The Washington Capitals and their 313 Goals
by Hawerchuk on Apr 11, 2010 3:54 PM EDT
Email
PrintScoring 313 goals in the modern defensive era is a huge achievement. The Capitals have the 7th-highest goal total relative to the league average since the 1967 expansion:
| Team | Year | GP | W | L | T | GF/G | LGE | %+ | PLY |
| BOS | 1970-71 | 78 | 57 | 14 | 7 | 5.12 | 3.12 | 63.9 | L 4-3 in QF |
| MTL | 1976-77 | 80 | 60 | 8 | 12 | 4.84 | 3.32 | 45.6 | SC 4-0 |
| EDM | 1983-84 | 80 | 53 | 18 | 9 | 5.58 | 3.94 | 41.3 | SC 4-1 |
| PIT | 1995-96 | 82 | 46 | 27 | 9 | 4.41 | 3.14 | 40.5 | L 4-3 in SF |
| BOS | 1973-74 | 78 | 52 | 17 | 9 | 4.47 | 3.20 | 40.0 | L 4-2 in SCF |
| BOS | 1971-72 | 78 | 54 | 13 | 11 | 4.23 | 3.07 | 38.0 | SC 4-2 |
| WSH | 2009-10 | 82 | 43 | 15 | 24 | 3.82 | 2.77 | 38.0 | ? |
| EDM | 1982-83 | 80 | 47 | 21 | 12 | 5.30 | 3.86 | 37.1 | L 4-0 in SCF |
| MTL | 1974-75 | 80 | 47 | 14 | 19 | 4.68 | 3.43 | 36.5 | L 4-2 in SF |
| MTL | 1977-78 | 80 | 59 | 10 | 11 | 4.49 | 3.30 | 36.1 | SC 4-2 |
There are really only four teams that have been able to do anything like this: the early 70s Esposito-Orr Bruins; the late 70s Canadiens dynasty; the Gretzky Oilers; and the Jagr-Lemieux Penguins, who didn't typically hit this level of dominance. The first three teams were able to repeat this performance year-after-year because they were just that much better than the rest of the league offensively. But the Capitals have not shown this level of shooting ability:
5v5 Sh% 2007-08: 7.1%
5v5 Sh% 2008-09: 7.2%
5v5 Sh% 2009-10: 10.9%
In other words, while the stars aligned for the Bruins, Canadiens and Oilers in terms of them acquiring superstars and their opponents sucking, the same has simply not been true for the Caps. With largely the same offensive personnel as in previous seasons, Washington has simply gotten very lucky in the shooting percentage department. Or for those of you who don't believe in luck, the Capitals players have "made their shots" this season. Either way, it's no more likely to happen again next year than it was to have happened last year.
33 comments
|
1 recs |
Do you like this story?
Comments
The Caps have been lucky, but they also have Ovie, Backstrom, Semin, and Green. They also have Knuble knocking in Ovie/Backstrom shots.
There has also been a lot more net crashing this year, which brings shot percentages up a lot. Look at Knuble, with 19.2&! I think the offense can do the same thing next year, but maybe not the year after that because of the possible loss of Semin/Knuble.
I agree totally about the stupidity of the average person. Horrifyingly, nearly 50% of people are even stupider!
Oh yeah, just take a look at how many guys have hit a career high, which is obviously the level that most players sustain. If you think Mike Knuble, a 37-year-old guy, is likely to exceed his career shooting percentage by 25% next season, well…Let’s bet on that.
It is possible that Washington increased their shooting percentage by 20% because they really did figure out a way to permanently get better quality scoring chances…They would be the first team in decades.
Or, more likely, they “made their shots” this season and this season only.
I’m not betting on Knuble getting better, but I am betting on him not getting significantly worse. His game is about positioning and stick handling redirections, so that’s someone that doesn’t have a sharp drop off like speed does.
And where would I go to look at career shooting percentages of individual players?
I agree totally about the stupidity of the average person. Horrifyingly, nearly 50% of people are even stupider!
I’d like to see those Individual numbers as well. I would theorize that the caps as a team have raised there shooting percentage due in part to there younger players having matured and stepped up there play. Lets not forget that Washington is still a young team and players like Fehr, Flesichman, Carlson…..perhaps even Backstrom and Green….have yet to hit there prime and could conceivably get better and post similar shooting percentages in subsequent years.
If Popeye got stronger after eating canned Spinach....Imagine what canned Bacon could do for Ovie.......
I’m sure Mr. Hawerchuk would be willing to take a wager on it.
Editor of The Copper & Blue, and leader of The Cult Of Hartikainen.
I wouldn’t wager that the Caps hit 313 again, but I would wager that they tap 300, though I could see it being only 290 next year too. That’s still quite an anomaly as it’s 32 more than the #2 team Vancouver currently has. But I think Ovechkin Backstrom and Carlson will all get a few more next year. Green I think is tapped out, 20 goals as a defenseman is impressive enough, I don’t think he will ever hit 30 again.
I agree totally about the stupidity of the average person. Horrifyingly, nearly 50% of people are even stupider!
While I agree with your overall assessment and the notion that underpins it (that deviations from the norm in shooting percentage are unsustainable), I also think your analysis is slightly incorrect on one crucial point:
With largely the same offensive personnel as in previous seasons, Washington has simply gotten very lucky in the shooting percentage department.
The Caps were giving significant ice time the last two years to Viktor Kozlov and Sergei Fedorov. Over those two years, Kozlov put up 7.3% and 8.5%, while Fedorov put up 5.9 and 9.3. Chris Clark was getting a regular shift and shooting 6.8 and 6.5 in 08 and 09, respectively.
Put simply, the Caps’ shooting percentage was pulled down a long ways by a few regulars who couldn’t hit the ocean from a pier. (I know that, technically, some of those percentages are “above” the average, but most forwards should shoot much higher percentages, thereby balancing out the low percentages of the defense.)
While I agree that Fleischmann and Knuble won’t continue to score at this rate of efficiency and an argument can be made that Ovechkin, Semin, Fehr, Chimera, Bradley and Backstrom will all slow down), the drop-off will likely not be from 10.9% down to 7.2. I’d say that a shooting percentage somewhere in the 8 to 8.5% range would be likely, but without analyzing it player-by-player, I’m just pulling that out of my butt.
I'm not JTRS compliant.
Overall, 19 guys have played for the Caps in each of the last three years. They shot 9.8%, 10.2% and 11.4% over the last three years. A group of 19 players regresses very hard to their personal mean.
The rest of the roster shot 7.8%, 7.7% and 12.2%. That’s swapping Nylander, Kozlov and Fedorov for Knuble, Morrison and Chimera. It’s pretty likely that those guys regress to about 10%, which makes the team’s overall performance not significantly different from last year.
I agree that there will be regression, I just think that greater analysis needs to be done to project their potential shooting % next year on a case-by-case basis. While I agree that regression to the mean is inevitable, it’s tough to determine a mean for players with shorter track records so the “grab bag” result will look skewed. For example, I feel comfortable regressing Fleischmann down from 19+% down to somewhere in the 12-14% range, but Fehr? I’m not sure he can’t repeat this performance given more PP ice time.
Admittedly, that’s just noise in the overall average, but I wouldn’t bank on the Caps’ offensive output dropping to 08-09 levels – something between 08-09 and 09-10 is likely.
Again, I agree with the premise and the general conclusion, I just think that players regress to their personal means and that those means are harder to determine for younger players, which the Caps have a great deal of. Regressing an entire team to its mean, while it gives you a ballpark figure, will be off because of differences in personnel and player development.
I'm not JTRS compliant.
What’s interesting looking at it from a player-by-player basis is that, of their big scorers, Ovechkin, Semin, Green, Laich and Morrison are in-line, and often below, their recent averages. Knuble is slightly above, Fleischmann is above, Fehr is only in his 2nd season but is clearly punching above his weight. Backstrom is higher, but Backstrom is only 22 and his performance this season may be sustainable (who thinks Drew Doughty will go back to 27 points next year?). Rerunning the season, I would guess they score 20 fewer, but no less.
by Tom Awad on Apr 11, 2010 9:42 PM EDT reply actions 1 recs
I think one of the questions that has to be asked is “Have the Capital’s changed something about their offense/style that would explain the rise in 5v5 shooting percentage?” I think if we were to do an analysis of where they were shooting and scoring from, we would likely see that compared to years past, a much higher percentage of shots and goals occurred within a few feet of the crease. If a higher percentage of shots are occurring closer to the net, where shot percentages rise, then this might explain the difference.
If we look at the team’s roster changes, specifically the subtraction of Kozlov and Fedorov, and the addition of Knuble, I think we do see a specific change in play style that might partially explain the difference. While in years past we might occasionally see Laich or Fehr trying to camp in front of the net, the presence of Capital near the net is almost assured this year.
It might not be a lot, but I think it contributes to that %2 difference in shot percentage.
Either way, it’s no more likely to happen again next year than it was to have happened last year.
I have tremendous respect for this blog, and you are one of the few guys that seemingly nudges the paradigm of hockey statistics with every post… but this qoute shocked me. While the two clear and dominant factors that determine a team’s goals scored are shots taken and prior shooting%, no real hockey mind would tell you they stand alone. It almost seems as you’d like to wave your hands and poof away aspects of the game that aren’t quantifiable. There’s a difference between explaining as much of the game as you can with statistics, and actually claiming complete knowledge and predictive certainty through such a simplistic model. This short post resounded of mathematical hubris, borderlining on arrogance.
Player progression, coaching style/system, and line synergy/cohesion are all potential factors that aren’t accounted for in your model. The Capitals clearly had a great, above mathematically expected shooting year. I am not blind to the numbers, but a much more fair assessment of the situation is this:
What’s interesting looking at it from a player-by-player basis is that, of their big scorers, Ovechkin, Semin, Green, Laich and Morrison are in-line, and often below, their recent averages. Knuble is slightly above, Fleischmann is above, Fehr is only in his 2nd season but is clearly punching above his weight. Backstrom is higher, but Backstrom is only 22 and his performance this season may be sustainable (who thinks Drew Doughty will go back to 27 points next year?). Rerunning the season, I would guess they score 20 fewer, but no less.
by Bushwood Bushwhacker on Apr 12, 2010 10:59 AM EDT reply actions
The first quote came off as a negative, based on the previous analysis, but on the other hand it’s actually an even handed assessment. They were able to do it this year, they have the same chance of doing it next year.
Suspend Colin Campbell!
Either way, it’s no more likely to happen again next year than it was to have happened last year.
What this quote implies is: Last year, this year and next year, the Capitals have the same expected shooting percentage, as they are no more or less likely to have a great shooting year (like 09-10) during any of those three seasons.
Whereas I would disagree. There were other factors that increased their shooting % this year, in addition to being lucky. It was not 100% due to the random number generator coming up goal that led them to this point. Like Awad said, I think the EV goals for this Caps team is about 20 goals less than they scored, not 40-50 like Hawerchuck would lead you to believe from the even strength Sh% stat he posted.
(The Caps scored 213 5on5 goals, and shot 34% better than they did last year.)
by Bushwood Bushwhacker on Apr 12, 2010 11:47 AM EDT up reply actions
A lot of money has been lost betting against regression to the mean. When a team has any kind of extreme performance (good or bad), they bounce back towards the middle the next season. You and I may disagree on the degree to which they will get pulled toward the mean, but if I think it’s double what you think, that’s hardly “arrogance.”
It’s not your end result that I find contentious, it’s the smoothing over the myriad details on your way to it. The phrase “regressing to the mean” is overused, misused and misunderstood everyday (I’m not saying by you), read by pseudo sports mathematicians as “a force that pulls everyone towards mediocrity.” There are very real, often unquantifiable factors that augment expected player outputs, and nowhere do I see them discussed or even nodded to as a unexplainable or unknown factor.
The certainty of the Capitals being “very lucky,” the inherent confidence in an overly simplistic model, and the lack of acknowledgment of many of the actual, flesh and blood hockey elements of the game was what I perceived as arrogance, not the stance itself.
by Bushwood Bushwhacker on Apr 12, 2010 12:14 PM EDT up reply actions
And “flesh and blood hockey elements of the game” is an argument that is used to justify confirmation bias by countless people.
Mike Knuble is a great hockey player, but he’s not worth 3-5 wins more than Fedorov. (He’s getting paid like he’s worth 1.25 wins above replacement.) You can argue that the whole is greater than the sum of the parts, but I don’t find it a very compelling argument given that no coach has shown the ability to consistently assemble such a team.
What’s likely here is that the quality of the Caps scoring chances regresses, and the rate at which they make their shots regresses. I don’t think that’s a surprising statement.
What about something as simple as player progression? Were Fehr and Fleishmann and Backstrom supposed to have the same expected Sh% this year as last year? Do players get better or worse or do they get luckier or unluckier? Don’t these players at the very least have predictable career arcs where they were supposed to have shot better this year than last year?
by Bushwood Bushwhacker on Apr 12, 2010 12:29 PM EDT up reply actions
I don’t think there was any good reason to believe that Ovechkin, Semin, Backstrom, Laich, Fleishmann, Fehr, Green and Bradley would, as a group, increase their shooting percentage 14% this year after having the same sh% in 2007-08 and 2008-09. The more people you include in a group, the more likely the group stays where it is.
When a rookie shoots 10%, then shoots 12% in his second year, then 15% in his third year, what do you predict his fourth year Sh% to be? (Not a rhetorical question. I’m curious what your methodology is.)
Do you consider talent development as something that actually exists? At what point is a player not getting luckier, but actually getting better (and possibly luckier at the same time)? And a logical question that follows is how can you identify a player that is getting better but not seeing results on the scoresheet (unlucky)?
Shouldn’t a 5,6,7% (ie .5-1% shooting percentage) be a reasonable uptick of Sh% for a player who is improving and heading into the prime of his career? Leaving some of your 14% due to luck and some to actual development?
I’m guessing you have numbers on player’s improvement by age, etc, I feel like I’ve seen them (though maybe it was somewhere else.) I won’t make you regurgitate. If you want to link them I’ll come back later after I’ve read them. I’m not just arguing here like some people like do to, I’m just looking to find the best answers to the hardest questions…
by Bushwood Bushwhacker on Apr 12, 2010 1:37 PM EDT up reply actions
It looks like Semin, Green and Laich all saw their shooting % drop between last year.
Meanwhile Ovechkin, Backstrom, Fleischmann, Fehr, Bradley and Knuble’s all increased.
Taking as a baseline that most players are likely to shoot around their career averages, that would mean we could expect the following shooting percentage changes next year:
Ovechkin DOWN 1%
Backstrom DOWN 2%
Semin EVEN
Green EVEN
Fleischmann DOWN 6%
Fehr DOWN 4%
Knuble DOWN 4%
Bradley DOWN 2.5%
Now if we account for more subjective flesh and blood hockey changes, perhaps we can make the following readjustments:
Ovechkin – likely to keep hovering around that career %
Backstrom – improving his # of shots every year; if he keeps doing that he can offset decreased shot percentage
Semin – likely to keep hovering
Green – likely to keep hovering
Fleischmann – a unique situation; when he is a center he shoots less…if he becomes one full time his goal totals could significantly drop. TBD
Fehr – break-out season could lead to more ice time and thus more shots (he had too little of each this season), which like Backstrom could offset any drop in shooting percentage
Knuble – perhaps, as surmised, his increased % is due to having the ideal linemate (six-shooting Ovie) for his rebounding style? Perhaps he can remain healthy for more games and pot even more (unlikely given his age)?
Bradley – shot below his career % last year and above it this year, won’t account for more than a 1-2 goal change
So through this rudimentary excercise, I’d surmise the following: of the Caps’ ten goal scorers, the biggest drop in goal totals you can expect to see next season is from Flash with Fehr and Knuble coming up behind closely. Between the three of them that might add up to 10-20 fewer goals. I’d expect little to no significant change from the Young Guns and Bradley.
This all assumes a lot about health, and that the other non-accounted for players will stay relatively the same, and that someone won’t have an off, awful year. I think what I’m getting at though is whatever drop in scoring is to come next year won’t be so significant. The Caps are still an offensive powerhouse.
The keyboard is mightier.
by breed16 on Apr 12, 2010 2:00 PM EDT up reply actions 2 recs
Yea, I think a lot of that is right on the money, qualitatively. I would really like to see each individual’s even strength shooting % and how they project next year, because despite the Caps’ crazy good powerplay, that is really where the variance lies for this team, this season, as Hawerchuck pointed out in the original post. 10.9% as a team is pretty crazy.
And like I’m trying to argue, I’d say the expected ES Sh% if we re-ran this season’s shots, we’d see closer to 9% than to the 7.5% you might expect from the previous two years. Because, essentially, to predict the outcome of THIS year after this year has already happened, you have to take the actual outcome into account. Right?
To simplify, our rookie shoots 10%, 12% and 15% in his first three seasons. If you were to replay the shots from his third season, do you expect 11%, his career average before the year, or something between 12-13% his career % including the third year?
by Bushwood Bushwhacker on Apr 12, 2010 3:20 PM EDT up reply actions
Knowing nothing else? Probably 13% – individual shooting percentage is driven by PP time, so if we know that a guy is getting more and more PP time, we would expect a lot of that ‘gain’ to be real.
If we knew how old he was, how much PP/ES time he had, what his shooting percentages were in the minors and junior, we could make a better estimate.
So couldn’t you assume that if a group of young players 21-24, who are inherently more likely to improve than to decline, shot 10% last year, and 11% this year, that if you were to rerun this season’s shots, they would be more likely to shoot at 10.5% than at 10%? Again, I’d point to a graph of Sh% by age if I had one. I’m guessing the graph plateaus around 26-29, like it does for other production related stats. At some point these players have to improve to that plateau, and it’s not just a roll of the die that puts them there. They get better.
Basically, I am still having issues with your original statement that the Caps were just as likely to have had this incredible season last year as they had this year. I’d argue this exact group of players is going to play hockey even better next year, if they were kept in tact. Would they score less? Most likely, yes. Their Sh% this year was incredible, and as a team, unsustainable. We agree on that. But there is just more to what is going on than what I feel like you are presenting.
by Bushwood Bushwhacker on Apr 12, 2010 3:09 PM EDT up reply actions
The big driver behind shooting percentage is PP time. It’s not clear to me that EVS% changes that much as a player ages, but I will check it out in more detail.
Cool. Those numbers would be interesting.
by Bushwood Bushwhacker on Apr 12, 2010 4:21 PM EDT up reply actions
Shouldn’t the driver be ratio of powerplay time to even-strength time?
Only YOU can prevent idiots from commenting!
by Knee high to a duck on Apr 12, 2010 6:46 PM EDT up reply actions
I think you’re underestimating Knuble’s production and overestimating Fedorov’s. Knuble’s GVT this year is 12.7. I’ve done a fair amount of analysis that shows a linear relationship between the absolute value of 2.75 goals (either for, or against or some combo thereof) and a win. By that metric, Knuble’s a “4.6-win-versus threshold” player. I’m not sure where Fedorov would have placed last year, but I have a hard time imagining him anywhere above a 1.6 win player. I think Knuble’s pretty solidly at 3 wins above Fedorov, and I think he’s probably at least 4 wins above Kozlov, who he was actually replacing on the roster.
I'm not JTRS compliant.
A win is 6 goals. Knuble had good shooting luck this season as we all seem to admit, so his contribution was higher than usual. He was projected for a 6.6 GVT, which is roughly in line with his contract.
D'oh
A win is 6 goals.
My bad (that’s what I get for commenting after a bottle of wine). I should have made clear that I’m talking about “points above average,” which I roughly extrapolated into goals. An average team post-lockout will get around 91 points and will score as many goals as they give up. Any deviation from that, either by adding goals for or taking away goals against, has a pretty precise linear relationship to points at the rate of about 2.75 goals/point.
I'm not JTRS compliant.
Comments For This Post Are Closed
Secondary Sidebar
User Tools
FanPosts
Community blog posts and discussion.
