to watch a game or two, but he said he was too busy working.
Many years ago, I knew a guy who had spent much of the 60s in Europe (army). He told me that when he was in France, everyday at about 4pm he would watch the French army practising their retreat (no lie).

Sometimes, you're pooched from the get-go.

Lowetide is offline  
Old
04-16-2004, 01:39 AM
  #20
igor*
 
Join Date: Mar 2002
Posts: 3,276
vCash: 500
Quote:
Originally Posted by kraigus
...
It strikes *me* as a bizarre argument, because it strikes me as being a chicken/egg issue: are their save percentages good because the team plays very well 5 on 5 (or 4 on 4), or is the team very good 5 on 5 because they're good goalies in ES situations?....
A great bit from TVs "The West Wing" ...

HOYNES
Have you ever seen a good hockey game?

SAM
No.

HOYNES
Me neither. I love sports, I just can't get next to hockey. See, I think Americans like to savour situations. One down, bottom of the ninth, one run game, first and third, left handed batter, right hand reliever, infield at double play depth, here's the pitch. But scoring in hockey seems to come out of nowhere. The play-by-play guy is always shocked. LePeiter passes to Huckenchuck who skates past the blue line. Huckenchuck, of course, was traded from Winnipeg for a case of Labatts after sitting out last season with ... "Oh my God, he scores!" So, what's going on?


Obviously the guy who wrote this (Aaron Sorkin I think) isn't being very flattering to the game we love :lol But its funny ... and pretty much spot on. At least I think so (said igor, knowing he'd be wildly unpopular for saying it out loud ).

The NHL game, especially now, with the wicked fast shots from the new sticks and the monstrous and technically sound goalies ... the majority of EV goals happen from a fortunate bounce, or a series of fortunate bounces. The last one being that one which ends up in the net. Maybe its just me ... watch The Score and count 'em for a couple of days, (in the playoffs they usually show all the goals even the ugly ones).

Its like a freakin' pachinko game out there, the puck must bounce off some part of Ryan Smyth a few times every shift. That guy does have a special knack for getting in the way. And IMO a low-percentage shot with traffic has a better chance of going in than a clean shot from anywhere but the slot, though probably no at the first whack.

Even the occasional one-shot even-strength goal ... usually its a whiff by the goalie, sometimes a laser shot to one of the tiny areas that the goalies equipment doesn't cover ... unless you're in the slot or the goalie has lost his net ... just no room. And even those occasional laser shots ... even Mario doesn't make those very often ... its like trying to hit a bullseye with a dart from 30 feet (even the draft monkey can't do that regularly!). Hard one-timers with the goalie moving are the best bet usually ... just aim for the middle of the net Mike-Bossy-style, get it off your stick in a hurry and try to get it airborne if you can ... at least there could be a rebound, and maybe the hockey Gods will direct it to a team-mate. Create enough chances and eventually you'll get some breaks.

There is a reason why the statistician in the example above did so well, even though he barely dipped his toe in the reality pond. It really is pretty random, this goal scoring. Isn't Poisson's equation also used to predict the odds of lightning strikes?

I can't recall an NHL goal-scorer describing it much different than that. Robitaille is a nice simple quote on goalscoring (he's a good quote for most things) ... "just shoot it at the net when you can. Sometimes some will go in." Somehow Ales Hemsky seems like the only NHLer who hasn't clued in to this. :-)

I've looked for it (oilswell, dawgbone, lowetide and others seem convinced that good players have a big affect on their goalies, I used to think so to, and they're all knowledgeable guys, so I started looking for it again recently). But for the life of me ... I just don't see it. When good EV players are on the ice (low EV goals against rate) ... the puck stays in the other end of the rink most of the time they are out there, that's why they are successful IMHO. Sometimes shyte happens when they are out there, a bad play or bad line-change by them or a team-mate, a good play by someone on the other team, an unlucky bounce ... whatever. But usually the puck stays in the other end when it gets there. And gets out of their own end quickly. MacTavish's obsession with 'a good first pass out of the zone' ... that seems pretty sensible to me. [BTW: The NHL actually tracked zone time by player, and presumeably still does, must be top secret though ]

And the OT thing. 4on4 OT is a blast IMO. Lots of open ice, room to skate. Players taking bigger chances ... because there's usually less to lose in OT, esp East vs West games. And most importantly more 2on1 s. Yet over the last two years the total save% in OT is only .0055 lower than the rest of the game ... and it is about the same as the 2nd period (with the longer change there). The shots on goal rate though, that goes through the roof ... 45% more per minute in OT.

If the NHL went to 4on4 the games would be a lot more fun to watch, at least I think so. Wouldn't be as open as 4on4 OT is now, but more flow, more talent on display. And the shot totals would definitely go up, and the goal totals would go up, and the save% (just my guess) would probably stay the same, or damn close to it one way or the other.

And goalies get traded, or more often large parts of the team in front of them does ... the logo might stay the same, but for most teams there is a lot of turnover.

Yet if the NHL goalies came around to my house to play igor-dice in my unlicensed basement casino :lol (its an elaborate craps game with weighted dice, and nobody knows for certain just how much which dice are weighted, and by how much). I'm pretty sure I could survive an audit from the gaming commission ... its a clean game. And after 5000 rounds of igor-dice ... I'd feel pretty comfortable writing a betting line for the next 1200 rounds with the same player. I'd have to take a pretty fat margin if I'd only seen 1000 or so previous rounds from a guy though. Anything can happen in this world, hell the draft-monkey might type a novel while we're playing, just not likely. :-)

But, in the unlikely event that any crazy beggar is still reading this. (I've noticed by the read-count that this is wildly unpopular). I'm going to end this tireless ramble now.

igor* is offline  
Old
04-16-2004, 01:53 AM
  #21
igor*
 
Join Date: Mar 2002
Posts: 3,276
vCash: 500
Quote:
Originally Posted by kraigus
... I know the grad student working on it. This is not a diss against him or the rest of the team - he's smart and a decent enough programmer (and he bought me coffee for helping him with some computer problems, he's a nice guy too) - but... he's never seen a hockey game in his life. Not even half a period. I have to wonder how valuable their program is gonna be. I tried to invite him over to watch a game or two, but he said he was too busy working.
Okay, one more post. Can't pass by this.

Your buddy might want to look for a paper by "Voyer et al". Was published in the journal of sport I think. I'm pretty sure that the guys who wrote that one didn't watch hockey either, but they sure know their math! It would take me a lifetime to get my head around their calcs ... but I think you'd get a kick out of the methods and conclusions. Plagiarism! ... its the key to spare time in college, doncha know.

Or maybe you could talk your buddy into having a chat with a hockey scout or something, just to save him from maybe spending a lot of time going in wrong directions. Do the Oilers have any scouts that live out that way

igor* is offline  
Old
04-16-2004, 03:18 PM
  #22
mudcrutch79
Registered User
 
Join Date: Jul 2003
Location: The Big Smoke
Posts: 3,903
vCash: 500
Alright, so I'm looking at the nhl.com event sheets, which provide a listing of all of the events that have taken place in the game. link Can anyone here explain to me what the hell the ! and * mean? I was wondering if perhaps they refer to deflections or blocked shots or goalposts, but they are counted in the shot totals, which blocked shots and goalposts aren't so I'm really not sure. I'm trying to throw together a macro of some sort to snag all of this info off the site, sort it out, and maybe put it into something useful so we'll see how that goes. That could maybe give us some more insight into the SV% data, as you can look at it from the situation in which the shot takes place, the shooter, and where the shot was taken from even.

So if anyone can explain those denotations to me I'd be much obliged. Oh and if anyone knows anything about real property, and cares to write an exam so I can do this instead...life would be very good.

mudcrutch79 is offline  
Old
04-16-2004, 05:43 PM
  #23
barto
Registered User
 
Join Date: May 2002
Location: Edmonton
Posts: 2,465
vCash: 500
Quote:
Originally Posted by kraigus
*snip*
I told him if he can make it work, they should sell it to the Oilers, but if it sucks they should sell it to the Flames. He didn't get it, unfortunately.
:lol

Attaboy, kraigus!! Thanks for the chuckle. Keep tabs on him, though, just in case it does turn into something useful!!

Bart

barto is offline  
Closed Thread

Forum Jump


Bookmarks

Thread Tools

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off



All times are GMT -5. The time now is 12:06 PM.

monitoring_string = "e4251c93e2ba248d29da988d93bf5144"
Contact Us - HFBoards - Archive - Privacy Statement - Terms of Use - Advertise - Top - AdChoices

vBulletin Copyright ©2000 - 2014, Jelsoft Enterprises Ltd.
HFBoards.com is a property of CraveOnline Media, LLC, an Evolve Media, LLC company. 2014 All Rights Reserved.