Thursday, July 20, 2006
On Powerplay
"Opinions are great—don't get me wrong. They're great for starting research projects. Then you go study and see if you can prove the opinion or not. But when placing multi—million dollar bets on future outcomes, opinions are wholly unsatisfactory. Opinions as conversation starters are fine. Opinions as conclusions are very bad. I started research projects to discern the objective “why.” I wanted to know why certain teams won and why other teams lost; why certain drafts produced big stars and others didn't. This was the naïve question at work."
--Paul Depodesta,
The Genesis, Implementation, and Management of New Systems
The Genesis, Implementation, and Management of New Systems
Continue reading "On Powerplay"...
Over the next little while, I am going to look at the Edmonton Oilers powerplay during the Stanley Cup Finals. The complete post is enormously large (surprise!). I started out with an opinion on the Oilers powerplay in the Finals, but had no idea if the evidence would support it. I have seen the numbers, the only numbers I was able to scrounge up, and now have more of a clue on how right and wrong I was. But I have no idea concerning the full validity or value of these numbers, or my thoughts. At this point, I am still the naive questioner searching for answers. Hopefully, these posts will bear some fruit. They took me an enormous amount of time to do, so I hope they are also of some value to others, even if the value is in how poorly the examination was performed.
During the Stanley Cup Finals between the Carolina Hurricanes and the Edmonton Oilers, there was a great deal of debate and discussion about the Oilers powerplay. Personally, I had been frustrated by the Oilers powerplay all season long, as it seemed entirely predicated on one, and only one, strategy: get the puck to the point. The strategy worked fairly well through the first three rounds, and the Oilers were able to score timely powerplay goals. In fact, they scored a powerplay goal in 13 of their first 17 playoff games.
| vs. Detroit | ||||||||
| vs. San Jose | ||||||||
| vs. Anaheim |
The powerplay had stalled in the Finals, however, as Carolina adjusted to what was, at least to my mind, the Oilers singular powerplay strategy. I was of course furious about this lack of diverse planning, as I had been all year long. Other, more moderate voices, however, suggested that the powerplay was getting better, and would come through to make a difference in the end. I didn't believe them. In between Game 5 and Game 6, I contacted Tyler and Avi, who I both knew to be statistical nutbars. An opinion was brewing in my head, and I wanted to be able to validate it. Here's how it all broke down, in present tense:
• I am intrigued by the assertion that the Oilers powerplay is getting "closer". I disagree.
• Other than Hemsky's goal in Game 5, when was the last time someone other than a player on the point took a shot on the Oilers powerplay?
• I would bet that over 80% of the Oilers shots on the powerplay throughout this series, if not throughout the entire playoffs, have been taken from the defensive side of the hashmarks.
• Based on what I know about hockey and basketball, but admittedly without much data to back me up, I want to identify those shots as being low percentage shots due to variables like distance and traffic.
• I also want to identify that a high volume of shots taken from the point also lowers their effectiveness, despite the quality of the shot-taker, due to their predictabilty.
• It also seems to me that the strategy is a waste of the talent within those players north of the hashmarks, but that probably demands a type of analysis that I am unqualified to perform.
• I want to know if anyone keeps a running tab of shots in the NHL, like they do in the NBA. I am sure stats companies do, but is any of it public? Like this:
• The NBA shot chart allows me to see the shots broken down by quarter, by distance, by location, and by shooter. It also tells me which shots were made, and which shots were missed.
• What is guiding my thought process here is indeed the NBA. If a team took 80% of its shots from beyond the 3-point arc, with two defenders always in their face, they would be crucified by players, coaches, managers, reporters, fans, and even casual observers. And they would be crucified, even if they had Michael Jordan and Larry Bird taking those shots. They would be attacked even more if they had a solid supporting cast in higher percentage shooting areas that weren't taking the easier shots. Essentially, I think this is what the Oilers are doing, even though not even Chris Pronger is as good as Jordan or Bird.
• I am also aware that the NBA and NHL are two very different games, and the analogy may therefore be tenuous at best. But it is the best way I could think of to describe it.
• I suppose that at the end of the day, the success of a powerplay can be broken down into sheer results. We got one more goal in Game 5 than we did in the previous three games, so I guess it did, indeed, get better. Maybe my question really should be, "Is a poor coaching strategy inhibiting the powerplay from being even better than it currently is or becoming?"
These were the thoughts, questions and opinions running around in my head. I knew already that the guys at Hockey Analytics had done some work on shot quality, showing that indeed distance does matter. I also knew that Tyler had done some work in relation to shot quality and goaltending, but that it was different from the answers I was trying to find. Tyler has now done some more work, focusing on the powerplay, and more powerplay, but again it doesn't really cover the things I want to know. It's probably best that he didn't cover them actually, as he knew I was working on this, and I'd be forced to kill him if he posted on it before me.
Anyway, I was seemingly on my own, and would have to figure out the answers to my questions myself. So I did what any sane person looking for information on the NHL would do. I typed in www.NHL.com. And that is where my problems began.
"My chief aim in this work has been thoroughness..."
--Immanuel Kant, The Critique of Pure Reason
So what exactly was I looking for? Well, to be precise, I wanted to know where shots were taken from, their distance, and who was taking them. For every game of the Stanley Cup Finals. Once I had those, I would create a series of tables, and analyze the results. I went to NHL.com, and couldn't find a shot chart. I went to TSN, SportsNet, and Yahoo! Sports as well. Nothing. The best I could find were the play-by-play reports on NHL.com. The play-by-play reports for each game gave me some valuable information, including the time of the powerplay shot, the shooter, the type of shot, and the distance of the shot. But there were many other problems with the play-by-play reports, such as:
• They identified three kinds of powerplay shots--shots that made it through, shots that were blocked, and shots that were missed. But they didn't provide the same information for each one of these shot kinds.
• They did not tell me exactly where the shot was taken from for any of the shot kinds. They only provided distances.
• For missed powerplay shots, they didn't tell you either the distance or location.
• For blocked shots, they didn't provide the shooter, the distance, or the location.
All of these things created problems. Only providing shot distances was inadequate, because it didn't give you a full picture on the quality of those shots. Obviously no number tells you the complete story, but it would at least help to know whether a shot taken "14 feet" away is being taken from directly in front of the net, or from along the side boards. Missed powerplay shots didn't have a distance or a location listed with them, which makes them almost completely useless to look at in relation to shot quality, and therefore powerplay efficiency. And it simply blew my mind that blocked shots didn't have a distance, location, or even a shooter listed with them. It listed the blocker of the shot, but that is all. So, even though I had my suspicions that Kevyn Adams was blocking a fair share of Chris Pronger's shots from the point, because he was listed as blocking quite a few shots and I knew he played on Pronger's side on the powerplay, I couldn't in fact prove it. Exasperating, to say the least.*
I went ahead with my plans anyway, knowing full well that I might not be able to find the answers I was looking for, but hoping that I could at least gain something else from the experience (I should also note that my plan after failing to find the shot charts was to create some shot charts myself, but the further I moved along the more convinced I became of the impossibility of such a venture). The very next thing I did, after finding the play-by-play reports, was try and determine the dimensions of the NHL hockey rink. I wanted to see if the distances gave me any idea of where the shots were coming from. I eventually found up-to-date dimensions on the NHL Officials page, and a bit of an explanation for them in the NHL Rule Book.
Looking at the map, I found that there is 75 feet between the backboards and the blueline. I discovered that the distance from the goal line straight out to the slot is 20 feet. The distance between the blueline and the goal line is 64 feet, so I started with the assumption that a decent shot is anything under half that distance, 32 feet or less. Having that information, as well as the incomplete information from the play-by-play reports, I began inputting the data into Excel spreadsheets.
I was almost done inputting the data for Games 1 to 5, and beginning to see some mildly interesting things, when I got an email from Tyler. He told me that the distances the NHL provides in the play-by-play reports are the subject of some controversy. Apparently Vic insisted that they are measured from the backboards, rather than from the goal line. Tyler disagreed, noting that it made no sense to calculate them in that manner. I hadn't even thought of it myself, completely assuming that they were being calculated from the goal line. Yet I knew there was a way I could verify how the shot distance was being calculated. In Period 3 of Game 1 of the Finals, Ales Hemsky had scored a beautiful goal on the powerplay. He had done so by dancing around a couple of players, and then deking out Cam Ward on the goal line to score. If Tyler and I were correct, the goal would be marked at a distance near or around 0 feet. If Vic was correct, the goal distance would be identified at around 11 feet. I took a peak at my spreadsheet...and saw that Vic was correct. Hemsky's goal was listed at 12 feet.
I took the night off after that one. While in hindsight I understand how the NHL would think to tabulate the distance in this manner--they might think that they would be forced into noting that Wayne Gretzky scored a ton of points from a negative distance--it also seems counter-intuitive. If the point is to score a goal on a net, couldn't and shouldn't the distance be marked as the space between said net and the player, regardless of where he is on the ice? Doesn't a combination of a distance from the player to the goal, as well as a shot chart to provide locational context, solve the problem in a simple manner? Of course the real problem was that I couldn't even find an explanation on how the NHL calculates shot distances. I still don't know for sure. All I have to go on is the numbers they do provide, and my memory of that one goal. Like with the rest of my statistical hunting for this project, I found the whole thing preposterous.
Once I got back at it, I had to redo some of my numbers, based on the fact that I now had to declare anything under 37.5 feet to be a decent shot. It changed my findings, in a dramatic manner. Shots in the 32-37.5 foot range, which I had initially marked down as poor shots, now made the quality cut. I countered this by adding a new range of quality shots, that being the distance from the backboards to the face-off dot in the offensive zone. 31 feet and less. This creates a problem at least once, but you'll see that later.
Near the completion of entering the redone information for Games 1-5, I discovered, thanks to Avi, shot charts. Ahhh. At last. Shot charts. From CBS Sportsline. Initially I was ecstatic, because they looked similar to the NBA charts.
I could see where the shots were taken from, see who took them, and see all it by period or grand total. If I moved my mouse over the X's and O's, I could even see what zone of the net the shot was taken on. My plan was to simply cross the data I had already found with these shot charts. I still wouldn't have all the information about missed shots and blocked shots, but I would at least have a clearer idea on the shots that got through to Cam Ward. I thought I might even be able to expand my scope, and see where Carolina were taking their powerplay shots from, as a point of comparison.
That plan didn't last for very long. It took me about an hour, but I eventually realized the shot charts on CBS didn't square with the official play-by-play reports of the NHL. For one thing, the shot times didn't align. It wasn't just that CBS kept track of the time by counting down, like European hockey clocks, rather than up, like the NHL clock. The times were just off, often noticeably off. For another, the CBS chart was giving credit to players for shots that the NHL did not give. The total number of shots were the same, but the shooters themselves differed. I was foiled yet again. If I had come upon the CBS site first, I probably never would have been aware of the discrepancies, and gone about my business in a happy but ultimately erroneous manner. But I had come across the official-play-by-play results of the NHL first, and was therefore back to square one. The irony wasn't lost on me, either, that the existence of incomplete, scattered, mostly useless, but official NHL statistics had ruined my chances of getting the information I needed from a more complete, valuable, but non-official source at CBS.
Fed up with it all, I stopped doing the work. I wouldn't pick it up again for a few weeks. By that time, the Oilers had lost the Stanley Cup.
*As I went to post this tonight, I decided to take another look at the stats on NHL.com. I took a look again at the offensive supersheets and the shift charts. I was suddenly terrified that the information had been there for me all along, and I had just missed it. The offensive supersheets list blocked shots and missed shots by who shot them, which is at least an upgrade from the play-by-play reports for the blocked shots. But you can't tell if the shots were taken on the powerplay or at even strength, and you don't know distance or location. You also don't know what period they were blocked or missed in. Even cross-referencing them with the shift charts won't work, because the players taking the missed and blocked shots on the powerplay are probably playing at the same time. I may be wrong, but I think the full information is still missing. To be honest, I hope it is now. If it is there for me to see, I'm going to have to re-write the remaining three posts. I think I'd rather die than have to do that.
"Pure mathematics is, in its way, the poetry of logical ideas. One seeks the most general ideas of operation which will bring together in simple, logical and unified form the largest possible circle of formal relations."
--Albert Einstein, New York Times, on the death of Emmy Noether
Legend
• Red Font Indicates Canes Player Blocking Powerplay Shot
• Blue Font Indicates Oilers Powerplay Goal
• Purple Font Indicates Missed Powerplay Shot
• Shots Blocked And Missed Do Not Indicate Where They Were Shot From
• Blocked Shots Do Not Indicate Who Shot, But Who Blocked
• 37.5 Feet Is Half The Distance From The Blueline To The End Board
• 31 Feet Is The Distance From The End Boards To The Line On Which The End Zone Faceoff Circles Are Placed
• I Do Not Differentiate Between a 5 on 4 and a 5 on 3 Powerplay
• I Count The Goalpost As A Shot, Not As A Missed Shot Like The NHL
• * Indicates My Own Determination Of Shot Distance
| 1st Period | ||||
| Shot | ||||
| Missed Shot | ||||
| Shot | ||||
| Blocked Shot | ||||
| Missed Shot | ||||
| Blocked Shot | ||||
| Shot | ||||
| Blocked Shot | ||||
| Blocked Shot | ||||
| 2nd Period | ||||
| 3rd Period | ||||
| Blocked Shot | ||||
| Shot | ||||
| Shot | ||||
| Shot | ||||
| Shot | ||||
| Goal | ||||
| Summary | ||||
| Shots Missed | ||||
| Shots Blocked | ||||
| Shots Through | ||||
| Total Shots On Net | ||||
| Shots Through From <37.5 Feet | ||||
| Shots Through From >37.5 Feet | ||||
| Shots Through From <31 Feet | ||||
| Powerplay Success/Game | ||||
| Powerplay Success/% | ||||
| Total Powerplay Time | ||||
| Game 1 Play-By-Play | ||||
| Game 1 Scoresheet | ||||
| Game 1 GameCenter |
| 1st Period | ||||
| Blocked Shot | ||||
| Blocked Shot | ||||
| Blocked Shot | ||||
| Blocked Shot | ||||
| Blocked Shot | ||||
| Missed Shot | ||||
| Missed Shot | ||||
| Shot | ||||
| Blocked Shot | ||||
| Missed Shot | ||||
| 2nd Period | ||||
| Shot | ||||
| Shot | ||||
| Missed Shot | ||||
| 3rd Period | ||||
| Missed Shot | ||||
| Shot | ||||
| Missed Shot | ||||
| Missed Shot | ||||
| Summary | ||||
| Shots Missed | ||||
| Shots Blocked | ||||
| Shots Through | ||||
| Total Shots On Net | ||||
| Shots Through From <37.5 Feet | ||||
| Shots Through From >37.5 Feet | ||||
| Shots Through From <31 Feet | ||||
| Powerplay Success/Game | ||||
| Powerplay Success/% | ||||
| Total Powerplay Time | ||||
| Game 2 Play-By-Play | ||||
| Game 2 Scoresheet | ||||
| Game 2 GameCenter |
| 1st Period | ||||
| Shot | ||||
| Shot | ||||
| Missed Shot | ||||
| Blocked Shot | ||||
| Missed Shot | ||||
| Missed Shot | ||||
| Blocked Shot | ||||
| Blocked Shot | ||||
| Blocked Shot | ||||
| Missed Shot | ||||
| 2nd Period | ||||
| Shot | ||||
| Blocked Shot | ||||
| Missed Shot | ||||
| Blocked Shot | ||||
| 3rd Period | ||||
| Shot | ||||
| Missed Shot | ||||
| Shot | ||||
| Missed Shot | ||||
| Summary | ||||
| Shots Missed | ||||
| Shots Blocked | ||||
| Shots Through | ||||
| Total Shots On Net | ||||
| Shots Through From <37.5 Feet | ||||
| Shots Through From >37.5 Feet | ||||
| Shots Through From <31 Feet | ||||
| Powerplay Success/Game | ||||
| Powerplay Success/% | ||||
| Total Powerplay Time | ||||
| Game 3 Play-By-Play | ||||
| Game 3 Scoresheet | ||||
| Game 3 GameCenter |
| 1st Period | ||||
| Blocked Shot | ||||
| Shot | ||||
| Missed Shot | ||||
| Shot | ||||
| Shot | ||||
| Shot | ||||
| Shot | ||||
| Blocked Shot | ||||
| Blocked Shot | ||||
| 2nd Period | ||||
| Missed Shot | ||||
| Missed Shot | ||||
| Shot | ||||
| Blocked Shot | ||||
| 3rd Period | ||||
| Summary | ||||
| Shots Missed | ||||
| Shots Blocked | ||||
| Shots Through | ||||
| Total Shots On Net | ||||
| Shots Through From <37.5 Feet | ||||
| Shots Through From >37.5 Feet | ||||
| Shots Through From <31 Feet | ||||
| Powerplay Success/Game | ||||
| Powerplay Success/% | ||||
| Total Powerplay Time | ||||
| Game 4 Play-By-Play | ||||
| Game 4 Scoresheet | ||||
| Game 4 GameCenter |
| 1st Period | ||||
| Shot | ||||
| Blocked Shot | ||||
| Missed Shot | ||||
| Blocked Shot | ||||
| Goal | ||||
| Goalpost | ||||
| 2nd Period | ||||
| Shot | ||||
| Missed Shot | ||||
| Shot | ||||
| 3rd Period | ||||
| Blocked Shot | ||||
| Blocked Shot | ||||
| Shot | ||||
| Blocked Shot | ||||
| Missed Shot | ||||
| Shot | ||||
| Blocked Shot | ||||
| Missed Shot | ||||
| Blocked Shot | ||||
| Summary | ||||
| Shots Missed | ||||
| Shots Blocked | ||||
| Shots Through | ||||
| Total Shots On Net | ||||
| Shots Through From <37.5 Feet | ||||
| Shots Through From >37.5 Feet | ||||
| Shots Through From <31 Feet | ||||
| Powerplay Success/Game | ||||
| Powerplay Success/% | ||||
| Total Powerplay Time | ||||
| Game 5 Play-By-Play | ||||
| Game 5 Scoresheet | ||||
| Game 5 GameCenter |
| 1st Period | ||||
| Shot | ||||
| Shot | ||||
| Blocked Shot | ||||
| Blocked Shot | ||||
| Shot | ||||
| Shot | ||||
| 2nd Period | ||||
| Shot | ||||
| Goal | ||||
| Missed Shot | ||||
| Shot | ||||
| Shot | ||||
| 3rd Period | ||||
| Goal | ||||
| Shot | ||||
| Blocked Shot | ||||
| Shot | ||||
| Shot | ||||
| Shot | ||||
| Goal | ||||
| Summary | ||||
| Shots Missed | ||||
| Shots Blocked | ||||
| Shots Through | ||||
| Total Shots On Net | ||||
| Shots Through From <37.5 Feet | ||||
| Shots Through From >37.5 Feet | ||||
| Shots Through From <31 Feet | ||||
| Powerplay Success/Game | ||||
| Powerplay Success/% | ||||
| Total Powerplay Time | ||||
| Game 6 Play-By-Play | ||||
| Game 6 Scoresheet | ||||
| Game 6 Game Center |
| 1st Period | ||||
| Shot | ||||
| Blocked Shot | ||||
| Blocked Shot | ||||
| Missed Shot | ||||
| 2nd Period | ||||
| Shot | ||||
| Blocked Shot | ||||
| Missed Shot | ||||
| Shot | ||||
| Shot | ||||
| 3rd Period | ||||
| Blocked Shot | ||||
| Missed Shot | ||||
| Blocked Shot | ||||
| Blocked Shot | ||||
| Summary | ||||
| Shots Missed | ||||
| Shots Blocked | ||||
| Shots Through | ||||
| Total Shots On Net | ||||
| Shots Through From <37.5 Feet | ||||
| Shots Through From >37.5 Feet | ||||
| Shots Through From <31 Feet | ||||
| Powerplay Success/Game | ||||
| Powerplay Success/% | ||||
| Total Powerplay Time | ||||
| Game 7 Play By Play | ||||
| Game 7 Scoresheet | ||||
| Game 7 Game Center |
"Who says there's an unemployment problem in this country? Just take the five percent unemployed and give them a baseball stat to follow."
Well, there you have it. All of my numbers for Games 1-7 of the Stanley Cup Finals. To recap, here are some of my more interesting findings:
| Powerplay Shots | ||||
| Shots Missed | ||||
| Shots Blocked | ||||
| Shots Through | ||||
| Total Shots On Net | ||||
| Powerplay Shots Through | ||||
| Shots Through | ||||
| Shots Through From <37.5 Feet | ||||
| Shots Through From >37.5 Feet | ||||
| Shots Through From <31 Feet | ||||
| *15 Of These Shots Occurred In Games 1 & 6 | ||||
| **14 Of These Shots Occurred In Games 1 & 6 | ||||
| Powerplay Success/Game | ||||
| Powerplay Success/% | ||||
| Total Powerplay Time |
• 57% of Edmonton's powerplay shots in the Finals didn't even make it on the net.
• 52% of the shots that actually got through to the net were from less than 37.5 feet.
• 48% of the shots that actually got through to the net were from more than 37.5 feet.
• 44% of the shots that actually got through to the net were from less than 31 feet.
• Excluding Games 1 and 6, 38% of the shots that actually got through to the net were from less than 37.5 feet.
• Excluding Games 1 and 6, 62% of the shots that actually got through to the net were from more than 37.5 feet.
• Excluding Games 1 and 6, 27% of the shots that actually got through to the net were from less than 31 feet.
• One of the through shots in Game 6--by Ethan Moreau--was marked at a distance of 6 feet. That means it was taken from behind the net.
| Bergeron | |||||
| Hemsky | |||||
| Horcoff | |||||
| Moreau | |||||
| Peca | |||||
| Pisani | |||||
| Pronger | |||||
| Samsonov | |||||
| Smyth | |||||
| Spacek | |||||
| Staios | |||||
| Stoll | |||||
| Torres | |||||
| Total |
• Chris Pronger took 25% of the Oilers shots that got through to the net.
• Every one of Pronger's shots that got through were from beyond 37.5 feet.
• Chris Pronger took 24% of the Oilers missed and through shots.
• Oilers defencemen took 40% of the Oilers shots that got through to the net.
• Oilers defencemen took 41% of the Oilers missed and through shots.
• Bergeron, Pronger, Spacek, Staois and Stoll took 52% of the Oilers shots that got through to the net.
• Bergeron, Pronger, Spacek, Staois and Stoll took 52% of the Oilers missed and through shots.
• Pronger and Spacek took 44% of the Oilers missed shots.
• Pronger, Spacek, and Stoll took 52% of the Oilers missed shots.
| vs. Detroit | |||||||||
| vs. San Jose | |||||||||
| vs. Anaheim | |||||||||
| vs. Carolina |
• The Oilers scored a powerplay goal in 5 of the 6 games in Round One.
• The Oilers scored a powerplay goal in 4 of the 6 games in Round Two.
• The Oilers scored a powerplay goal in 4 of the 5 games in Round Three.
• The Oilers scored a powerplay goal in 3 of the 7 games in Round Four.
• The Oilers overall powerplay success decreased through every round of the 2006 playoffs.
• Excluding Game 6, the Oilers powerplay success rate was 5.5% in the Stanley Cup Finals.
"Now the partisan, when he is engaged in a dispute, cares nothing about the rights of the question, but is anxious only to convince his hearers of his own assertions. And the difference between him and me at the present moment is merely this--that whereas he seeks to convince his hearers that what he says is true, I am rather seeking to convince myself; to convince my hearers is a secondary matter with me."
--Plato, Phaedo
Taking a look back at my original queries and opinions, I would say that I was both bang on and totally off, with a healthy dose of "who the hell really knows" thrown in the mix.
Question: Did the Oilers powerplay get "closer" to being good as the series against Carolina went along?
Answer: I certainly don't think so. As the numbers in Part IV show, the Oilers powerplay success rate dropped through every single round of the playoffs. They started out with a 21.6% success rate against the Wings in Round 1, and ended up at 11% by the end of the Finals. The graph in Part IV also doesn't indicate any growing strength in the seven games against the Canes. The Oilers did score three powerplay in Game 6, but they went back to scoring 0 in Game 7. The powerplay success rate in the Finals is especially alarming if you consider that the Game 6 success bumped up the percentages by 5.5%. The Oilers shouldn't be criticized for scoring powerplay goals in Game 6, but it really was the exception to the rule in the Finals.
Question: Other than Hemsky's goal in Game 5, when was the last time someone other than a player on the point took a shot on the Oilers powerplay?
Answer: It actually happened a minute and thirteen seconds before Hemsky's goal, on a shot from Raffi Torres. Clearly an absurd claim on my part. But a 33 foot shot from Mike Peca at 16:23 of the first period, in Game 4, was the last time a shot from less than 37.5 feet was taken before that. That's about 55 minutes of space in between shots from somewhere other than the point.
Question: Did over 80% of the Oilers shots on the powerplay throughout the Finals come from beyond 37.5 feet?
Answer: No. That number is much lower, at about 48%. But several points should be made here. The first is that the number goes up by 14 points, to 62%, if you take out the 15 shots the Oilers shot from less than 37.5 feet in Games 1 and 6. If you take out Games 1 and 6, and look at shots from less than 31 feet, the numbers change even more. Excluding those two games, 73% of the Oilers shots were taken from beyond 31 feet.
It should also be noted that this doesn't count missed and blocked shots. 57% of Edmonton's powerplay shots in the Finals didn't even make it on the net. 37 shots were blocked, and 27 were missed. Jaroslav Spacek had 6 missed shots, and so did Chris Pronger. All of Pronger's shots that actually got through to the net were from beyond 37.5 feet. Is it reasonable to assume that all of his missed shots were also from beyond that distance? Of the 5 shots Spacek got through to the net, 4 were from beyond 37.5. Do I make a similar assumption with him? That's an extra 9 or 10 shots added to list, which would surely raise the percentage of shots from beyond 37.5. And what about those 37 blocked shots that I have no distance or shooter for? How many were taken from beyond 37.5 feet? It's all speculation, obviously, but I think the numbers are misleading, and much closer to my 80% guess than they are with the incomplete statistics.
Question: Did over 80% of the Oilers shots on the powerplay throughout the 2006 Playoffs come from beyond 37.5 feet?
Answer: I don't know. The only thing I bothered to look at for the other three rounds was the overall playoff success rate. The lack of statistical data was going to prevent me from even seeing a complete picture of the Finals. I wasn't about to waste another 50 hours looking at more incomplete information.
Question: Can you verify that shots at a distance beyond 37.5 feet are low percentage?
Answer: No. But others can, and have. Take a look at the graph from Alan Ryder's work on shot quality.
The numbers are from the 2002-2003 season, but I see no reason why the arc would have changed dramatically. Statistically speaking, distance matters, and the closer the shot, the better.
Question: Can you verify that a high volume of shots taken from the point also lowers their effectiveness, despite the quality of the shot-taker, due to their predictabilty?
Answer: Again, no. I think the way to do this would be to monitor the Oilers' success on the powerplay throughout the entire season, as well as the playoffs. Then you could properly determine the percentage of shots taken from the point all year (assuming that you could actually get this data, which as I have shown, you can't), how well an opposing team adjusted to the strategy, and whether that affected their powerplay success rate. But the Oilers powerplay success rate did drop in each consecutive round of the playoffs, and that has to count for something, even if it is a small something.
Question: Can you verify that a powerplay strategy primarily focused on the point shot is a waste of the talent within those players north of the hashmarks?
Answer: I didn't even try. I wouldn't even know where to begin on that one. Maybe a cost/benefit analysis? You've got me. It is a neat concept though, and raises some interesting questions. Are kids with "better hands" placed at the forward positions, or do they develop "better hands" because they are placed in the forward positions? It also relates to the question before this. How important is shot distribution? When should the puck be placed on the stick of your best shooter? And so on.
Question: Does anyone keep a running tab of shots in the NHL, like they do in the NBA? And is it public?
Answer: Part II deals with this, and the answer is yes and no. CBS Sportsline does have a shot chart, and it has similar functions to the shot charts the NBA publishes. But the information on them doesn't align with the information on NHL.com, so their value is questionable. My guess is that CBS and the NHL use different companies to keep track of things, which creates the discrepancy. It does illustrate the subjectivity of the NHL statistics many take to be Gospel, however. Those statistics are still being tabulated by human beings, and I'm 100% certain they are making subjective judgments on things like shots, in the same way that baseball scorers make subjective judgements on things like hits and errors.
Question: Was a poor coaching strategy inhibiting the Oilers powerplay?
Answer: While admitting that I don't have all the data, yes. This is obviously a complicated question with many different aspects, and my whole goal has been to try and sort through it all. My failure to answer this question as completely as I could is, at least to me, no fault of my own. As a point of fact, the end result of this entire post may not so much be an indictment of the Oilers powerplay, but rather an indictment of the NHL's incomplete statistical gatherings. I do not know the location of any of the Oilers powerplay shots. I do not know how the NHL gauges shot distance. I do not know the distance or the location of missed powerplay shots. I do not know the distance, location or shooter of blocked powerplay shots. Without this data, I am prevented from seeing what I would like to see. I am blinded from truly knowing whether I am right or wrong about the Oilers powerplay.
But there are some things I do know for certain. I do know that the Oilers powerplay success rate declined through each consecutive round of the playoffs. I also know for sure that every single powerplay goal the Oilers did score in the Finals was from less than 31 feet, the distance from the slot to the back boards. I know for certain that Chris Pronger took the majority of missed and through shots on the Oilers powerplay, doubling the amount of the two next closest forwards (Ales Hemsky and Ryan Smyth). I know for certain that fellow defenceman Jaroslav Spacek had the next highest amount of missed and through powerplay shots. I know for certain that every single one of Pronger's through shots were taken from beyond 37.5 feet. I also know for certain that in Game Six, when the Oilers scored 3 of their 5 Finals powerplay goals, they took 9 shots from 31 feet or less. That was half of their total shots through for that game, and a substantially larger number of shots from that distance than they took in any other Finals game. These things I know.
The hockey powerplay is unique in North American professional sports. No other major professional sport on this continent allows the rules--and the breaking of them--to create such a regularly distinct offensive advantage and opportunity. The free throw exists in basketball, and it is likely the closest comparison, yet it is functionally different from the powerplay. There, a player gets one to three uninterrupted shots from a mandated distance. But regular play is stopped, and the shooter works alone. In hockey, the penalized player leaves the ice, and regular, team play, resumes. As such, it is difficult to draw analogies on the best strategies for success. Yet like in basketball, and even football, certain things are critical if teams wish to succeed. Things like ball and puck distribution come to mind, as well as maximizing your odds by taking shots at the right time, with the right players.
It is my contention that the Edmonton Oilers failed in this regard during the 2006 Stanley Cup Finals. They placed too much emphasis on having their powerplay shots come from certain players, at certain distances. Those distances were not statistically advantageous, leading to a decline in powerplay success rate. In addition, because of the repetition of the act, the Carolina Hurricanes were able to key in on certain players and a certain distance from the ice, again leading to a decline in powerplay success rate. It was the equivalent, in my mind, of having one or two players consistently taking a 3-point shot in basketball. The distance leads to a poor success rate, and after a while, the opposing team gets a lot of hands in the faces of those shooters. The success rate plummets. Another equivalent might be a football team consistently attempting to throw fifty-yard passes, to the same receiver. Again, the distance is inopportune, and after a while a team just starts keying in on the route.
It is my belief that the Oilers should have developed a better powerplay strategy, focused on the activities of better puck circulation and shot distribution, eventually culminating in taking high percentage shots from high percentage places. That is why I made the suggestion, after Game 4, that the Oilers dress Robbie Schremp for the rest of the playoffs. Tyler disagreed with me, stating among other things that as a replacement player Schremp's value was less than other players in the lineup, and that looking at the 82 games in the regular season proved that the Oilers powerplay was simply better than the Canes penalty kill. Of course he did this all without any actual VORP numbers, and by ignoring the fact that Carolina hadn't played Edmonton 82 times that year, and that adjustments had clearly been made to shut down the Oilers singular powerplay strategy, but I digress. I suggested in his comments that Schremp shouldn't be thought of as a real player, but as a metaphor. Here is my explanation for that statement. My point was that what the Oilers needed at that time wasn't another defenceman or forward who liked to take howitzers from 60 feet away, but a forward with good hands and a fearless ability to shoot from close distances. I really didn't care who it was, as long as it happened. I know that ultimately without a reformulated plan of attack from Craig Simpson and Craig MacTavish, the personnel wouldn't matter. But my belief was that Schremp was egotistical enough to ignore the plan of attack, and shoot from 25 feet or less anytime he touched the puck.
I stand by my suggestion even now, though I believe I should have suggested that Tyler think of Schremp as a change in philosophy, rather than as a metaphor. There is an orthodoxy in the game of hockey, as there is in every sport. I believe this orthodoxy, this stubborn refusal to change or modify beliefs and strategies in the face of predictive or actual failure, cost the Edmonton Oilers the Stanley Cup. I believe that more shots from less than 37.5 feet, more variety in the shooters taking those shots, all mixed with the occasional blast from the point, would have allowed the Oilers to score 2 to 4 extra goals in the Finals. My belief is especially strong considering the many 5-on-3 chances the Oilers had in the Finals, which I didn't even begin to discuss in this post. 2 to 4 extra goals would have made a big difference, likely a big enough difference to bring the Stanley Cup back to Edmonton, where it so rightly belongs. Chris Pronger would have still left town, but we'd have all been too drunk to care.
# posted by andy grabia @ 3:03 am 
Comments:
It also seems to me that there may be a simple cost/benefit analysis going on here too. Instinctively the defensive team knows that shots closer to the goal are more dangerous and thus collapse quicker to make those shots a) harder to get and b)harder to get through. Shots from the point have less chance of scoring but they're easier to get. So do you go with the (lots of shots)x(low probability) or (few shots)x(high probability). And remember players have to make these calculations in real time, while opponents harass them while under the pressure of winning/losing the Stanley Cup, on ice of varying degress of quality. For example, it may make more sense to try the extra pass early in a period when the ice quality is good and to blast away on the point, late in a period when the ice gets choppy.
Of course had you simply found someone who has all the data already you would've saved yourself a lot of trouble...
Carolina in contrast had:
42% > 37.5 ft
58% < 37.5 ft.
It would almost be worthwhile to petition the NHL for two pieces of information on distance ([x, y] or [r, theta]). Doubt it will ever happen though.
Of course had you simply found someone who has all the data already you would've saved yourself a lot of trouble...
Sure, but isn't the point that I shouldn't have to hunt it down on the internet from people I have never met, who may or may not be unreliable sources? And even those people are unlikely to have all the info I needed.
Without all the numbers, the smell test told me that firing away from the point wasn't going to do a Goddamned bit of good all by itself, which is kind of what it was. It seemed to me (and you have indeed proven) that nearly all of our PP goals came from below the top of the circles, meaning that the forwards needed to be doing a lot more shooting than they were, something Game 6 proved succinctly. In the other games, I recall a couple of times when, for example, Horcoff or Hemsky made a pass back to the point that wound up clear out of the zone, even though they had at least decent chances of getting through, and screaming at my TV in holy rage. I think you've really confirmed what I was thinking all along, that the Oilers' PP was too reliant on Pronger (how many of those Spacek shots did he feed? I'm willing to bet well over half), and needed to key a bit more on Hemsky, who's got loads of natural talent, but still doesn't quite use it to its fullest, even with the leap he took last year. I also think Horcoff needs to be less timid on the PP next season, and I hope that this thing you've demonstrated is something that the coaching staff has realized over the course of the offseason.
While losing Pronger cost us our first marquee (outside Northern Alberta) star in several years, and a solid two-way physical defensive presence, it may be a blessing in disguise, for the PP, anyway, as it limits our options from the point, and the return (Lupul) further forces the coaches' hands in playing more of the PP down low. I think they'll still try to go with Stoll, Tjarnqvist, Bergeron, and Staios blasting from the point to at least some degree, but not nearly to the point it hit this past season, for simple practicality.
Finally, because it's relevant and still kinda funny, the Flames' power play seems to operate on a similar principle. "Phaneuf shoots again, and he missed the net!"
"Paradise Lost"..."War and Peace"..."On Powerplay"...
Finally, because it's relevant and still kinda funny, the Flames' power play seems to operate on a similar principle. "Phaneuf shoots again, and he missed the net!"
It's true that Calgary's PP was overly reliant upon shots from the point as well. Whether this is due to a lack of strategy or a lack of skilled personnel at forward (or both) is hard to say.
What I do know is that Kristian Huselius was actually the most efficient at getting points with the man-advantage on the Flames roster - so, Im also hoping the Flames re-jig the ol' PP next season, with a focus on getting more high percentage shots from the forwards down-low.
I can't comment on strategy, but definitely, there seemed to be a lack of skilled personnnel, though Reinprecht's sudden emergence in those last few weeks in Phoenix suggests there's a bit of the former in there, as well. Tanguay should fix that to some degree, and I suspect that if Huselius, like Hemsky, gets a bit more love on the PP, it could drastically affect the outcome, though I'm hoping Playfair's not that clever. ;).
A final point. I know this could probably be determined with A LOT of work but you don't know that the goals scored on the PP (<31 feet) were not from rebounds or blocked shots from the point.
I do actually. I can do it from memory. The first Hemsky goal was on that deke in Game 1. The second Hemsky goal was on the one-timer from the left circle in Game 5. The Pisani one was a weird one that I think did originate from the point. The Smyth and Horcoff goals were clean shots by both men.
As for the variables, they exist in every sport. But that doesn't mean the results aren't worth looking at. You could probably create a formula to adjust player performance based on said variables like ice quality, opposition performance, and what not. But even I admit that my queries here are only first steps, not final solutions.
As for your assertion that the OK adjusts, I agree. It in large part shapes my thoughts on how unsuccessful Edmonton was. Carolina adjusted to our powerplay, and it cost us. But in terms of pure numbers, let's not ignore that the powerplay is at an advantage. It should be able to determine the flow of the game for those two minutes. That is especially the case on the 5-on-3. I don't have any numbers to back me up, but my recollection is that the Oilers continued the same strategy of passing to the point on the 2 man advantage. This makes little sense to me. The 5-on-3 should be all about rapid puck movement, and lots of shots from down low. There simply aren't enough bodies to cover all the men. Yet the Oilers barelt got a sniff on the 5-on-3.
Great work, Grabia. This is a really well-put together piece.
Have you sent it to the NHL in any form? I'd love to see if they respond or offer better alternatives to collecting pertinent data.
No I haven't. I thought about it, but I wanted to see if others could fill in any blanks before I did.
And thanx.
tibia money
tibia gold
tibia item
runescape
money runescape
gold tibia
money tibia
gold runescape
gold runescape
accounts tibia gold tibia money runescape
money runescape
gp buy
runescape gold tibia gold tibia item buy
runescape money runescape
gold runescape
items tibia
money tibia
gold
Post a Comment
<< Home
It also seems to me that there may be a simple cost/benefit analysis going on here too. Instinctively the defensive team knows that shots closer to the goal are more dangerous and thus collapse quicker to make those shots a) harder to get and b)harder to get through. Shots from the point have less chance of scoring but they're easier to get. So do you go with the (lots of shots)x(low probability) or (few shots)x(high probability). And remember players have to make these calculations in real time, while opponents harass them while under the pressure of winning/losing the Stanley Cup, on ice of varying degress of quality. For example, it may make more sense to try the extra pass early in a period when the ice quality is good and to blast away on the point, late in a period when the ice gets choppy.
# posted by 
: July 20, 2006 4:39 am
: July 20, 2006 4:39 am Of course had you simply found someone who has all the data already you would've saved yourself a lot of trouble...
Carolina in contrast had:
42% > 37.5 ft
58% < 37.5 ft.
It would almost be worthwhile to petition the NHL for two pieces of information on distance ([x, y] or [r, theta]). Doubt it will ever happen though.
# posted by : July 20, 2006 10:03 am
: July 20, 2006 10:03 am Of course had you simply found someone who has all the data already you would've saved yourself a lot of trouble...
Sure, but isn't the point that I shouldn't have to hunt it down on the internet from people I have never met, who may or may not be unreliable sources? And even those people are unlikely to have all the info I needed.
Without all the numbers, the smell test told me that firing away from the point wasn't going to do a Goddamned bit of good all by itself, which is kind of what it was. It seemed to me (and you have indeed proven) that nearly all of our PP goals came from below the top of the circles, meaning that the forwards needed to be doing a lot more shooting than they were, something Game 6 proved succinctly. In the other games, I recall a couple of times when, for example, Horcoff or Hemsky made a pass back to the point that wound up clear out of the zone, even though they had at least decent chances of getting through, and screaming at my TV in holy rage. I think you've really confirmed what I was thinking all along, that the Oilers' PP was too reliant on Pronger (how many of those Spacek shots did he feed? I'm willing to bet well over half), and needed to key a bit more on Hemsky, who's got loads of natural talent, but still doesn't quite use it to its fullest, even with the leap he took last year. I also think Horcoff needs to be less timid on the PP next season, and I hope that this thing you've demonstrated is something that the coaching staff has realized over the course of the offseason.
While losing Pronger cost us our first marquee (outside Northern Alberta) star in several years, and a solid two-way physical defensive presence, it may be a blessing in disguise, for the PP, anyway, as it limits our options from the point, and the return (Lupul) further forces the coaches' hands in playing more of the PP down low. I think they'll still try to go with Stoll, Tjarnqvist, Bergeron, and Staios blasting from the point to at least some degree, but not nearly to the point it hit this past season, for simple practicality.
Finally, because it's relevant and still kinda funny, the Flames' power play seems to operate on a similar principle. "Phaneuf shoots again, and he missed the net!"
# posted by : July 20, 2006 10:15 am
: July 20, 2006 10:15 am "Paradise Lost"..."War and Peace"..."On Powerplay"...
Finally, because it's relevant and still kinda funny, the Flames' power play seems to operate on a similar principle. "Phaneuf shoots again, and he missed the net!"
It's true that Calgary's PP was overly reliant upon shots from the point as well. Whether this is due to a lack of strategy or a lack of skilled personnel at forward (or both) is hard to say.
What I do know is that Kristian Huselius was actually the most efficient at getting points with the man-advantage on the Flames roster - so, Im also hoping the Flames re-jig the ol' PP next season, with a focus on getting more high percentage shots from the forwards down-low.
I can't comment on strategy, but definitely, there seemed to be a lack of skilled personnnel, though Reinprecht's sudden emergence in those last few weeks in Phoenix suggests there's a bit of the former in there, as well. Tanguay should fix that to some degree, and I suspect that if Huselius, like Hemsky, gets a bit more love on the PP, it could drastically affect the outcome, though I'm hoping Playfair's not that clever. ;).
# posted by : July 20, 2006 10:39 am
: July 20, 2006 10:39 am A final point. I know this could probably be determined with A LOT of work but you don't know that the goals scored on the PP (<31 feet) were not from rebounds or blocked shots from the point.
I do actually. I can do it from memory. The first Hemsky goal was on that deke in Game 1. The second Hemsky goal was on the one-timer from the left circle in Game 5. The Pisani one was a weird one that I think did originate from the point. The Smyth and Horcoff goals were clean shots by both men.
As for the variables, they exist in every sport. But that doesn't mean the results aren't worth looking at. You could probably create a formula to adjust player performance based on said variables like ice quality, opposition performance, and what not. But even I admit that my queries here are only first steps, not final solutions.
As for your assertion that the OK adjusts, I agree. It in large part shapes my thoughts on how unsuccessful Edmonton was. Carolina adjusted to our powerplay, and it cost us. But in terms of pure numbers, let's not ignore that the powerplay is at an advantage. It should be able to determine the flow of the game for those two minutes. That is especially the case on the 5-on-3. I don't have any numbers to back me up, but my recollection is that the Oilers continued the same strategy of passing to the point on the 2 man advantage. This makes little sense to me. The 5-on-3 should be all about rapid puck movement, and lots of shots from down low. There simply aren't enough bodies to cover all the men. Yet the Oilers barelt got a sniff on the 5-on-3.
Great work, Grabia. This is a really well-put together piece.
Have you sent it to the NHL in any form? I'd love to see if they respond or offer better alternatives to collecting pertinent data.
No I haven't. I thought about it, but I wanted to see if others could fill in any blanks before I did.
And thanx.
tibia money
tibia gold
tibia item
runescape
money runescape
gold tibia
money tibia
gold runescape
gold runescape
accounts tibia gold tibia money runescape
money runescape
gp buy
runescape gold tibia gold tibia item buy
runescape money runescape
gold runescape
items tibia
money tibia
gold
# posted by : August 01, 2008 12:42 am
: August 01, 2008 12:42 am Post a Comment
<< Home
