Pat McLean’s been on an absolute tear lately, writing some great stuff to read. It’s been like reading Bill Simmons, only if Bill Simmons wasn’t a 39 year old frat boy. (I actually went and looked up Simmons’ age, if only because it’s hard for me to believe that Simmons is only a year or two younger). If you’re not a regular reader, check it out.
* * *
A while back, Vic Ferrari commented on something he called a PDO number. A PDO number is a team’s ES S% plus their ESSV%. The idea, and I know that this will make some people uptight, is that there’s a lack of sustainability in that number - if it’s high, it’s going to come down; if it’s low, it’s going to go back up. We’re just past the quarter pole in the NHL, so I thought I’d check out how that idea holds up over the years for which we have information (2003-04 onwards). Here are the twenty best and worst PDO numbers from the first quarter since the 2003-04 season, along with the team results for the remaining three quarters:
Overall, the effect is really startling. At left, I’ve summarized the differences between the best 20 teams and the worst 20 teams in terms of the percentages for the first quarter and then the final three quarters. While there’s still some differences in the final three quarters, they’re much, much smaller.
Just as an aside, dealing with the whole mention of luck. There’s a fine BP article from a few years back by former BP staff writer and current Tampa Bay Rays
As a brief aside, it’s important to clarify what is meant by batting average being subject to great deal of “luck.” This is not to say that all major league hitters are equal when it comes to AVG, and the differences evident between them are entirely random. Rather, players have a theoretical AVG-ability that varies from player-to-player, but the sample size of a season is too small to accurately reveal that every year. The high volatility of AVG from year-to-year–the statistical “noise,” if you will–is sufficiently large enough to obscure the differences between many major league hitters of similar ability.
When I’m talking about luck, that’s the issue I’m talking about. It may be that there are other factors in play, the kinds of things that people love to talk about - Horcoff taking his shots from the outside, Smid making great defensive plays but it seems to me that in the long run, these things seem to disappear. If Shawn Horcoff will, in the fullness of time, start again getting his nose dirty in the crease and Laco will start losing the backdoor play again, I have a hard time getting worked up about their failures in that area for the time being. If the team percentages seem to snap back to much closer to the NHL average over the course of of an entire season, it seems silly to me to get too worked up one way or the other about what’s happening at any given point in time. That’s a separate issue though and one I’m sure I’ll revisit at some point in the future.
Additionally, this all comes back to the idea of outshooting that I like to harp on. As the rate by which a team gets outshot at evens increases, you need to rely on the percentages more and more. I’ll come back to this below; the 2007-08 Capitals and 2007-08 Oilers are an interesting comparison here.
Coming back to the lists of teams above, I really find the teams that make the list to be fantastically interesting - there’s just way too much cool stuff there. Let’s start with the 05-06 Senators. Shortly after the first quarter ended in 2005-06, the following headline was published in the Globe and Mail:
ARE THE SENS THE BEST EVER?
The ‘talented bunch’ from Ottawa is manhandling its opposition this season. The Senators are an offensive juggernaut, and their defence isn’t too shabby, either.
Whoops.
The percentages disappeared for the Senators in the final three quarters and, given that was the foundation of their claim to the title of “BEST HOCKEY TEAM EVER!!!!!1111ONEONEONE!!!111″ (in fairness to the writer of the story, Tim Wharnsby, the story really didn’t match the headline), so went the Senators claim to greatness. It also strikes me as more than a little bizarre that all three Senators teams between 05-06 and 07-08 had PDO numbers that place them in the top 20 of the 120 team-seasons that have been played since 2003-04. It’s not a wonder that every winter seems to bring stories about what’s wrong with the Sens; they crank up the expectations with the PDO number every year in the first quarter and then, once luck evens out, things get ugly.
I’m happy to see the 2007-08 Flyers on the list of teams for whom the bubble popped. I had them as a team that was due to fall off a cliff around this time last year and sure enough, it did. If I was a Flyers fan, I’d probably blame Joffrey Lupul.
The 2005-06 Carolina Hurricanes: I’ll say it again - I can’t believe that the Oilers lost to that bloody team. The hockey gods have a wicked sense of humour - Oilers win as serious underdogs (Detroit) and mild underdogs (Anaheim and San Jose) and then lose to the team that they’re probably better than. Cam Ward has been junk ever since he stopped Pisani with about four minutes left in G7. Hilarious.
The teams that performed poorly in the first quarter are fun to look at too. There’s the 2007-08 Washington Capitals, handily outshooting the opposition at ES, but with little to show for it. Bruce Boudreau comes in, Washington keeps outshooting the opposition but the percentages snap back to normal and suddenly Boudreau wins the Jack Adams Trophy. Does it diminish the credit that he gets if we know that the PDO number almost always comes back to closer to 100% for the teams that have had brutal first quarters? I would think that it must. The Oilers and Caps are fun to compare here too - the Oilers were -16 and the Caps -9 at evens in the first quarter. For the final three quarters, the Oilers were -3 and the Caps +27. The Caps were better positioned to take advantange of the percentages turning because the puck was spending more time being directed at the other team’s goalie than it was at their own.
The 2005-06 Sharks are interesting too - they were terrible to start the year, made a trade for Joe Thornton and didn’t look back. While adding Thornton obviously improved their team and, I would argue, made it more likely that they’d have a high ES S%, they weren’t nearly as bad as they appeared in the first quarter.
Mirtle was picking on the 2005-06 Columbus Bluejackets (something I think that we’ve all done, given that they’re an horrific franchise) around the end of the first quarter that year, writing:
I was going to write a post on the Blue Jackets as the worst team in hockey over the weekend. I caught their game with the Oilers on Friday at work, and while the score was only 3-1 for Edmonton, Columbus looked so disinterested and disorganized that the game’s outcome was never really in doubt.
Columbus had a terrible PDO number that year through the first quarter; they ended up posting a 101.0% through the final three quarters and, while they didn’t really sniff the playoffs and other aspects of their game improved as well, they ended up third in the Central, suckering their fans into buying tickets for another season.
Here’s how things look this year, at the end of the first quarter:
There’s all sorts of fun stuff in there. The Canucks and Oilers make an interesting pair to look at together - they have very similar ratios of SF/SA, but the pucks have gone in for the Canucks and not for the Oilers while, less surprisingly, Vancouver has stopped pucks that the Oilers goalies haven’t. Looking at the two teams and at the tables above, I think that the only reasonable expectation is that this gap will narrow with time.
The Bruins are one of the darlings of the media right now; everyone is fawning over them. Kevin Paul Dupont of the Boston Globe just wrote a story praising virtually everything about the team, which featured some of the following lines:
1. Claude Julien. His coaching thus far has been masterful, among the keys the standard he set with the yanking of a gritless Phil Kessel from last year’s playoffs…Julien’s X-and-O scheme is simple, straightforward, and the minutes he doles out are earned, not granted.
…2. Tim Thomas. Forget all the bunk about his unorthodox style. Thomas stops the vast majority of pucks that come his way. That’s the job, right?
…
5. David Krejci and Kessel. Krejci’s spatial awareness and timing are sublime, and they seem to grow by each increment of 2-3 games. Kessel, post-playoff benching, is developing legit feistiness (a distant cousin of crustiness) to his attacking game, and has turned into a responsible citizen in three zones.
Dupont sees an awful lot of stuff there. We’ll see how the final three quarters go for the B’s, but I would expect that at least some of this praise will be tempered by a following PDO number, followed by columns about how they need to get back to playing like they did in the first quarter.
Detroit is going to be scary if their PDO number moves back towards 100. Similarly, Toronto is an interesting team - they outshoot the opposition at evens by a healthy amount but have the worst PDO number in the league, thanks largely to their inability to stop the puck. Maybe Brian Burke’s first miracle, as certified by the Toronto press corps (I think we should call them “The Disciples”), will be to push their PDO number towards 100.
The Bluejackets appear to have turned into a legitimate ES team. Ken Hitchcock and Scott Howson have had them moving in the right direction for a while now and they posted a healthy ES + in the first quarter without relying on a good PDO number.
Anyway, this is, to my mind, pretty interesting stuff. We’ll check back in at the end of the year and see if the trend of the past four years held up.
A PDO number is a team’s ES S% plus their ESSV%. The idea, and I know that this will make some people uptight, is that there’s a lack of sustainability in that number - if it’s high, it’s going to come down; if it’s low, it’s going to go back up.
Not just teams, either: players too. Though common sense (as always) needs to be applied; there’s a wider range in which players should “naturally” sit (according to ability). E.G. if Mario Lemieux’s PDO# in 1992 is 103%, you don’t hold your breath for it to come back down in ‘93.
Todd Bertuzzi should be expected to be below 100%, because he doesn’t create EV offense, turns the puck over too easily between the blue lines, etc. But right now his PDO# is 94.5%, and there’s a lot better chance that it’s 99% over the remainder of the season than of it sitting right there for the duration. (Though I will regrettably note that he continues to invent new ways to disappoint fans of his teams, so, it’s possible.)
Meanwhile, most Canucks should be expected to have over 100% by simple virtue of their well above average goaltending (pending Luongo’s recovery). However, Willie Mitchell’s Norris Trophy aspirations will probably take a hit when his number inevitably drops from it’s present level of 107.6%. Likewise, count on Kyle Wellwood’s committment to his team and career to be questioned once again when he plummets from 108.2%, where he is right now.
You’re stealing my thunder Matt. Mirtle had this to say about Sharks defenceman Douglas Murray last year at around the first quarter:
That’s part of the reason why I find the on ice/off ice statistics Behind The Net offers so useful: Just because you don’t end up with a lot of points doesn’t mean you’re not assisting your team’s offensive production.
Just look at San Jose’s Doug Murray, who has only four assists in 25 games this season, but who, when he’s on the ice, sees his team’s production go from 1.95 goals per 60 minutes to 4.30, which is 10th best in the NHL.
Murray’s first quarter PDO number was an INSANE 125.0 or so. No. 2 was like 116, IIRC.
There seems be a lot of talk in the media over how poorly Montreal has played so far this year.
I wonder how much this will intensify once their PDO number begins to regress back to 100…
Granted, they’ve been woeful on special teams in terms of the percentages. Maybe the two effects will balance one another out.
My biggest takeaway from this is just how good the San Jose Sharks have been so far this year. Winning to the degree that they have without having lopsided percentages is frightening.
It’s also interesting to see that teams with a low shots ratio seem to have a better chance at a high PDO number. 8 of the bottom 10 are 1.00 or greater, 4 of the middle 10 and only 2 of the top 10 are outshooting teams. It doesn’t seem to hold as well for the historical numbers mind you, but I so find that interesting.
Great stuff. I’ve been waiting for this post since you mentioned you were working on it awhile back.
Tyler actually refers to PDO as “the stat that allows me to slam James’s site” in casual conversation.
Interesting stuff TD. I’d like to hear more about it in terms of applications to this season.
Tyler actually refers to PDO as “the stat that allows me to slam James’s site” in casual conversation.
Do you have a daughter? Much younger sister?
Maybe I shouldn’t finish that thought…
Great stuff here Tyler… I’ve only really used it in a micro sense, mostly do to a lack of time.
Thanks for the kind words Ty, appreciate it.
I have to get my head around this, its interesting stuff but it makes my brain hurt.
Interesting that Buffalo and Nashville sustained high numbers throughout 2006/2007 but Buffalo actually went from very high shooting plus good save to very high save and good shooting.
Meanwhile Nashville was right about on the money for both. Not much movement.
Both pretty good clubs.
Tyler actually refers to PDO as “the stat that allows me to slam James’s site” in casual conversation.
Tyler edited this post to tell me that future posts on this site in which I make zero contribution but act like a jerk will be edited. He’s a pretty lenient guy with respect to the give and take of the comments section but, given that a very high percentage of my posts feature absolutely no content other than a shot at somebody, and given that I rarely display any interest in discussing the point at hand without behaving like a tool, a line has been drawn.
Enjoyed this post very much.
Matt: Does anyone post (or host) PDO numbers per player? Perhaps we could prevail upon Mr. Desjardins at Behind The Net to include that stat.
All: Are there fundamental differences between ES ST% and ES SV%?
I completely understand the regression to the mean concept on the ST% side of it–there is a lot of luck involved in the WHEN and WHERE of juicy rebounds. But part of my brain still rebels at the idea that ES SV% is always destined to return to the mean.
I’m guessing some outlier goalies like Hasek probably strung together mind-blowing ES SV% quarter after quarter. (How do Luongo’s quarters compare to the NHL mean BTW?)
Part of the “luck” factor with respect to TEAM ES SV% might simply be playing time or injuries. For example, in the case of the 2005-06 Thrasher (one of the bottom 1st Q outliers above) The team was missing its starting goaltender (Lehtonen) for the 1st and 2nd Q of the season. The team stuck with AHL and ECHL level net minders for much of the 1st half. After Lehtonen returned, he played almost every game the rest of the season. According to your table Atlanta’s ES SV% improved by over 20 points in 2-4th Q.
I guess what I’m getting at is that the poor ES SV% in this case was a matter of injury–>poor quality goaltending–>poor SV%. In the case of goaltending I suspect that injuries will comprise a much bigger component of the “luck” factor in “team performance” than in the case of ST%. I also wonder if certain outlier caliber goalies enable their teams to avoid a regression to the mean in that stat.
Falconer -
I don’t know anyone who does. I routinely check ‘em out when I look at the ES numbers at timeonice but that’s in my head. I intend to do the individual first Q numbers at some point and I have last year’s handy as well.
I completely understand the regression to the mean concept on the ST% side of it–there is a lot of luck involved in the WHEN and WHERE of juicy rebounds. But part of my brain still rebels at the idea that ES SV% is always destined to return to the mean.
I’m guessing some outlier goalies like Hasek probably strung together mind-blowing ES SV% quarter after quarter. (How do Luongo’s quarters compare to the NHL mean BTW?)
Well, it’s regression towards the mean, not regression to the mean right? Plus, as you point out below, there’s some selection at work, as guys who get bombed or who can’t score don’t get any more games.
With Luongo and Hasek, I’d expect them to be able to put together quarters that are consistently above the NHL mean. It’s not huge though - there’s just not that much spread in save percentage, even though a single point for a starting goalie is pretty big.
Part of the “luck” factor with respect to TEAM ES SV% might simply be playing time or injuries.
Completely agreed.
First, if Pat McLean is the Bill Simmons of the Oilogosphere, how come I haven’t read one 1500 word story about him going to Vegas with his buddy Sully and enduring a grade III stomach punch, only to get rid of Sully and win $500 at the sports book like Daniel-san against Cobra Kai proving the Ewing Theory applied to Sully and also Roger Clemens did steroids?
Second, interesting number. And this is going to be a really kindergartener question (which is why I tried to be funny in the first paragraph), but I’m not sure what this stat has to do with anything. Basically, we’re seeing good teams with lousy PDO numbers and lousy teams with good PDO numbers and it doesn’t seem to correlate to anything, really. I understand that it seems to regress to the mean but that’s nothing we didn’t already know: if Tom Gilbert shoots over 11% in his rookie NHL season, well, he’s probably not going to do that again.
interesting stuff. I can’t help but wonder if the bottom five top five were compared as opposed to splitting it down the middle, if that would show exceptions.
MC: Thanks for showing me how to get the season totals. I’ve been putting the Corsi numbers into a spreadsheet game-by-game this season and posting them periodically on my Thrashers blog. I’ve been trying to get the fans used to seeing them.
re: ES SV% by player. How do you guys interpret these numbers? If I guy has a low ES SV% do you see him as just being “unlucky” who is the victim of some bad bounces in a small sample of ice time, or do you take it as a sign that he is a weak defensive player?
Lord Bob: what it means is that TEAM ES SF/SA ratio is pretty darn important. You might get “hot” or “lucky” for a quarter and over- or under-perform you shot ratio but in the long run you come back to that reality.
@Lord Bob: Well, you must not be the target audience because it seems that you already get the point. The point I’m driving at is that the media and fans get suckered by the percentages. I was in the habit of listening to Stauffer’s show last year and he was on and on and on about the Flyers. He wasn’t the only one: people everywhere were talking about the Flyers miracle. It was all imaginary.
@rent-a-goalie: I’m not splitting it down the middle. These are the top and bottom 20 out of 120. (30 teams x 4 years)
@The Falconer: I’m not sure yet what to make of it. I don’t think that individual players have enough control over it for it to mean anything. Let me throw some individual correlations at you to show what I’m talking about.
466 players were on the ice for at least 200 ES shots in both the first and second half while wearing the same number and playing for the same team. Of that group, the correlation in on-ice SV% from the first half to the second half was .06. The correlation in on-ice shooting percentage was 0.28. The correlation in SF/SA ratio was .60.
The last correlation, the one dealing SF/SA ratio kind of intrigues me - it’s about the same as the correlation for on base percentage year over year in baseball, as you’ll see if you look at the story linked above. That’s respectable. I’m not willing to conclude that it’s an individual ability driving that - I’ll want to see how it performs over a period of years first, because guys are, by and large, in similar situations throughout the year. Still, it strikes me as intriguing.
If you split the group into forwards and defencemen, some more interesting stuff shows up. They have very similar correlations, from one half to the next, in save percentage and in SF/SA ratio (SV%: .06 for the forwards, .05 for the D; SF/SA ratio: .61 for the forwards, .59 for the D) but the forwards have a much better correlation in terms of on-ice shooting percentage (.34 to .14), which kind of makes sense if you think about it.
Tyler: Thanks for posting the correlations. That .06 for SV% is pretty closer to random noise–not unlike the year-to-year correlation for team success in the shootout.
On the other hand, the SF/SA is pretty impressive at .60, I ran a bunch of correlations on team stats on a year to year basis and I was struck at just how weak the relationship was across years in many areas.
I also ran a bunch of correlations on year to year player ES TOI, PP TOI and SH TOI as well as ES, PP and SH scoring rates. The data shows that defensemen usage (TOI) tends to be more consistent than forwards across seasons. The scoring rates (pts per hour) were less robust than I would have hoped. Perhaps the most disappointing (for me at least) is that ES Scoring is a rather poor predictor of future PP scoring rates.
The upshot of this all seems to be that Usage (TOI) and Net Shots are among the more consistent metrics available on the player level. Of course Shots and Minutes are all elements that have a much larger sample size than scoring (relatively rare) or PP scoring (ever more rare).
My dream is to have something like PECOTA available for hockey but sometimes I despair that the problems of chance and poor measurement are so large as to make such a project a fools errand.
The thing about Pecota is that it actually predicts very wide bands for players and then breaks it down into percentiles. Take a look at Manny’s Pecota card, for instance. They had him at somewhere between .249/.342/.411 and .311/.416/.558, with a weighted mean of .281/.380/.486. That’s a pretty astonishing spread - one of those players is an MVP candidate while the other is a useful fourth outfielder. I think that, even in baseball, there’s a pretty wide spread that can occur in a player’s results.
My point is, I don’t know if you should necessarily worry that this isn’t possible in hockey. If anything, I’m more encouraged seeing some of the data that’s coming out and what we can tell. The question will really be once we have some more years of data, whether we can start to get a better sense of the size of the difference between Ales Hemsky being on the ice versus Zack Stortini. I’m optimistic.
@JLikens: Montreal’s been weird this year. It seems like they’ve been better at EV — not hard, really; they’ve been terrible since the end of the lockout at generating offence on a level playing field — and I think Tanguay’s been part of that, but we’re back to the Bad Kovalev of ‘06-’07 this year, and this time, he doesn’t have Sheldon Souray’s career year to prop his numbers up. (Actually, I’d argue he’s been worse than he was in ‘07; can’t seem to connect a pass or hit the net more than a handful of times in the six or eight games I’ve seen this year.)
@McLea: *snicker*
Now I can kinda want to know what McLea actually said.
Kent - you’ve probably seen it all before
Lord B. - the scary thing is I have only a tiny inking of who Bill Simmons is - combination of age and ignorance
He wasn’t the only one: people everywhere were talking about the Flyers miracle. It was all imaginary.
Imaginary is surely the wrong word there. Believe me when I say I understand the point of the comment and the post, but (and particularly if you’re concerned about making people “uptight”) it could easily have been earned through nice finish and good goaltending. Being unsustainable over the long haul is “lucky” in a specific context only.
Jarome Iginla might score two goals tomorrow night on well-placed shots set up by beautiful passes, where the goalie had no chance. The fact that he’s not going to score 164 goals this year, or even 64, does not mean that those 2 goals were undeserved; there’s no ‘imaginary’ about it.
I liked your answer to Lord Bob otherwise, though. That’s really all this is; a way to quickly eyeball whether fantastic (or awful) results are deriving from things we know are unsustainable, no matter how well-earned.
I should add, in a form that’s intuitive. Perusing the variation in a stat that’s based on the # 100 is a lot more accessible than checking EVShoot% against 8.5%, or EVSV% against 0.915.
re: ES SV% by player. How do you guys interpret these numbers?
I try to avoid it.
I have no doubt that an individual player can impact the quality of shots & scoring chances against (the opposite take is self-evidently absurd, I think), but as Vic notes in this excellent post:
The area between the two is part randomness (we have 100 dice seasons, but only one NHL season after all), and partly the effect of players impacting shot quality as defenders, and partly the effect of the guys who play more against better shooters and play makers. That’s a lot of theory competing for a tiny patch of real estate.
Imaginary is surely the wrong word there.
Yeah, I knew that as I was writing it. Don’t know why I persisted.
Thanks for the explanation, Mudcrutch. I’m so used to not understanding this sort of thing intuitively that when I do understand what you’re talking about I assume I don’t.
Fantastic post and discussion as usual.
Falconer’s point about SV% being a slightly different animal than S% is the first thing I thought of, since a team’s S% is driven by 20+ players whereas a team’s SV% is highly influenced by one.
I do recognize that goaltender is a strange position in that, on one hand, you have this one single player that can influence the outcome of a team’s success in a way that no other position can, maybe in all of professional team sports. But on the other hand, in actual reality they rarely do, since most NHL goaltenders are so close to the mean.
So therefore, it’s been my belief that a team basically shouldn’t even care much about goaltending UNLESS they have a chance to be at the fringes, i.e., a good bit worse than average or a good bit better. Because then it matters, quite a lot.
So my question is, Tyler, are you saying that there really aren’t significant fringe goalies? Let’s acknowledge the fact that one season isn’t an adequate sample size from which to glean an assessment of skill. Fine. Is there any way we can get a list of all active goalies’ career ES SV% (preferably with number of shots faced listed)?
It seems intuitive to me that a team really could expect to persistently under- or over-perform league average SV% if they accurately assessed goaltenders. But ironically, it probably takes such a large sample to get a good assessment that, by then I guess you gotta worry about the goalie sustaining that skill you just figured out is real.
Thanks for the link to Vic’s Dice Rollers vs ES SV% Chart. Good stuff.
One more question on my mind is this–when you looked at 1st half and 2nd half correlations which metric was strongest?
1) NHL defined SF/SA
2) Corsi Numbers
3) Fenwick Numbers
Answers and more here, Falconer. Be sure to go through the comments as well.
Corsi & Fenwick are more repeatable and better predictors of EV+/- than SF/SA in the medium term, although the longer you go the more they converge… so over long periods it’s probably best to go with SF/SA. Then at the very least you avoid arcane discussions about the true/actual value of missed or blocked shots.
Also IIRC, in most of these things Vic and others have run for the whole league, Corsi# is marginally better than Fenwick#. I still prefer the F# because the Flames, who I follow the most closely by far, consistently have an absurdly high ratio of Shot Attempts Blocked For/Against, and general Shots and F# are more reflective of scoring chances *in their games* than Corsi.
Thanks for all the answer guys.
Matt: I did read that post a long time ago but I wasn’t as focused on some of these question as I am now. So if I’ve got this right the causal arrows appear to run: Net Shifts End OZ/DZ–>ES SF/SA–>ES GF/GA in terms of predictive model?
Is there any way I can pull up the “where shifts start” and “where shifts end” data from the 2007-08 season for Atlanta and other teams? I can get the current year, is there any way to specify season on Vic’s Time On Ice site?
Just catching up as fast as I can in 10 minutes or less but I wanted to say great original post and also great 1st comment Matt. I don’t know why I think of these particular ideas as common sense when it seems apparent that so few hockey fans believe them but I love the work you do and especially when it meshes with my own perspective of the game. I regret that I’m such a fringe player in the new ideas league but never regret my decision to click through to this and the other great sites.
Falconer: ATL shift-end numbers for 07/08
Muck around with the URL based on the ones you do know, and you’ll hit something soon enough.
The hierarchy re: predictive model you have noted is correct. Believe it or not, the best predictor there is for TEAM EV outscoring in the 2nd half is where your shifts were ending in the 1st half.
For individual players, it’s about the same as for Corsi/F#, according to Vic’s pretty picture at least.
Matt or Tyler: I seem to recall that in the first couple of seasons of the RTSS the www.nhl.com boxscores had a Zone Time stat.
Since Faceoff Location is basically a proxy variable for zone time, have any of you tried to dig up the old “zone time” stats for those seasons and checked the correlation with Team GF/GA? It probably lumps all ES,PP,SH situation together which would make it less useful, but it could be a nice little piece of evidence for the link between puck location and shots and goals.
Yeah, one of ‘us’ has, you’ll be shocked who. There’s a little search box at the top of all Blogger blogs including IOF; you may find it handy.
It was all imaginary.
Diction issues aside, isn’t this dramatically overstating it? The Flyers, even after coming back down to earth in terms of their PDO, still finished with the 5th most points in the East and made it to the Final Four. This was a pretty good hockey team last year, no matter how you slice it. A non-insignificant piece of their early success was in fact sustainable over the long-haul, notwithstanding a regression toward the mean in their PDO.
A non-insignificant piece of their early success was in fact sustainable over the long-haul, notwithstanding a regression toward the mean in their PDO.
What piece is that?
My feeling on the Flyers last year is that:
A) The fact that they finished with the 5th most points in the East and won two playoff series doesn’t necessarily show that they were a pretty good team. They could easily have been average to below average.
B) They finished last season with a -13 goal differential at ES.
C) Their power play was outstanding.
On that last note, it’s interesting to see that the Flyers’ PP success was also driven by a very high S%. Have any of y’all done research on how PDO numbers behave on special teams? Do we see similar regressions? Is sustaining an above or below average PDO more possible, less possible, or the same as ES?
I would think we’d see similar regressions, but then again, I don’t know. A team’s PP S% is probably more influenced by a couple (possibly very skilled) players.
What piece is that?
I dunno, whatever it is beyond PDO that goes into making a team succeed or fail. It appears to be some rather complicated linear combination with three of the biggest components, I’d argue, being ES SF/SA, PP%, and SV% (with the last one being represented in PDO as well, of course).
A) it’s possible, but somewhat unlikely I think. Sure, they played well at the right time to win two playoff rounds (which always has to be the case), but they also won a tonne of games in a row to get in (though of course that they had to must mean they weren’t that great to begin with, right?). watching edmonton-carolina battle it out for the cup notwithstanding, i cant believe im wasting all this time watching a tournament of mediocrity trudge along where the victors are not at least given a rebuttable presumption of superiority over the losers (trivia caps also notwithstanding - as an aside, how would the trivia caps analysis work in a best of 7 series. at that point, does shot ratio (trivia ability) overtake the randomness (dice rolls)?).
B) yes, they also got outshot fairly badly at ES as well (though roughly by the same amount as mtl, who i would also argue was better than a mere average team, for roughly the same reasons). phi was not a good team at ES, certainly. but with a fantastic PP and solid PK, and in a league where a disproportionate amount of time is spent in either of those situations, i think that makes for an above average, or what i deemed “pretty good” earlier, team.
C) their PP shooting percentage was retardedly good. so was mtl’s. rightly or wrongly, that stuff goes along in determining team quality beyond all the great ES that gets looked at.
Have any of y’all done research on how PDO numbers behave on special teams? Do we see similar regressions? Is sustaining an above or below average PDO more possible, less possible, or the same as ES?
i’ll let the spreadsheets confirm, but i’d have to imagine that you even see a stronger regression w/r/t to PP S%, just because the sample sizes are going to be smaller and you’re going to see some wacky high percentages at times. i’ll also ask, have we looked at the correlation between ES S% and PP S%? There’s gotta be one there, right, because the players (scoring forwards) driving the results at ES are also going to be doing the trick on the PP as well?
Hi With PDO..Where did you find ES (even strength shots For and against totals to do your c alculations?? I have looked everywhere and can’t find seperate shot totals for even strength and special teams thanks dan
come by virginia slims ultra reflection cigarettes
shoddy doral cigarette cheaply doral cigarette|cheap doral cigarettes
low-priced doral cigarettes shabby doral cigarettes
cheaply gpc cigarette reduced gpc cigarettes
tight gpc cigarettes budget-priced gpc cigarettes
virginia slims ultra