The number after the team name in parentheses is goal differential for the season.
The Actual Playoff Picture
1. Washington (+27) - 107
2. Philadelphia (+36) - 106
3. Boston (+51) - 103
4. Pittsburgh (+39) - 106
5. Tampa Bay (+7) - 103
6. Montréal (+7) - 96
7. Buffalo (+16) - 96
8. New York Rangers (+35) - 93
1. Vancouver (+77) - 117
2. San Jose (+35) - 105
3. Detroit (+20) - 104
4. Anaheim (+4) - 99
5. Nashville (+25) - 99
6. Phoenix (+5) - 99
7. Los Angeles (+21) - 98
8. Chicago (+33) - 97
The Old-School-Style Playoff Picture
1. Philadelphia (+36) - 98
2. Washington (+27) - 97
3. Boston (+51) - 96
4. Tampa Bay (+7) - 92
5. Pittsburgh (+39) - 91
6. Montréal (+7) - 88
7. New York Rangers (+35) - 82
8. Buffalo (+16) - 82
1. Vancouver (+77) - 109
2. San Jose (+35) - 96
3. Detroit (+20) - 94
4. Anaheim (+4) - 92
5. Chicago (+33) - 87
6. Phoenix (+5) - 87
7. Nashville (+25) - 86
8. Dallas (-6) - 86
The Soccer-Style Playoff Picture
1. Philadelphia (+36) - 147
2. Boston (+51) - 146
3. Washington (+27) - 141
4. Pittsburgh (+39) - 140
5. Tampa Bay (+7) - 135
6. Montréal (+7) - 132
7. New York Rangers (+35) - 125
8. Buffalo (+16) - 124
1. Vancouver (+77) - 163
2. San Jose (+35) - 143
3. Detroit (+20) - 138
4. Nashville (+25) - 135
5. Phoenix (+5) - 135
6. Anaheim (+4) - 133
7. Los Angeles (+21) - 133
8. Chicago (+33) - 131
Some initial scattered thoughts:
---No matter what you do, Vancouver is better than you.
---Weird that the old-school system actually admits a team with a negative goal differential, in Dallas replacing the Kings.
---Both new systems make ties in points in the standings less common (although that seems counter-intuitive in the old-school system).
---Without the artificial "division-winners take the top three spots" rule, in both the old-school-style and soccer-style systems, the three division-winners ended up in the top three spots, in both conferences. In the current system, the rule was needed to place Boston ahead of Pittsburgh.
---The West really is that much better than the East.
---Other than that one substitution (Dallas for LA), both systems give us the same playoff teams as the current system this season.
---I'm thinking that the current system doesn't necessarily give you the wrong playoff teams and matchups, it just adds more randomness into the system, so you might get the wrong teams and matchups. The current system brings teams that are close in the standings even closer than they should be. It looks like this season, hockey got lucky. What I'm saying is that the teams that made it by a hair this season would have made it by a few more hairs using these other systems.
Showing posts with label shootouts. Show all posts
Showing posts with label shootouts. Show all posts
Monday, April 11, 2011
Monday, April 4, 2011
Wednesday, March 30, 2011
Revisionist Eastern Conference Playoffs 2010
So, with my methods well understood (I hope?), let's look at some data. Last season, you may well recall missing the playoffs by a point. Here's how that broke down:
1. Washington - 121
2. New Jersey - 103
3. Buffalo - 100
4. Pittsburgh - 101
5. Ottawa - 94
6. Boston - 91
7. Philadelphia - 88
8. Montréal - 88
--------------------
9. NEW YORK RANGERS - 87
I don't want to sound like a bitter fan, claiming we deserved a playoff spot last season, but let's take a look at what our new, sane point systems would have heralded. Here's the breakdown using the old-school points system (each game is worth two points):
1. Washington - 109
2. New Jersey - 95
3. Buffalo - 92
4. Pittsburgh - 88
5. Ottawa - 88
6. Philadelphia - 81
7. Boston - 77
8. NEW YORK RANGERS - 77
--------------------
9. Montréal - 76
And here's how it breaks down using the soccer-style system (each game is worth three points):
1. Washington - 164
2. New Jersey - 143
3. Buffalo - 135
4. Pittsburgh - 134
5. Ottawa - 128
6. Philadelphia - 123
7. NEW YORK RANGERS - 121
8. Boston - 116
--------------------
9. Montréal - 112
Rather than looking at the benefits of the soccer system vs. the old-school system, here's what's interesting to me: these new systems seem to agree with each other more than either one agrees with the stupid system we use now. Both new systems agree that the Canadiens are the team that should have fallen short last season, and they agree that Philly deserved a higher ranking. They also agree that Buffalo should have beaten Pittsburgh in points (under the stupid current system, Buffalo was handed the #3 spot due to winning their division, but Pittsburgh actually had more points).
In fact, let's take a look at the one place the two new systems disagree: the #7 and #8 spots. They're actually tied in points: Boston is only handed the #7 spot because of my tiebreaker concession: they have more shootout wins. If we were to do away with the shootout entirely, that tiebreaker would fall to the season series victor, which was the Rangers, 3-1. Which means that without the shootout tiebreaker, the two new systems actually agree 100% on what the standings should have been, and it is not what they were.
I'm not saying we wouldn't necessarily have ended up with the goddamn Flyers representing the conference anyway. I am saying that it's interesting to note that the two sane systems I'm proposing agree with each other in terms of what the playoff picture should be, and that what the NHL actually does seems to disagree. And that we missed out on yet another Rangers-Devils first round.
I wonder if other seasons, or the Western Conference, will see a closer relationship between the old-school system and the soccer-style system than between either of those and the current, stupid system. More to come, when I get a chance.
1. Washington - 121
2. New Jersey - 103
3. Buffalo - 100
4. Pittsburgh - 101
5. Ottawa - 94
6. Boston - 91
7. Philadelphia - 88
8. Montréal - 88
--------------------
9. NEW YORK RANGERS - 87
I don't want to sound like a bitter fan, claiming we deserved a playoff spot last season, but let's take a look at what our new, sane point systems would have heralded. Here's the breakdown using the old-school points system (each game is worth two points):
1. Washington - 109
2. New Jersey - 95
3. Buffalo - 92
4. Pittsburgh - 88
5. Ottawa - 88
6. Philadelphia - 81
7. Boston - 77
8. NEW YORK RANGERS - 77
--------------------
9. Montréal - 76
And here's how it breaks down using the soccer-style system (each game is worth three points):
1. Washington - 164
2. New Jersey - 143
3. Buffalo - 135
4. Pittsburgh - 134
5. Ottawa - 128
6. Philadelphia - 123
7. NEW YORK RANGERS - 121
8. Boston - 116
--------------------
9. Montréal - 112
Rather than looking at the benefits of the soccer system vs. the old-school system, here's what's interesting to me: these new systems seem to agree with each other more than either one agrees with the stupid system we use now. Both new systems agree that the Canadiens are the team that should have fallen short last season, and they agree that Philly deserved a higher ranking. They also agree that Buffalo should have beaten Pittsburgh in points (under the stupid current system, Buffalo was handed the #3 spot due to winning their division, but Pittsburgh actually had more points).
In fact, let's take a look at the one place the two new systems disagree: the #7 and #8 spots. They're actually tied in points: Boston is only handed the #7 spot because of my tiebreaker concession: they have more shootout wins. If we were to do away with the shootout entirely, that tiebreaker would fall to the season series victor, which was the Rangers, 3-1. Which means that without the shootout tiebreaker, the two new systems actually agree 100% on what the standings should have been, and it is not what they were.
I'm not saying we wouldn't necessarily have ended up with the goddamn Flyers representing the conference anyway. I am saying that it's interesting to note that the two sane systems I'm proposing agree with each other in terms of what the playoff picture should be, and that what the NHL actually does seems to disagree. And that we missed out on yet another Rangers-Devils first round.
I wonder if other seasons, or the Western Conference, will see a closer relationship between the old-school system and the soccer-style system than between either of those and the current, stupid system. More to come, when I get a chance.
Monday, March 28, 2011
Different Standings Systems
Hey, so, here's the first in what may or may not be a series of posts about possible systems of standings we could use instead of the really stupid system we use now. To start, let's talk about the really stupid system we use now.
Right now, if a game is tied after regulation, it goes into a 5-minute, 4-on-4 overtime period. If it remains tied after that overtime, it goes into a shootout, which is guaranteed to determine a victor. In the standings, the winner is given two points no matter what. The loser is given zero points if the game ended within regulation, but one point if the game was tied after regulation. This, as mentioned, is stupid.
It is stupid for exactly two reasons. One: the shootout is not hockey, and should not be used to determine points in the standings of a league that plays hockey. This is not immediately mathematically obvious, but is generally agreed upon by almost every single hockey fan that does not own a Sidney Crosby Snuggie. Two: some regular season games are worth more points than others, which is mathematically obviously a terrible idea.
In order to try to reduce some of the obvious stupidity, the GMs agreed to use "regulation plus overtime wins" as the first tiebreaker in the standings. In other words: if two teams are tied in points at the end of the regular season, the higher spot in the standings will go to the team that got more of those points not from the shootout. Good, that's a start. It actually resolves zero of the two issues I listed above, but it's a start nonetheless.
So, to further alleviate (or eliminate entirely) the really unbelievable stupidity, here are two other potential point systems we could be using, each significantly less infuriatingly stupid than the one we're using now. Both systems feature the benefit of using the same exact game structure we use now, just counting points differently.
System A: Old-School, or "do what we did before we started doing the stupid thing we do right now." If the game ends in regulation or in the overtime period, the winner gets two points, and the loser gets zero points. Because the winner won the damn game, which was worth two damn points. If the game remains tied after the overtime, each team gets one point (the two points are split among the two teams, who have tied). Then, because the NHL hates ties, have the shootout anyway, and use it to determine the "winner" of the game. This team gets no additional points in the standings, but records one "shootout win." We can use the shootout wins as the first tiebreaker, so they are not actually meaningless, but they're significantly less meaningful.
System B: Soccer-Style, or "just do what FIFA does, they are smarter than you." If you want to give two points to the winner and one point to the loser of every game that goes into overtime, fine. But then games are worth three points. So make a game worth three points! If the game ends in regulation, the winner gets three points and the loser gets zero. If the game ends in overtime or a shootout, the winner gets two points and the loser gets one. Holy shit two plus one is three you guys.
So, there are plenty of reasons that either of these systems would be all hells of kinds of superior to the one we use now. Most of them simplify to "the system we use now is stupid," so I won't go into that now. Here's what's cool, though: we can play "What If?" With the magic of arithmetic, we can find out how previous seasons since the lockout would have ended, if the NHL were using either of these obviously better systems instead of its astoundingly stupid one. Here are my methods:
First, the old-school system. If a game ended in regulation, we currently do exactly what we want in the old-school system (2 points to the winner, 0 to the loser), so no change needs to be made. If a game ended in the overtime period, we currently award 2 points to the winner and 1 to the loser, and we should award 2 to the winner and 0 to the loser, so each team needs to lose one point per loss within the overtime period. Then, if a game ended in a shootout, we still award 2 points to the winner and 1 to the loser, and we should give one point to each instead. So each team needs to lose one point per shootout win. So, for each team, if we subtract the number of times they lost within an overtime period and their number of shootout wins from their total points, we will get the number of points they deserve in the old-school system.
As for soccer-style, if a game was tied after regulation, we currently do exactly what we want (2 points to the winner, 1 to the loser). However, if a game ended in regulation, we currently give 2 points to the winner and 0 to the loser, whereas we should give three to the winner (and 0 to the loser). So each team needs to gain one point per regulation win. For each team, if we add the number of times they won in regulation to their total points, we will get the number of points they deserve in the soccer-style system.
Coming up next (possibly): some results! (I've calculated them for the Eastern Conference for every season, I haven't done the Western Conference yet at all. I'll post these results if I have time today or tomorrow.)
Right now, if a game is tied after regulation, it goes into a 5-minute, 4-on-4 overtime period. If it remains tied after that overtime, it goes into a shootout, which is guaranteed to determine a victor. In the standings, the winner is given two points no matter what. The loser is given zero points if the game ended within regulation, but one point if the game was tied after regulation. This, as mentioned, is stupid.
It is stupid for exactly two reasons. One: the shootout is not hockey, and should not be used to determine points in the standings of a league that plays hockey. This is not immediately mathematically obvious, but is generally agreed upon by almost every single hockey fan that does not own a Sidney Crosby Snuggie. Two: some regular season games are worth more points than others, which is mathematically obviously a terrible idea.
In order to try to reduce some of the obvious stupidity, the GMs agreed to use "regulation plus overtime wins" as the first tiebreaker in the standings. In other words: if two teams are tied in points at the end of the regular season, the higher spot in the standings will go to the team that got more of those points not from the shootout. Good, that's a start. It actually resolves zero of the two issues I listed above, but it's a start nonetheless.
So, to further alleviate (or eliminate entirely) the really unbelievable stupidity, here are two other potential point systems we could be using, each significantly less infuriatingly stupid than the one we're using now. Both systems feature the benefit of using the same exact game structure we use now, just counting points differently.
System A: Old-School, or "do what we did before we started doing the stupid thing we do right now." If the game ends in regulation or in the overtime period, the winner gets two points, and the loser gets zero points. Because the winner won the damn game, which was worth two damn points. If the game remains tied after the overtime, each team gets one point (the two points are split among the two teams, who have tied). Then, because the NHL hates ties, have the shootout anyway, and use it to determine the "winner" of the game. This team gets no additional points in the standings, but records one "shootout win." We can use the shootout wins as the first tiebreaker, so they are not actually meaningless, but they're significantly less meaningful.
System B: Soccer-Style, or "just do what FIFA does, they are smarter than you." If you want to give two points to the winner and one point to the loser of every game that goes into overtime, fine. But then games are worth three points. So make a game worth three points! If the game ends in regulation, the winner gets three points and the loser gets zero. If the game ends in overtime or a shootout, the winner gets two points and the loser gets one. Holy shit two plus one is three you guys.
So, there are plenty of reasons that either of these systems would be all hells of kinds of superior to the one we use now. Most of them simplify to "the system we use now is stupid," so I won't go into that now. Here's what's cool, though: we can play "What If?" With the magic of arithmetic, we can find out how previous seasons since the lockout would have ended, if the NHL were using either of these obviously better systems instead of its astoundingly stupid one. Here are my methods:
First, the old-school system. If a game ended in regulation, we currently do exactly what we want in the old-school system (2 points to the winner, 0 to the loser), so no change needs to be made. If a game ended in the overtime period, we currently award 2 points to the winner and 1 to the loser, and we should award 2 to the winner and 0 to the loser, so each team needs to lose one point per loss within the overtime period. Then, if a game ended in a shootout, we still award 2 points to the winner and 1 to the loser, and we should give one point to each instead. So each team needs to lose one point per shootout win. So, for each team, if we subtract the number of times they lost within an overtime period and their number of shootout wins from their total points, we will get the number of points they deserve in the old-school system.
As for soccer-style, if a game was tied after regulation, we currently do exactly what we want (2 points to the winner, 1 to the loser). However, if a game ended in regulation, we currently give 2 points to the winner and 0 to the loser, whereas we should give three to the winner (and 0 to the loser). So each team needs to gain one point per regulation win. For each team, if we add the number of times they won in regulation to their total points, we will get the number of points they deserve in the soccer-style system.
Coming up next (possibly): some results! (I've calculated them for the Eastern Conference for every season, I haven't done the Western Conference yet at all. I'll post these results if I have time today or tomorrow.)
Subscribe to:
Posts (Atom)