I kind of buried the post, but I wanted to do a follow up on my post arguing that the Oilers current hot streak (and implicitly, their somewhat respectable looking record) is a mirage, I took a look at some of the shootout numbers. Frequent Oilogosphere commenter AsiaOil has been the impetus for a couple of these posts and this one is no different. Here’s the comment that triggered this:
I bet you thought that the shootout string could not last at 5-1 right? Not every situation ends in a 50-50 split in the short to medium term. Have you ever considered that the Oilers are just simply good at the shootout? Hemmer, Gagner, Horcoff and Pisani are solid and Garon has been a wall this year. The percentages certainly favor the Oilers - but by how much? How much is luck? How much is skill? You don’t know and neither do I - that’s my point.
I’m bear baiting because the numbers boys are often just a bit too smug about their powers which are clearly quite limited given the basic statistical tools applied and the nature of the game. A lot of the individual analysis you guys do is great and very interesting - but far too simplistic when applied to complex, dynamic situations such as the future performance of an NHL hockey team over a season.
I’ll start by acknowledging that I think that the complexity of forecasting NHL performance when looking at a reasonably sized chunk of the immediate future is entirely overstated by AsiaOil - if you’re really paying attention and looking at the right areas, I suspect that you can get most of the things in the ballpark, most of the time. There’s always going to be randomness but it doesn’t overwhelm the skill, which can be discerned through the results that guys put up. Of course, certain things - injury and goaltender performance in particular - are a lot harder to predict. Those two factors can swing a team pretty wildly in one direction or another.
The reason I find AO’s bear baiting so baffling (and the reason I’ve written two separate posts on it within a week) is that I think that he’s picked the wrong time to lecture the numbers folk on how their analysis is “far too simplistic when applied to complex, dynamic situations such as the future performance of an NHL hockey team over a season”. Maybe I’m wrong and he’s right, I just don’t think that the Oilers particularly support his claim.
Anyone can surf over to behindthenet.ca and see the Oilers even-strength performance for themselves: 27th in ESSF/60, 21st in ESSA/60, 14th in ESGF/60, 7th in ES shooting percentage, 26th in ES save percentage and 27th in ESGD/60. But for an amazing year shooting percentage wise, they’d be absolutely horrible at ES. As it is, they stink. I don’t see how this makes his case.
On to the PP (I’m using 5v4 as a proxy here because that’s how Gabe has his site set up but the Oilers are 30th overall, so it’s not like the 4v3 and 5v3 numbers are going to bail out this bag of stink): let’s see, 28th in PPGF/60, 30th in PPSF/60 and 21st in PP shooting percentage. I get that Sheldon Souray and Joni Pitkanen being out has hurt but neither of these guys has done much on the PP when they’ve played and it’s not like the Oilers PP has been any better with those guys in the lineup either.
The PK has probably been the highlight of the Oilers season so far (again, 4v5 numbers here) and it’s thoroughly middle of the road. I suspect that a lot of the numbers guys would have said “Yeah, sounds about right” if told in August that the Oilers would be averagish on the PK and terrible on the PP and at ES.
The only reason that I can see for such bear baiting and general taunting of the numbers guys for being wrong lies in the shootout results, although he’s also written that there are “…just too many intangibles like injuries, chemistry and luck” to predict performance - incidentally, I think that the numbers guys tend to be the biggest boosters of luck out there - I can’t see how those would really bear on the shootout results and I think AO acknowledged that it was the shootout performance behind the Oilers stunning run to a three way tie for ninth with two teams that have games in hand on them in the post I’m citing. In fairness to the numbers guys, the position that AO was advancing was that the PP would be very good and that they’d be pretty decent at ES. The numbers guy consensus, to the extent that there was one, was a hell of a lot more on the mark. In fairness to AO, if I understood him correctly, he wasn’t saying that his view was any more valid than the one advanced by the pessimists, only that there was too much faith being placed in “…a bit of dodgy math on the back of napkin and a truckload of assumptions” and that he saw things that were positives. I’m sure that he’ll correct me if I’ve misrepresented his position.
As I assume that most people reading this are aware, the Oilers have taken part in 11 shootouts this year, winning 10 of them. I’m not so sure that the Oilers winning percentage in the shootout is the really amazing thing - as I’ll set out below, the number of shootouts in which they’ve played is the really phenomenal part. Their success here is really two pronged: first, the record and second, the number of games. Both bear some examination.
Assuming, for the sake of argument, that the Oilers are a 50% true talent team in the shootout, a 10-1 streak could be expected to occur about 5 times in a 1000 11 game streaks. The odds of a five game winning streak for such a team would be about 3 in 100. To answer the first question raised by AsiaOil’s comment, if I had turned my mind to the question of whether the Oilers 5-1 start in the shootout was likely to continue, I would have thought it unlikely unless I believed that the Oilers were about an .870 winning percentage team on true talent in the shootout, which I lean towards thinking isn’t true.
This raises the question: what would an .870 true talent winning percentage team look like in terms of their shooting percentage and save percentage? I’ve simplified things to take a look at that question; what I’ve done is run trials based on three shooters per team and a given team shooting percentage and save percentage. I suspect that this simplifies things too much but it illustrates my general point and seems to mesh relatively well with the results of the past five years. The x axis here is save percentage; the y is shooting percentage. If you find the most applicable set of save percentage and shooting percentage, it appears to be a good approximation of the winning - most teams seem to fall within +/- 1 win of their expected shooting percentage according to the chart. The orange area is the area in which a team’s shootout results would need to fall in order to expect the Oilers current winning percentage on the shootout.
From that, I’d take the position that, in and of itself, the Oilers shootout performance is pretty amazing. I won’t say that they’re outperforming this chart, because it doesn’t appear that they are. I’ve got some doubts as to whether this performance is indicative of the actual skill of the various Oilers who are involved in the shootout - the chart below sets out their results prior to this year and this year’s results - but I’m willing to give them the benefit of the doubt and assume that their results in the shootout this year were earned on the basis of skill. I don’t know that I believe that - I suspect that they’ve been awfully lucky (Hemsky’s goal against the Wild and Pisani’s winner against Vancouver are stuck in my mind) but I don’t know that there’s enough here to say that this is really luck for the Oil.
That, however, isn’t the end of the questions that arise out of Edmonton’s love affair with the shootout. The winning percentage by itself isn’t enough to have the Oilers in a three way tie for ninth - there’s been a simply incredible volume of shootouts. The sheer volume of games that they’ve played that have been tied through regulation and overtime is necessary to cover for their craptacular performance in games that don’t make it to the shootout - 6-15-1. We’re in a bit of a better position here because we can compare the Oilers performance against NHL history - there’s so little information on the shootout that it’s hard to tell whether those results are really indicative of skill or just chance.
As of the Vancouver game, 11 of the Oilers’ 34 games have been tied after 65 minutes, a rate of 32.3% of games. Over the course of a full season, this would be the third highest rate in NHL history. The Oilers rate is all the more amazing because of the relatively recent introduction of 4 on 4 overtime. The highest rate posted by a team since the NHL went that route is that of the 2003-04 Minnesota Wild, who ran up 20 ties - Edmonton is on pace for 26 games tied after overtime. If I can anticipate an argument here, I don’t think that the evidence would support the argument that there’s some skill in taking a game to a shootout - after correcting for the incidence of ties (which tend to increase as goals per game fall and parity increases), the correlation year over year in tie games by team is 0.02. The general theory amongst the people who write about sports through a lens of numbers is that if a result isn’t repeatable, it’s probably not due to a skill - that would suggest that while the Oilers performance in the shootout may or may not be due to their skill, the frequent opportunities to use those skills has its genesis in luck.
Really, I would argue that what we’ve seen to this point in the season is a confluence of two things: a remarkable rate of success in shootouts, which could be due to luck or skill, and an historically remarkable rate of games tied at the end of overtime. I don’t speak for any of the other statsarazzi but if asked at the start of year, I would have said that such a thing was possible but extremely unlikely. As an Oilers fan, I don’t take much solace from what’s happened so far this year because I believe it extremely unlikely to continue.
I’m of the view that all of the doom and gloom predictions for the season look justified to this point in time because even with the shootout performances and the frequency with which those performances have occurred, the Oilers are a team that’s tied for 9th in the conference. One of the teams with which they’re tied has four games in hand, the other has two. The Blackhawks sit two points back in 12th place with two games in hand, the Predators sit four points back in 13th place with four games in hand and the Coyotes are tied with the Predators while having three games in hand on the Oilers. If not for Edmonton’s current edge in games played over a lot of teams, this likely wouldn’t even be a topic for discussion because the rose tinted glasses set wouldn’t be drawing courage from standings that show the Oilers on the fringes of contention.
So really, in conclusion, if AO thinks that the Oilers near historic rate of games that have made it to the shootout and Oiler results in the shootout proves the guys who said that this team was nowhere near good enough in the summer wrong, well, it’s a free country. Even if it continues, they’re probably not a playoff team and the numbers would suggest that they’re unlikely to continue playing this many shootouts. I guess I’ve got nothing to really say if people want to draw solace from the fact that the Oilers have done so well in the shootout but I’d argue that those results don’t tell us all that much about the Oilers as a hockey team and that all of the numbers that do tell us something suggest that this team is a hell of a long way from a playoff appearance, let alone contention for the Stanley Cup. The application of simplistic analysis to complex, fluid dynamic situations that I engage in from time to time here may be utterly without worth but the 2007-08 Oilers aren’t the case on which to hang that conclusion. But for an astounding number of shootouts, something that looks pretty random to me, they look for all the world like a flat out bad hockey team.
Just like I said in August.
Your chart indicates that the Oilers’ historic performance has them bang on at .500 in the SO. They were .454 those first two years, so they underperformed slightly before and are overperforming slightly now. Enough to call the variations luck? As for the change in percentages, it’s possible that it’s streakiness, but it seems like it’s gone on too long to really be entirely due to luck.
While I can’t disagree that the freaky incidence of shootouts that’s turning this from a 6-17-11 (0.338) team pre-1999 to a 16-16-2 team now isn’t sustainable, I still think they’re playing better than a .338 team would, though until they learn to hold a damned lead in the third period, I don’t think I’ll have many supporters in that stance, especially since the numbers won’t give me anything until the Oilers smarten up in the third.
If you knock out the SO goals, they’re at +82 and -102 for the season, which should have them 3 or 4 games below .500 - 23 points in 34 games would likely be a bit unfair to them, something like 27-29 would probably be more in the ballpark.
Glad you stressed the issue of how improbable this many shootouts is. They’ve played 12 4v4 OT periods (which in itself is a ton), and 11 of them have been scoreless.
The illustration here is obvious; that’s like playing an entire game 4v4 and having the final score be 1-0. You could have the ‘04 Flames face the ‘07 Wild, each playing tighter than Tom Thumb’s ass, and that wouldn’t happen.
A lot of the individual analysis you guys do is great and very interesting - but far too simplistic when applied to complex, dynamic situations such as the future performance of an NHL hockey team over a season.
This drives me crazy. How terribly unimaginative. A complex system? Ooh, let’s just give up and crawl back to our caves.
Anyone know the number of games lost due to injury, btw? I agree with you Tyler, but they’ve had terrible luck with injuries, too.
Hi Tyler - actually you stated my position rather well and quite fairly - cheers. I would only add that I was also not overly optimistic at the start of the year either - predicting 8th-10th - but being quite open to the possibility of shit happening which few deemed likely or even possible. It was the certainty that filled the pre-season predictions of suckage that was well….well it smacked of hubris given the state of our tools.
Well 10-1 in shootouts in just over 30 games certainly qualifies as shit happens and nobody’s analysis spit that out. Who knows what else can/will happen over the the rest of the season - but that’s why they play the game. A lot of the time it’s not logical, it’s not fair and makes no sense - but if it did it would be a computer game. Shit happens is in fact quite common - every year certain teams inexplicably outperform while others underperform - you can essentially bank on it.
As for Andy’s comment - environmental complexity and the inability of our little human brains to understand it is reality my friend - the worst thing you can do when confronted with this situation is think you are smarter than you actually are. I read and appreciate Tyler’s and Vic’s stuff as much or more than anyone - but there are simply limits when it’s used as a predictive tool in something as complicated as a hockey season.
…..at least until the shit happens algorithm is perfected
Tyler maybe you or some of the other guys can think about the question I posed on HF - but the thread has degraded into a whose smarter then whom argument.
I wanted to see if there was some way to capture the value of the Pisani’s and Horcoff’s of the NHL. The guys whose play transcends their individual numbers and significantly impacts their linemates and overall team results. I’ve termed these guys synergy players - but none of the existing individual stats seem to capture their value at all - any ideas or is this simply beyond our grasp?
Andy, I don’t think AO is saying “give up”, he’s saying “the system you have isn’t perfect”.
It’s true - there’s a lot of room for improvement, and I think mc acknowledges that. Any “numbers guy” who doesn’t is an idiot. It’s valid to point and say “your system doesn’t account for x”, at which point the numbers guy can say “yes it does, here, look” or “oh, I hadn’t thought of that, let me try y.”
And besides, a mysterian position is also a perfectly valid one. It doesn’t mean “give up and crawl back into your caves,” it means “you’re trying to dividing by zero gives you a valid result, my friend, and that just doesn’t work.” I agree with you though, it’s too early to say if mc’s dividing by 0 to get his results.
Speaking as somebody who believes that some day we can have a true machine intelligence, I think that the problems the “numbers guys” are trying to solve are solvable. (Compared to modelling human intelligence or the weather, it seems that predicting the outcome of what happens when 40 guys are put in a pretty small area with a piece of rubber seems relatively trivial.) But I think our current models are fairly crude, particularly with respect to predicting a team’s future.
AO: I think you can account for Horcoff and Pisani by looking at what isn’t there as much as what is there. The Pisani-effect would be seen in things like ESGA while he’s on and off the ice, as well as the performance of the PK when he’s on it. I’m not sure how Horcoff’s performance isn’t showing up even in just the boxcar stats this year. Horcoff’s career shows a clear progression.
Matt: what makes it even more incredible is that the one GA in 4v4 OT was on the powerplay.
I wanted to see if there was some way to capture the value of the Pisani’s and Horcoff’s of the NHL. The guys whose play transcends their individual numbers and significantly impacts their linemates and overall team results. I’ve termed these guys synergy players - but none of the existing individual stats seem to capture their value at all - any ideas or is this simply beyond our grasp?
I would say that, if it’s real, it’s simply beyond our grasp. I’m not so sure that it’s real though - it seems just as likely to me that the explanation is that those guys are good hockey players who make their line better.
I’m not so sure that I’d proclaim Fernando as being able to carry Nilsson/Gagner yet either. As of right now, he’s ESSF 17 ESSA 16, which is a stunning change for Gagner, whose presence on the Horc line resulted in an ugly 7/29 shot ratio. 4 goals on those 17 shots…not going to continue. 1 goal against on those 16 shots…also not going to continue. I’d give it some more time before annointing Pisani as a synergist.
Wouldn’t a “synergy” player’s effect be covered by on/off ice comparisons for the team and his linemates? (Change in goal/shot/penalty/whatever differential) Or is that too simplistic?
Also, I realized my initial post was totally bogus, since the above charts don’t have every Oiler’s SO stats from the first two years (i.e. no Smyth, Sykora, Three-Headed Monster, etc.) Here’s the real numbers:
2005-06: S% = 35.8, SV% = 57.4, PW% ~= 0.43, W% = 0.438 (7-9)
2006-07: S% = 27.3, SV% = 66.7, PW% ~= 0.40, W% = 0.500 (3-3)
So they hit the nail on the head in ‘06, and outperformed in a small sample size in ‘07.
MacT in today’s Urinal:
“I’m just hoping we can win some games in regulation and not risk the numbers turning around,” he said.
A few things you have to remember when you’re looking at luck:
There are 30 teams, what might look like skill to 1 team, when fitted into the broader picture looks more like luck.
Take for example the chance of going 10-1 in the shoot-out (5/1000), however, the probability that NO TEAMS go 10-1 in the shootout is 86%, meaning there’s a 14% chance that 1 team goes 10-1 (not great, but not bad either).
Secondly, luck and skill are a part of almost all results. The real question is: which one is the dominant factor?
In the financial world, when analyzing the “results” of a business or company — you can look at the obvious:
- short-term results: Profit & Loss (aka Income Statement)
- mid-term results: Balance Sheet
However, these alone show simply a snapshot in time.
More savvy analysis looks at ratios and rates. They tell the forecaster:
- are things improving or worsening
- and at what rate will it do so
Short-term results can still be happy — but depending on how deep a financial hole was there — may or may not improve the long-term.
There are even ratios such as z-score which has a pretty good confidence in predicting success or failure of a venture within certain timeframes.
I’d draw the analogy that we can be happy with the short-term results of the Oilers, but that their fundamentals aren’t significantly improved. And unless they start to at least make some improvements there — they are in danger of letting the inverse of synergy or bad luck start to dominate the results.
I don’t think that there is any dichotomy in that. And AG’s clip from MacT is evidence that he senses it too.
What I wonder is how the return of Pitkanen, then Pisani and Souray and even Thoresen now, have changed the results for the Oilers.
What I mean is this - without these guys how many of these games would have been straight out losses? Likely a whack more. So what is needed now to turn the SO wins into regulation wins and more of the losses into OT/SO situations?
They will be adding Moreau. Will they trade for another player or two? I think so.
The other thing is that they are mediocre, if that, but outside of Detroit and Ottawa are there any good teams out there? If Detroit falls then the WC will be wide open, imo, even moreso then in past years.
Not that I think these guys are going anywhere.
Having said that I’m enjoying it. SO wins are better then no wins.
Interesting discussion, I find myself agreeing with ao and the “numbers guys” in turn. For sure 10-1 in shootouts qualifies as “shit happens” — or as I put it on lowetide site, “unfuckinsustainable” — but the Oil’s record of 12 regulation ties in 32 games may be less of a fluke.
Forget about OT v. SO for a moment, any games tied after 60 become three-point games and the winning percentage (or should I say points percentage)of the teams involved balloons from .500 to .750. It follows that the Oilers “should” have 6 bonus (a.k.a. “bogus”) points from this system, and they have 10. So that’s 6 points for beating the system by getting to overtime, and 4 extra points for beating the odds in the shootout.
In 2005-06 MacT took a .500 club to the playoffs by similarly beating the system. That year the Oil went 28-28 in regulation, and 13-13 in OT/SO. As a result the Oil finished 13 games “over .500″ (whatever that means) with 95 points, and slipped into the postseason instead of Vancouver, who had a superior record of 34-32 in regulation and 8-8 in OT. I needn’t mention (but will anyway) that this .500 team subsequently came within a game of the Stanley Cup.
So the Oil finished 2005-06 at 28-28-26 in regulation, with a remarkable 32% of games tied after 60. This year’s rate of 37% “ties” doesn’t seem quite so remarkable in that light. I think our coach “gets it”, and much as I hate the stupid system, my hat is off to him. The fact that the Oilers are giving away free points to the rest of the league matters far less than the high percentage of those points they are “earning” for themselves.
Bruce:
I think you are underestimating the 05/06 Oilers dramatically, Bruce. That team was only the underdog at home three times all season (twice vs DET, understandably, and once vs DAL). The Oilers will be underdogs at home 34 or 35 times this season (they’ve only been the bookie’s favourites once so far this season, but some weak teams are due).
The 05/06 Oilers were favourites at home right through the playoffs, except for against DET obviously, where they had their asses owned everywhere except the scoresheet.
I think that is where the differentiation happens here. Me and MC, Riv etc were telling everyone that the 05/06 Oilers were a good team, to stop panicking … because the Oilers were outplaying the other guys a lot, and if that’s the sort of thing you notice, then that’s what you see. Now we’re at the other end of the spectrum, and that’s not to just be awkward, it’s because they aren’t outplaying the bad guys very often at all.
If you only see the Ws and Ls, then you’re going to be riding the emotional wave like Sara Silverman off her meds. This because nothing will make sense, the only reasons that mesh will have to be pulled straight out of the ether. And fabricated from shit found while rooting through looking for patterns, without a toe close to the reality pool. This is just true.
This latter group aren’t the “stats guys” because they can’t be, it won’t be possible for them to rationalize the shit they imagine, because, and here’s the kicker “IT NEVER FUCKING HAPPENED”. Pucks go in off of guy’s knees, it is what it is, stuff happens. So there is no way for them to build a convincing argument from evidence simply because there is none there to support it. Inevitably they end up pulling goofy shit like “synergy” out of the air.
Sorry for the second double posting, this one’s deliberate due to piss-poor formatting in the above. I wish there was a trash can option on this site.
***
I think you are underestimating the 05/06 Oilers dramatically, Bruce.
Say what? I said, and the record proves, that in the regular season they were a pure .500 team that understood the value of getting to overtime.
The 05/06 Oilers were favourites at home right through the playoffs, except for against DET obviously, where they had their asses owned everywhere except the scoresheet.
… and, by extension, between the pipes. In the 2005-06 regular season the obvious problem was the substandard goaltending of Fricklin and Frakkanen, where the Oilers would routinely outshoot a team like Chicago 36-25 and get beaten by some guy named Michael Leighton. The problem was therefore identifiable and fixable, and Lowe identified and fixed it before the playoffs. But Oilers made the playoffs only by virtue of those free points (not to mention a shit streak not dissimilar to the current hot roll by one Mark Morrison).
Me and MC, Riv etc were telling everyone that the 05/06 Oilers were a good team, to stop panicking … because the Oilers were outplaying the other guys a lot, and if that’s the sort of thing you notice, then that’s what you see. Now we’re at the other end of the spectrum, and that’s not to just be awkward, it’s because they aren’t outplaying the bad guys very often at all.
… except between the pipes. Oilers’ goalies have been excellent over the last month and have stolen quite a number of points. On the season, they have allowed 1 goal in 59 minutes of OT, and in the shootout just 6 goals on 36 shots, which goes a long way to explaining why Oilers got 2 points instead of 1, 10 times out of 12.
To some extent it’s a mirage, at +82/-102 they clearly are not as good as the +249/-242 team of 2005-06 (shootout “goals” excluded from both). But most of that -20 differential happened early in the season, they have since righted the ship and have become pretty competitive — the four teams they “tied” last week are a combined 34 games over .500.
By hook and by crook they have pulled their way into some semblance of contention.
I for one am kind of enjoying that, even if my eyes tell me they’ve been lucky a lot of times. So I will choose the middle ground of not setting unrealistically high expectations for this young, developing team, while at the same time refusing to submit to doom and gloom that the raw numbers suggest.
If you only see the Ws and Ls, then you’re going to be riding the emotional wave like Sara Silverman off her meds.
Nope, I see Ws, Ls, and MacTs
mc, you’re a magician. Thanks for cleaning up my messes.
Everybody else, pls. ignore the lead-in paragraph to the post immediately above.
Vic - I usually respect your stuff but that reply was just plain snarky and pointless. With respect to the 06 Oilers - you and everyone lese painted them as barely a playoff team yet they almost won it all. Bit of revisionism going on there my friend - NOBODY saw that result coming and the few who did were painted by this particular group as naive fanboys in late 05. Also - your blessed odds makers were wrong a lot in 2006 as the Oilers were pretty mediocre at home losing 21 games - so what’s the point? In any case the vegas boys play the big picture and all the individual player and single team results is just noise in the overall result that they work to make money on.
Basically you are missing my point entirely - I’m not saying the Oilers are a good team - they are not by any metric conceivable. What I was saying back in September was that the team had enough talent that unexpected and unpredictable events could produce results better than predicted by last season’s box scores. I cannot understand how this simple idea can be so threatening to everyone’s world - but it was mocked mercilessly.
As for synergy - I am not applying it to team results at all - I’m simply interested in defining something which is difficult to measure. Get your calipers and slide rule out and tell my how that slow, little runt with no shot name Gretzky did what he did? Synergy is not some bit of wierd voodoo that requires sitting under crystals. Measuring and estimating synergistic effects is a primary research concern in my field of environmental management. One plus one does add up to three under certain conditions - IT DOES FUCKING HAPPEN - and it’s been proven scientifically for ages. So saying that synergistic effects are voodoo or “goofy shit” is just uninformed.
So my point in Sept is the same as now - that the Oilers are a talented enough team that good fortune could make a significant difference to their results - and so we should wait and see how the season turns out before assigning them a top 3 slot in the 08 draft. So far that has been pretty accurate. Now fortune may indeed turn her gaze elsewhere in the final 40 games - but that does not make what I said in Sept any less relevant or valid.
With respect to the 06 Oilers - you and everyone lese painted them as barely a playoff team yet they almost won it all. Bit of revisionism going on there my friend - NOBODY saw that result coming and the few who did were painted by this particular group as naive fanboys in late 05.
Look in the archives of this site AO - I’ve pointed out how false this criticism of yours is before. I can’t speak for anyone else but I was saying all along that this team was legitimate if they got some goaltending. Christ, some CBJ fans over at HF made a picture mocking me for saying that the Oilers needed only a goalie; there was some schmuck at CalPuck with the quote “Edmonton plus average goaltending is a Stanley Cup contender” in his sig line - that disappeared around mid-May.
…that the Oilers are a talented enough team that good fortune could make a significant difference to their results
To start with, I don’t think that this is what you were arguing initially at LT’s - if it was, I misunderstood you, as did everyone else. I took your argument to be that this Oilers team was good enough to contend for a playoff spot with a reasonable amount of luck and good breaks - I still don’t think that this is accurate and I’d argue that even now, they aren’t really in contention - you may remember I was running playoff odds back in 2005-06. If I ran those odds right now for the Oilers, I’d be surprised if they were over 10% and that’s with an incredible amount of luck in the form of bonus points through the first half.
I look at it this way. At the start of the year, there’s 15 possible places that the Oilers can finish in. Each of those places has a probability of occurring. With this year’s team, I was (and am) of the view that the probability of 1-8 is so slight that even with unbelievably extreme benefits from the SO (although Bruce makes a fine point worthy of further examination about going to OT that I’ll look at more closely), it wasn’t really worth considering. I still don’t really think that it is.
The numbers men usually average over a season, and don’t allow for a team that develops over time. The Oilers are developing. Developing chemistry, especially with all the changes in the summer. Developing players, a few of whom are just getting their first smell of the coffee. Look how the Pens developed last year. Or the Ducks and Sharks in 05-06 - was that predicted by numbers? There are too many variables with this group of players to pigeonhole them with numbers and as far as I can see, thats how it should be!!!
Oiler Mag: The Celtics turned over a bunch of their roster and haven’t lost 3 times this year.
As for the Asia argument, I’m at a loss to explain it. I pegged the Oil for 8-10 and then 10-12 after the injuries hit and that was based on all the injuries to their few NHL players.
How many forwards on the Oilers have played more than 200 NHL games and could reasonably expect to make a significant step forward this season from their established level of ability?
How would that number compare to the other 29 NHL teams? I don’t think we need bother to look it up.
AsiaOil - with all due respect, any clown with a spreadsheet could have summarized the amount of pointless snarky comments (PSC’s) from Vic in threads questioning statistical analysis and predicted with a high degree of certainty that one was on way….he has a PSC/60 comment number of 3.5, so by comment 19 he was really overdue. This might be explained simply as bad luck (power failure?), or poor linemates (Window’s Vista? Dial up?), but things should even out by the end of the thread:)
Asiaoil:
The message boards are rotten with people sniping at me for defending the 05/06 Oilers. Your point is flat out wrong.
Mostly, I just think that there is wonderful discussion on clutchness and chemistry on the fan boards, absolute tonnes of it. Go crazy talking about it over there, there is no reason to bring it to this part of the internet.
Granted, in this case MC79 seems to have run you a bit. The lesson is: Don’t poke the bear.
The Oilers are developing. Developing chemistry, especially with all the changes in the summer. Developing players, a few of whom are just getting their first smell of the coffee.
I’d have a far easier time believing this if the Oilers weren’t on a shooting percentage binge. Teams just do not have the shooting percentage that the current defenders are trying to paint as chemistry. That’s why I have an awfully hard time with this explanation.
Oilman:
Well, that’s just hurtful.
You’re a curious little bugger, Daryl.
Do you wager on sports at all?
Saying stuff like “before the numbers right themselves” is kinda bullshit. What are the odds that the oilers go 11-1 in the shootout the next time they enter it? The same as if they were to go 1-0 at the beginning of the season. Ignoring of course who they are playing, goalies, shooters and all the rest of it..
Its like computing the odds of getting a heads when flipping a coin after you have miraculously just got ten heads in a row? The same. Still 1/2.
The Oilers have no worse or better chance of winning the next shootout simply because they have a record of 10-1.
Vic…it’s Darren, can’t say that I haven’t been called Daryl before - but it stings as much every time….as for sports betting - I don’t touch it. I was in a football spread pool a few years ago and one weekend picked every winner on the weekend - 14 for 14, and got 1 for 14 against the spread…I know when the odds are not in my favor.
“So the Oil finished 2005-06 at 28-28-26 in regulation, with a remarkable 32% of games tied after 60. This year’s rate of 37% “ties” doesn’t seem quite so remarkable in that light. I think our coach “gets it”, and much as I hate the stupid system, my hat is off to him. The fact that the Oilers are giving away free points to the rest of the league matters far less than the high percentage of those points they are “earning” for themselves.”
You know, I don’t often chirp in here because you guys are so far beyond me in your analysis. But hells bells, I swear the above comment has been ringing true to me for a while. If pretty much anybody can calculate the increased odds of winning by getting to OT, why not MacT and his crew? And if his strategy is such, wouldn’t it make sense that he spends an inordinate amount of practice time working on the SO to again - increase his odds? Hmmm…
I was at the game tonight. Turco has this cool psych move in the SO where he turns sideways to the shooter and jumps square when the shooter gets to the blue line. I swear it was freaking out all our guys. It was like they were saying “WTF?” as they were bearing in. Did they show this move on TV?
You know, I don’t often chirp in here because you guys are so far beyond me in your analysis.
Glad to have brought the discussion down to your level, David.
I too was at the game tonight — well, the last 14:35 plus OT plus the shootout, don’t ask, it’s a long story — and saw Turco’s schtick. He’s always done that, and he’s always got into shooter’s heads too. Without looking it up, he’s probably won more career shootouts than any goalie.
So we lose the shootout tonight, and for the second time in 9 days we lose to the fucking Stars, and yet this Oiler fan wasn’t unhappy at all. Because we didn’t actually lose. Another game, another tie, another point in the bank with a chance to go for the second. So that’s 5 ties in a row, and 3 of the 5 times Oilers got the second point, so that’s 8 points in 5 games and we’re on the move up the standings. Without … beating … anyone … but such is the fucked-up-ness of Gary Bettman’s NHL where there are no more tie results, but plenty of tie games. It just takes a crafty coach like Jacques Lemaire or Craig MacTavish to figure out how to count the cards, and guess what. There’s no more ties, except every fucking game is a fucking tie.
Thanks, Gary.
David, no, it wasn’t apparent on TV.
I think goalies are on to Gagner, he needs a new trick. And Northlands^WRexall needs new ice. At least commentators have finally realized that it’s not “the best ice in the NHL!” any more.
Say you and I are at a bar and it’s midway through the third period tied 1-1. You say “Oilers are 10-1 in shootouts, if they can only get through the OT” and I say “their luck has to run out sometime, I think it’s tonight” are the odds 50/50?
If so, then why do we know in our hearts that when Jeff Fassero starts a season 8-2 despite a career 51-55 record that is the time to deal him in our Roto league?
Please explain.
LT….would you have dealt Brady Anderson early in 1996?
Lowetide said: … are the odds 50/50.
I’d stay away from any bet on the shootout. Hockey results are repeatable for obvious reasons, a lot happens in a hockey game, just a matter of separating the player’s results from each other. And if you’ve ever listened to Roger Neilson or Bob Johnson, and are looking at the things that matter, the spread of the results are almost exactly as they should be as well.
I had a quick look at shootout results just now and it’s madass. I just cut and paste the shootout results for the last three seasons from NHL.com into a spreadsheet, and ran this we macro for generating results from coin flippers.
Sub coinflip2()
For k = 34 To 133
For j = 2 To 31
Randomize
flips = Cells(j, 29)
heads = 0
weighting = 1 - Cells(j, 31)
For p = 1 To flips
If Rnd >= weighting Then heads = heads + 1
Next p
Cells(j, k) = heads
Next j
Next k
End Sub
If you assume that the weighting is like a coin (.5) the spread of results is a mess. If you use the 05/06 SO win% for the team to predict the 06/07 season it brings the top 7 or 8 teams in line with expectation, but the rest get even worse.
There are just far two many teams that absolutely suck at the shootout in a given year, over and above luck, and the repeatablity is poor. This in all two and a half shootout seasons so far, same pattern each year. There just aren’t nearly enough teams that are ‘a bit better than expected’ and far too many that are ‘pure crap’. Go figure.
I think it’s like most baseball stats, so you can grab any five teams and predict the average of their results for next year based on the last three. And look good doing it. But the spread will be way wider than
luck alone allows. Teams will inexplicably soar or fall into the tank.
Obviously practical issues as well, like a coach who doesn’t like his chances in the shootout is going to have his team going balls out in OT.
My guess is that it’s not like during a game, there’s too much time to think for both goalie and shooter. Someone with a psychology background might be able to make some dents in predicting these, I dunno. Or maybe digging deeper as MC suggests, into individual players stats, though I seriously doubt that will make a fig of difference. Granted, it may, I’ve only invested 30 minutes here. But at this first blush it seems so far off the mark I don’t think that it can be rationalized.
My point: Never bet on a shootout, Lain, not unless you know something that I don’t. Just hope for the best and take them for what they are, an entertaining exhibition of skill. Because it just may be that Turco turning sideways in the crease, or a myriad of other things, is having a material effect here.
My point: Never bet on a shootout, Lain, not unless you know something that I don’t. Just hope for the best and take them for what they are, an entertaining exhibition of skill. Because it just may be that Turco turning sideways in the crease, or a myriad of other things, is having a material effect here.
Wouldn’t the fact that it’s not a complete coin flip make it a better bet? Maybe I’ve missed the point (probably?), but if there are things about the Oilers or Stars that are making them win more shootouts than chance would allow, doesn’t that suggest that they are better bets in the shootout?
I still can’t believe that people are harping on the ‘06 team. There were a bunch of guys who kept repeating how that team was just about outchancing everyone on any given night.
All that team needed was netminding.
On a final note, MC’s musings about an unsustainable shooting pct have certainly rang true in the last two games.
Just two goals combined for the Oilers but a tonne of chances that are all of a sudden not finding twine.
Yet somehow the team remains competitive, so it’s not just the hot shooting that is driving the recent run. That’s five straight ties against solid competition (three games against the top two teams in the West, all five against teams in the top six in the conference for points percentage.) The shooting percentage may be unsustainable and the shootout success almost certain so, but I’m seeing nothing on the ice that suggests this team is an also-ran. Twenty games ago, yes. Now, not so much. And while games in hand remains a bit of a deficit, thanks in large part to 13 regulation ties the Oil have climbed right out of the points hole they were in.
Ummm….so right now we are 10-2 in the shootouts and are projected to play 28 by the end of the season. 28 SO games really does seem like a ridiculous amount considering that the league avg has been around 10 with the most by one team at 18 (achieved by NJD last season).
But for argument’s sake, let’s say that we really are gonna have the chance to strut our (supposedly) potent stuff in 16 more shootouts… I’ve been wondering, what’s the probability that we’d have a 10-2 stretch somewhere along the line?
So just for fun, I did an experiment to try to figure it out.
The idea is very straightforward: I assume a certain SO winning percentage iterated over 10,000 seasons, with each season having 28 games decided in SO’s. I then examine how many of those seasons contain at least one 10-2 stretch.
This is then repeated for various SO winning percentages.
Note that the SO wins & losses were simulated using MC’s idea of simply combining some assumed shooting% and goalie save%. However, in this experiment, this is really not necessary - just the SO winning% assumption would have been sufficient. But I’ve included those numbers for completeness’ sake.
Anyway, for what it’s worth, here are some results:
SO Winning% = 0.679 (assume: Shoot% = 55%, Save% = 0.600)
Probability of having a 10-2 stretch: 60.3%
SO Winning% = 0.623 (assume: Shoot% = 50%, Save% = 0.600)
Probability of having a 10-2 stretch: 41.3%
SO Winning% = 0.587 (assume: Shoot% = 47%, Save% = 0.600)
Probability of having a 10-2 stretch: 30.3%
SO Winning% = 0.564 (assume: Shoot% = 45%, Save% = 0.600)
Probability of having a 10-2 stretch: 23.8%
SO Winning% = 0.501 (assume: Shoot% = 40%, Save% = 0.600)
Probability of having a 10-2 stretch: 11.0%
SO Winning% = 0.437 (assume: Shoot% = 35%, Save% = 0.600)
Probability of having a 10-2 stretch: 4.0%
ooops…what I actually meant is the probability of having a stretch of 12 SO games with at least 10 wins
dubya:
Sorry for not being clear, I was looking at the chances of teams beating the odds. So in a 30 team league you would expect 3 teams to have a shootout win% better than 90% of coin flippers. And 3 teams better than 80% but less than 90% … etc, etc, down to 3 teams with less than 10%. Did it this way because different teams see different numbers of shootouts.
Generally speaking, MC is absolutely right with his general assertion, as far as I can tell. It’s mostly just luck. In fact if you apply a bit of reason (using SO win% from the previous season for the coin weighting, instead of 50%) it actually has the opposite effect that you would expect. If you want to pick which five teams are going to see their fortunes change in the shootout in the back half of the season … the brief history of the shootout is telling us that the best way to do that is by selecting DAL and then pulling the other four teams out of your ass.
And there will likely be far more teams that completely go in the tank at the shootout, far more than the coin flippers will. This seems inevitable, if this brief history continues to repeat at least. And there is probably a rational way of predicting which teams will falter, but I can’t see it. Just seems random, not in a predictable ‘roll of the dice” way, but in an erratic “hand of God” way.
Yet somehow the team remains competitive, so it’s not just the hot shooting that is driving the recent run.
Are they really that competitive though?
Right now we can say with conviction that the Oilers are ahead of the Kings. Every other WC team behind them has a better GD and enough games in hand to pass them. After all, the Oilers are dead last in their division and that shows no sign of changing IMHO.
I think you may be onto something with your theory about MacT and OT.
It’s fool’s gold stuff though. The Oilers play 32 times against NW division opponents and cannot afford to give away free points to them on a regular basis. Then they have 40 games against the rest of the WC and probably 25 of those are important to the playoff race. (ie. not against the Red Wings or Sharks) A team that’s serious about making the playoffs will want to take two and give away zero in nearly half of those games at least to have a decent chance. Lastly there are the 12 games against the EC teams, where the freebie OT point doesn’t matter either.
So the Oilers probably have 27 games where they can truly afford to give away the free point with little or no impact on their playoff aspirations.
If this is MacT’s strategy, I highly doubt it’s a winning one for a playoff spot. You could run through all the possible combinations and permutations, but 30+% of games going to OT means you’re bleeding a pretty decent chunk of points inside the conference to teams you’re competing with for playoff spots.
It’s even more likely to be cosmetic, when you’ve only managed to take +2 points in 6 out of 35 games.
Ummm….so right now we are 10-2 in the shootouts and are projected to play 28 by the end of the season.
One bet you can take safely to the bank is that the Oilers will not participate in any 28 shootouts. They may well continue to milk tight games into overtime and it is barely possible they could play 28 three-point games, but there is no way those overtimes will be scoreless at the rate they are now.
At this point the Oil have reached OT on a league-leading 13 occasions. More surprising is that only one goal has been scored in those 13 OTs, 64 minutes of 4-on-4 hockey played at a 1-0 clip? Now that’s unreal.
At this point of the season only one other team in the league has played so many as 10 overtime games (Anaheim). So let’s summarize how far out of whack the Oilers are with the norm:
League 501 GP (through tonight)
90 regulation ties = 18%
34 overtime decisions = 38% of OT games, 7% of all games
56 shootouts = 62% of OT games, 11% of all games
Oilers 35 GP
13 regulation ties = 37%
1 overtime decision = 8% of OT games, 3% of all games
12 shootouts = 92% of OT games, 34% of all games
So …
1) The Oilers reach way more OTs than anybody (double the league average),
2) are way below the norm for OT decisions (one-fifth to one-half the league average depending on how you measure), therefore
3) the % of games reaching shootouts is absolutely lights out (triple the league average).
And that’s before considering the unlikelihood of a 10-2 record in those shootouts.
It’s unsustainable. They will likely lead the league in total shootouts and shootout wins based on the huge early lead, but I wouldn’t count on them extending that margin. It’s more important that they continue to milk games like last night’s down the stretch and into OT, where they can take their chances. Eventually, some Oiler might even get the idea that they should try to score in overtime, where after 64 minutes the redhot shooting percentage that mc talks about up top stands at a very cold 0.0%. And that too is unsustainable.
What a strange admixture of extremes. I think much of it can be written off to small number statistics, but to a lesser extent it’s strategy. As I mentioned above, MacT has worked this system to his advantage before and has a good idea how to maximize his return. Lucky or not, the coach deserves a little credit for crafting a .500 points percentage from a -20 team. The Oil have survived what easily could have been an early season disaster. No doubt they have to get better, but there are promising signs in that direction as well. Maybe not enough to convince the unbelievers, but enough to make things interesting.
If this is MacT’s strategy, I highly doubt it’s a winning one for a playoff spot. You could run through all the possible combinations and permutations, but 30+% of games going to OT means you’re bleeding a pretty decent chunk of points inside the conference to teams you’re competing with for playoff spots.
In answer to your first statement, RQ, I have detailed above how the ‘05-06 Oilers made the playoffs with 41 wins (including just 28 in regulation) over a Vancouver team with 42 and 34. Under Bettman’s dumbass system regulation wins are overrated (or if you prefer, undervalued) — they just get you the same two points you can still get in OT. Sure they’re nice, but the reward for scoring a tie-breaking goal late in regulation is far less than the penalty for allowing one. Assuming OT as a coin flip, a late-in-regulation GWGF is worth 0.5 points, and a late GWGA costs 1.5 points and is a disaster.
As for points bled within the conference, it’s both true and relatively inconsequential. In one sense it’s hard to quantify: who actually gets the third point? The team that wins the mini-game after both split the points of the “tie”? Or the loser who gets one “loser point” instead of the loser’s traditional zero? You can make the argument both ways.
I’ll make it this way. Assuming an even split of those points, such as the Oilers 13-13 OT/SO record in ‘05-06, the Oilers are getting half of the extra points from their games, and the other half are being split among 14 teams in their conference and therefore being heavily diluted in value. Free points are only lost to one team, but are gained against ALL teams. And as you say, against the other conference playing to OT is a free space on the bingo card. Benefits far exceed costs. The system is heavily weighted in favour of the team that plays lots of OT.
Case in point: Phoenix, like Edmonton, is officially 16 wins, 16 regulation losses. Both stand at .500, the Oilers are three points ahead due to three “other” losses, whereas Phoenix hasn’t lost any games in extra time; of course they have three games in hand.
Now in regulation time Phoenix is a respectable 14-16-2, while the Oilers stand at an abysmal 6-16-13. Phoneix is a perfect 2-0 in overtime, the Oilers a less than perfect 10-3. But it doesn’t matter, a win is a win, two points. And every bonus point Edmonton bleeds to a team like Vancouver — like the one they blew in the dying minutes Saturday — hurts Phoenix just as much as it hurts Edmonton!! The Oilers got their two points anyway, so their incompetence in blowing the late lead was pretty much unpunished. Sure it helped Vancouver, but that hurt the other 14 teams in the conference, including 13 innocent bystanders. Every three-point game has the same result, a free point generated from the ether that passes for Gary Bettman’s brain that helps the competing teams and hurts 13 to 28 innocent bystanders. And when it’s in the interests of both teams to nurse a tie game to OT, how can that not affect the competitive integrity of the game? The dying minutes of a game like last night’s where each team propects their point and “creates” a third is effectively collusion. This system is seriously broken.
How often have you waited for the results of a non-Oiler game simply hoping that one of the teams wins in regulation, you don’t even care who. And as soon as it reaches OT you turn the TV off in disgust, you still don’t care who wins but you’re pissed that neither team lost.
So it’s in the best interests of all teams to play a lot of overtimes. Oilers have been more successful than other teams in achieiving this dubious honour in two of the last three seasons. It helped them in ‘05-06, and it’s helping them now.
In regular time and even in O.T. to a degree you can expect a certain level of performance based on past performances. In the shootout you are dealing with too small of a sample size - 3 shots +/-. A goalie’s SV% is generated over a much larger sample than a shootout, and a player S% is as well. Add to that the diversity of shooters involved from night to night I don’t believe there is enough data to predict any outcome in a shootout.
The great equalizer in the NHL has been the 3 point game, and even more-so the shoot out, both of which allow inferior teams to gather points through boring, sustainable play (Trapping, Locking, Cycling) against superior teams. Sam Gagner has what 3 game winners? but statistically he is just a misery in R.T..
As much as I doubt the coaching I have to agree with Bruce. The Oilers are currently exploiting a system which rewards mediocrity, much like the Minnesota Wild have in the recent past.
Let’s face it, the essence of this argument is that when the Oilers lose, it’s because they’re supposed to; if they win, it’s a fluke.
For everyone who sees the stats and instantly writes the Oilers off, there are the 2007 Arizona Diamondbacks, who won their division while spitting in the face of one of sabermetrics most hallowed theorems.
http://www.hardballtimes.com/main/article/no-mirage-in-arizona/
There’s a long way to go in the NHL season, but, so far, when the Oilers lose, they lose big. When they win, they win by the narrowest of margins. And they rarely rout anyone.
Just like the D’backs did in ‘07
bill needle says:
Let’s face it, the essence of this argument is that when the Oilers lose, it’s because they’re supposed to; if they win, it’s a fluke.
That’s just nonsense.
Personally, I expect the Oilers to win if they outshoot and outchance the opposition. If they don’t win the scoring chance battle, then I expect them to lose. They generally get outchanced, but they’ve had some pretty decent performances against some good teams as well. They’ve probably been a little unlucky in regulation so far - this team isn’t very good, but 6 wins in RT after 35 games is just ridiculous.
As far as the DBacks go, the crux of that argument is that Melvin manipulated personnel usage to yield those results. If you think there is something in common here for the Oilers, then you should be able to find some evidence. It should show up in the 1st, 2nd and 3rd period GD as well as in the TOI and matchups for the younger players. Go ahead and prove it if you like, but my suspiscion is that there’s a whole lot more random chance contributing to the Oilers’ 35 games than to the D-Backs’ 162 game performance.
Interesting link, BN. The bipolar bullpen makes perfect sense, assuming the manager knows what he’s doing. At least baseball doesn’t have a system where extra-inning games have winners but no losers. Imagine what that would do to the GB column!
Reviewing the Oilers early season performance, they are a predictably ugly 2-6 in games decided by three goals or more. Looking more closely, in the first 15 games they were 0-6 in such games. Over the last 20 games they are 2-0, a pair of back-to-back four-goal wins over Anaheim which are outside the form chart altogether. All the rest of their recent games have been close (give or take that ugly 4-2 loss in Denver).
To recap:
First 15: 5-10-0; 30 GF, 51 GA (SO 3-0)
Next 20: 11-6-3; 53 GF, 52 GA (SO 7-2)
In regulation:
First 15: 2-10-3; 30 GF, 51 GA
Next 20: 4-6-10; 53 GF, 51 GA
I don’t see anything there which doesn’t suggest the Oilers have righted the ship after a horrible start. Indeed, the positive ratio of real goals for and against in the last 20 suggests they have been slightly unlucky to register a mark of 4-6 in games decided in regulation. (All those wasted goals against the Ducks!) But of course that’s just half the games, they have tied the other 10, racking up a record of 7-0-3 in those games for 17 points.
The key is to avoid regulation losses, something Oilers weren’t doing in October … they weren’t even close. (Or when they were, 78 decided to make a line change, something that did not endear him to his coach.)
Last year the 15 teams with the fewest regulation losses (30 or fewer) ALL made the playoffs. Interestingly, those exact same 15 teams were the top 15 defensive clubs in the league, a remarkable correspondence which flies in the face of the supposed move to offence in the “New NHL”. I am absolutely certain the points system is at the root of it; play it tight, get it to OT and you’ll be fine. There was no such correspondence, not even close, for top offensive teams or teams with a lot of regulation wins, because they are not commensurately rewarded.
Oilers remain a long way from being a top defensive club, but the regulation GAA has improved from 3.40 after 15 games to 2.91, and in OT it’s an impressive 0.94. They’ve allowed some late tying goals (Detroit, Vancouver) that haven’t hurt them thanks to the shootout, but for the most part they have avoided giving up that late winner that winds up in the (first) L column. They still have issues in flow of play: I’m sure MC or Vic will point out how they were -50 in SoG in the first 15 games and -125 in the 20 games since, and maybe that’s a recipe for disaster. But the goaltending has been solid (again, after a rough start, esp. by Roli), the scorers opportunistic, and the coach savvy enough to beat the system. Maybe it’s all a mirage, but here we are, in the hunt and going in the right direction.
The issue here is that the Oilers are usually getting outplayed and outchanced, even at home. And sooner or later that comes home to bite you in the ass.
If you’re not one to notice scoring chances, then you are in the vast majority, Bob Johnson is clucking his tongue in heaven, and you’re probably more optimistic about this Oilers team than I am. And now would probably be a bad time to start looking at that.
I mean if the scoring system was as originally proposed, 3 points for a regulation win and 2 for an OT/SO win, one point for an OT/SO loss … then arithmetic tells us that the Oilers would be in 29th place overall, no? But of course they wouldn’t really, because the incentive structure for the coaching staff would have changed too.
Right now a lot of teams, maybe even most, see their rate of scoring chances drop a lot if they are clinging to a one goal lead, the chances of the other team scoring drop as well, just not quite as much. Moreau calls it “playing to the score”, fans don’t like it, but it works.
It means that the leading team is a bit less likely to get the regulation win, also less likely to get a regulation loss, and more likely to get a regulation tie.
Bruce keeps pointing this out with the Oilers, but it’s true for lots of teams, the Oilers have been in a lot of situations where it has ended up tied at regulation so far, so it’s more noticeable I think. So you end up with way too many ties, and some fairly joyless third periods unless you’re a big fan of one of the teams involved.
Nice link to the baseball article, Bill, sensibly written, unlike most of the baseball stuff I’ve read. A little while ago Cosh linked to an article on the Cardinals and Yankees “beating the pythagorean winning percentage” that was tragic.
I mean if the late 90s Sabres, who weren’t a very good team overall, but had Hasek in his prime … if they had pulled up a terrible goalie from the ECHL as Dominick’s backup. Started him in half the games. And if the ECHLer had managed to keep the Sabres in the game, tied or better, after two periods, then Hasek was sent in to play the third. Well the Sabres would win a lot of close games, and lose a lot of blowouts.
Would they be “beating their goal differential”? Absolutely, kicking it’s ass. But what does that even mean?
Would they be exceeding expectations in W’s and L’s? Not if you were aware of the context. Or at least an equal chance of this being a yes or no, depending on the how the bounces went.
I mean if the scoring system was as originally proposed, 3 points for a regulation win and 2 for an OT/SO win, one point for an OT/SO loss … then arithmetic tells us that the Oilers would be in 29th place overall, no?
Yes they would. e.g. Phoenix would have 8 additional points for their regulation wins.
I have been a staunch proponent of the system that I call the “three-point must” since about five minutes after they came up with the loser point in 1999. Split the points 3 and 0 or 2 and 1, but split the same 3 points, don’t create new ones out of nothing. I’m not sure I was aware that it was the system that was “originally proposed” — when the bonus point system was announced as a fait accompli it took me that same five minutes to figure out the balanced solution all on my own — but it sure as hell is the only one that makes sense from a competitive balance perspective. Why the NHL didn’t adopt it escapes me entirely.
It’s all well and good that my team happens to be taking advantage of the unbalanced mess that Gary Bettman and Brian Burke inflicted on the game, but nobody would be happier than me if they fixed it to reward teams who win in regulation, instead of just punishing the teams who lose in 60 minutes. That’s a pretty powerful disincentive for resolving games.
I’d just like to point out a few things:
1) Oilers have played much of the year without many of their top defensive players (Pisani, Moreau, Pitkanen, Souray, Tarnstrom, and now Torres) and have had to lean heavily on their kids. I think MacT and co. deserve some credit in getting this rag-tag bunch to Christmas at or near a .500 record. If on day 1 of the season you told me that the Oil would be plagued with injuries again this season, I would not have put money on them being within striking distance of a playoff spot in late December. Yet, here they are…
2) According to Jeff Sagarin at USA Today, guess which team has played the toughest schedule to date? That’s right, your Edmonton Oilers:
http://www.usatoday.com/sports/sagarin/nhl0708.htm?loc=interstitialskip
They’ve played:
16 of 32 vs NW Division (5-10-1)
5 of 8 vs Det, StL (2-3-0)
6 of 12 vs SJ, Ana, Dal (4-0-2)
6 of 20 vs Phx, LA, CBJ, Chi, Nas (4-2-0)
2 of 10 vs the EC (1-1-0)
I think the Oilers can destroy the SW Division and go .500 vs. the New York area teams for a record vs. the EC of (5-2-1)
Have to beat the WC bottom feeders (8-4-2)
9 games left vs. WC elite (3-5-1)
That puts us at 32-27-7 with 16 games vs. our Divisional rivals. Supposing parity means that 91 pts will be good enough to get you to the dance, Oilers need 20 pts out of those 16 games (9-5-2).
I think the Lowe needs to make a couple of trades, a few guys need to get/stay healthy, and a couple of other teams will have to go through some injury troubles of their own, but if all the stars align, these buggers might just squeek into 8th spot.
Bruce,
Yeah, that would be fairer, and would liven up some third periods. But it would hurt parity, or the illusion of it, by removing a lot of the random luck involved in the standings.
I agree with you. But as an Oiler fan, this is the worst possible time for you to be making this argument.
I think that making the nets bigger, or at least the top part bigger, would help make hockey better too. But again, our friend Parity would take another shot in the nuts. I wish that the NHLPA had negotiated more control over rule changes into the CBA, as it is, it’s left entirely in the hands of the owners, and parity is good for business. In the long term entertaining hockey is good for business too, but I don’t expect the governors to vote in this change, do you?
“We’re playing everybody even. I think if we keep going, if we get two points here, one point there, we’ll rise in the standings. Once they start believing that they’re a good team, then good things usually happen.”
– Bruce Boudreau,
“rookie” head coach,
Washinton Capitals
Sound familiar? It’s really not all that diffficult to figure out for anyone with a passing knowledge of game theory. As the season wears along watch for an increasing percentage of three-point games on a league-wide basis. In the first 246 games (20% of the sked) there were 39 OT games (16% of all games), in the second 246 there were 49 (20%). As the season wears on and standings budgets get tighter, watch more and more teams go for the free lunch.
I don’t have the exact numbers anymore, but last year the percentage of three-point games were similarly “low” around this time, went through the 20% level around the All-Star games, and despite an ever-growing divisor were still flying right up the slope to the end of the season, finishing at 22.8%. Meaning NHL games were worth an average of 2.23 points, and the mean winning percentage was .557.
This year those numbers are down a little (so far), 2.18 and .544. By way of comparison, there have been 2.37 points awarded per Oilers game, and the combined winning percentage of the two teams is .593. Given that the Oilers are always one of the two teams, that high joint winning percentage is a good thing.
Where MacT is way ahead of the curve is in understanding the value of three-point games right from the beginning of the season, none of this slow build-up that the league as a whole seems to do. So far the bonus points have kept our nostrils near the water line. And, as Boudreau said, “once they start believing that they’re a good team …”
Sorry if I keep hammering on the same point — either it’s obvious or it’s subtle, I’m not really sure which — but it’s telling that no matter what sort of statistical or analytical lens I use on this issue the answers keep coming back the same. From a competitive balance perspective the system is broken, and affects the in-game strategy and the integrity of competition. From a GAAP perspective, the system is inbloodysane.
So let’s try the big picture lens. From a business perspective, the old saying “we’re with you win or tie” applies in an increasingly results-driven marketplace to both teams in an OT situation; the deliberately-broken system encourages close games and standings parity. They either don’t understand or don’t care that the system has a negative effect on offence and quality of play in regulation time. By rewarding the two teams that make it to overtime with an extra point to play for, it’s possible to have your cake and eat it to, you can’t lose — you just need a warped imagination and/or a cold corporate asshole in a position of influence to wish it so.
I agree with you.
Well what do you know, Vic. I agree with you too. Is this a first? The only difference between our “big picture” statements is you said Parity with a capital P, I will defer to your wisdom in this case.
Now as for bigger nets …
In the long term entertaining hockey is good for business too, but I don’t expect the governors to vote in this change, do you?
No.
But as an Oiler fan, this is the worst possible time for you to be making this argument.
I don’t care. 1) I’m a lowly fan and the NHL has long since proven they don’t listen to us. 2) They can’t change the rules ’til next year at the earliest. 3) Much as I am a proud Oiler fan, I’m an even bigger hockey fan, and will always advocate what I see as the good of the game. And Gary Bettman’s sleight-of-hand with games, standings, statistics and records ain’t it. 4) See 1)
[dismount soapbox]
Vic, comment #53: “Right now a lot of teams, maybe even most, see their rate of scoring chances drop a lot if they are clinging to a one goal lead, the chances of the other team scoring drop as well, just not quite as much. Moreau calls it “playing to the score”, fans don’t like it, but it works.”
If a team was significantly better than the norm at playing to the score, they should outperform their pythagorean expectation.
I basically copied what Bill James did in this paper (pdf) for all NHL teams from 1968-69 to 2003-04, using Alan Ryder’s Pythagenpuck formula. Instead of comparing outperforming teams to other single teams with similar GF/GA, I grouped outperformers and underperformers by how many points they were better/worse than expected and examined how they did the following season. I don’t have the numbers right in front of me (they’re on my other machine), but I can report that the biggest outperformers/underperformers were only off their expectation by a miniscule amount the following season. The ability to outperform the pyth. expectation in the NHL is not repeatable.
If some teams are better at playing to the score (New Jersey? Minnesota?) they still happen to continue outscoring opponents at a rate consistent with the pyth. expectation.
I’ve been meaning to do a post on this for awhile but I know it will be a lot of work.
Vic - I defended the 05/06 Oilers as much as you (and I certainly dont deny that you did) but you also need to fess up that almost everyone who frequents this blog thought Conkanen was a very good idea as well (me included). No one mentioned SCF and Oilers in the same breath without being riduculed as clueless - and there’s a big difference between that and saying they were a decent team that needed a goalie.
Yes I can see perfectly well that MC is going at me regularly - and I am also astute enough to know that he would not spend so much time try to undercut my postion if it didnt threaten his universe. Stats in hockey are useful - but thinking that last season’s boxscores can allow you to predict the future of a particular team in a particular season is vain in the extreme - and also a good way to lose money if you choose.
I am also astute enough to know that he would not spend so much time try to undercut my postion if it didnt threaten his universe
Yes, you’re right up there with Terry Jones and Jim Matheson. In all fairness, I wouldn’t classify your stuff as the kind of low hanging fruit that those guys put out but to say that you threaten my universe is maybe putting it a bit strongly. Shit, you’re having a hard time even getting the facts right in this thread:
I defended the 05/06 Oilers as much as you (and I certainly dont deny that you did) but you also need to fess up that almost everyone who frequents this blog thought Conkanen was a very good idea as well (me included).
Again, what’s your point? I don’t think that any of us claim infallibility. I don’t like to talk about “us” but god knows I was off the Conkannen bandwagon by January or so, judging by the posts here.
No one mentioned SCF and Oilers in the same breath without being riduculed as clueless
As I mentioned in the first AO thread on this here site, I wrote the following on CBC’s website before the playoffs started:
If the scoring comes back to where it was most of the season, there’s a strong argument that the Oilers are the second-best team in the conference.
Unfortunately, there’s an even better argument that the Red Wings are the best team in the conference by a sizable margin. The failure of the scoring down the stretch means that the Oilers get the only team against whom they aren’t arguably even odds to win the series. In a short series, though, anything can happen. It’s the 25th anniversary of the greatest upset in Oilers history, when they crushed a Montreal Canadiens team that was one year removed from a streak of four consecutive Stanley Cups.
The difference between those two teams during the regular season was comparable to the difference between these two. The Oilers are huge underdogs but they’re good enough to make this happen. They’ll need some bounces but if they get them in the first round, they’re as good a shot to go the Cup finals as anyone in the West.
I don’t know what the hell else you can ask for.
…also a good way to lose money if you choose.
I’d be more than happy to put my methods up against yours, but as far as I can tell, yours boil down to “Shit happens, so let’s just expect the best.” For what it’s worth, since I’ve started really paying attention to the numbers and working through the playoff series with my methodology, I’m something like 22-8 or 23-7 at calling the playoff series.
That doesn’t prove anything - could be a hot streak (although I’d assume the guys who think the Oilers are for real would be tipping their cap to me) - and I don’t think I’m smart enough to beat the house over the long run. That said, in a market without transaction costs, I’m comfortable saying I wouldn’t lose money.