Why aren't losses factored into sorting teams in the standings?

[Nenikoj]

I've been a hockey fan for well over twenty years and there's something about hockey which I've always wondered about. In US sports like the NBA and MLB, the teams are sorted by percentage and games behind, while in the NHL, the teams are sorted by points. The percentage & games behind method seems to make more sense, as it factors in all games played, while the NHL's points system treats a regulation loss no differently than if the team didn't play.

For example, let's say the Raptors are 20-19 and the Celtics are 21-22. The Raptors would be listed ahead of the Celtics, as the Raptors have a better percentage and a better games above .500. In the NHL, if the Canadiens are 20-19-0 and the Maple Leafs are 21-22-0, the Canadiens would be listed behind the Maple Leafs. Let's say the Bruins are 10-9-0 (20 pts) and the Sabres are 9-8-0 (18 pts). The Bruins lose a game while the Sabres don't play. The Bruins are now 10-10-0 and are still at 20 pts. So the regulation loss didn't affect their position in the standings at all.

I know it isn't because of the presence of overtime losses (or previously ties), as an overtime loss is the net equivalent as half a win and half a regulation loss.

I also know it doesn't really matter during the season. But if the standings are going to be published on a regular basis during the season, why sort the teams using a method in which a regulation loss is treated the same as not playing? Why not use a method which is more balanced like the systems used in MLB and the NBA?

Perhaps there are some other hockey fans here who have some insight on why the NHL does this? Thanks

Click to read more...

 

[Batrous]

I've been a hockey fan for well over twenty years and there's something about hockey which I've always wondered about. In US sports like the NBA and MLB, the teams are sorted by percentage and games behind, while in the NHL, the teams are sorted by points. The percentage & games behind method seems to make more sense, as it factors in all games played, while the NHL's points system treats a regulation loss no differently than if the team didn't play.

For example, let's say the Raptors are 20-19 and the Celtics are 21-22. The Raptors would be listed ahead of the Celtics, as the Raptors have a better percentage and a better games above .500. In the NHL, if the Canadiens are 20-19-0 and the Maple Leafs are 21-22-0, the Canadiens would be listed behind the Maple Leafs. Let's say the Bruins are 10-9-0 (20 pts) and the Sabres are 9-8-0 (18 pts). The Bruins lose a game while the Sabres don't play. The Bruins are now 10-10-0 and are still at 20 pts. So the regulation loss didn't affect their position in the standings at all.

I know it isn't because of the presence of overtime losses (or previously ties), as an overtime loss is the net equivalent as half a win and half a regulation loss.

I also know it doesn't really matter during the season. But if the standings are going to be published on a regular basis during the season, why sort the teams using a method in which a regulation loss is treated the same as not playing? Why not use a method which is more balanced like the systems used in MLB and the NBA?

Perhaps there are some other hockey fans here who have some insight on why the NHL does this? Thanks

It's all about making sure the most competitive teams are reflected in the standings. A team that has lost half its games in OT/Shootout is probably going to be more competitive than a team that has a couple more overall wins. The point system is a way to measure which teams overall were the most competitive in the season and therefore most worthy of a playoff spot / lower draft position.

 

[Nenikoj]

It's all about making sure the most competitive teams are reflected in the standings. A team that has lost half its games in OT/Shootout is probably going to be more competitive than a team that has a couple more overall wins. The point system is a way to measure which teams overall were the most competitive in the season and therefore most worthy of a playoff spot / lower draft position.

Are you sure? If team A has 44 pts in 43 games (e.g. 22-21-0) while team B has 46 pts in 47 games (e.g. 23-24-0), hasn't team A performed at a higher level? In that case, what would be the justification for listing team B ahead of team A?

 

[Canadiens1958]

Standings

Baseball and basketball never recorgnized ties. Hockey did.

In hockey points give credit for achievments = wins and previously ties. Percentages do not recognize achievements rather they assume that teams will perform at a constant rate.

Baseball with its GBL approach is the most flawed since a team can always win the unplayed games on its schedule and surpass the perceived le. The team with more loses but ahead in GBL can not unlose the games it lost especially if it is not scheduled to play the team with fewer loses.

 

[JadedLeaf]

Are you sure? If team A has 44 pts in 43 games (e.g. 22-21-0) while team B has 46 pts in 47 games (e.g. 23-24-0), hasn't team A performed at a higher level? In that case, what would be the justification for listing team B ahead of team A?

Standings in the middle of the season really don't matter though. You don't win anything for being in first at the 45 game mark. Most people use point percentage when looking at mid season standings anyway so it would eliminate that problem. At the end of the season once every team has played 82 games is when the standings matter and that issue wouldn't apply here.

 

[Nenikoj]

Baseball and basketball never recorgnized ties. Hockey did.

In hockey points give credit for achievments = wins and previously ties. Percentages do not recognize achievements rather they assume that teams will perform at a constant rate.

It's already been determined that the presence of ties (or OTLs) is not the reason, as an OTL is the net equivalent of half a win and half a regulation loss.

Percentage does not assume anything. It merely looks at what the team has done to date.

Baseball with its GBL approach is the most flawed since a team can always win the unplayed games on its schedule and surpass the perceived le. The team with more loses but ahead in GBL can not unlose the games it lost especially if it is not scheduled to play the team with fewer loses.

Please provide an example.

 

[Nenikoj]

Standings in the middle of the season really don't matter though. You don't win anything for being in first at the 45 game mark. Most people use point percentage when looking at mid season standings anyway so it would eliminate that problem. At the end of the season once every team has played 82 games is when the standings matter and that issue wouldn't apply here.

If so, then why sort the teams during the season using a method in which a loss is equivalent to not playing when there are established methods in which a loss is factored in?

 

[Canadiens1958]

Example

It's already been determined that the presence of ties (or OTLs) is not the reason, as an OTL is the net equivalent of half a win and half a regulation loss.

Percentage does not assume anything. It merely looks at what the team has done to date.


Please provide an example.

Not aware of OTL or ties in MLB or the NBA.

Percentage does not convey games played or ties or post regulation loses generating points.

Example.

Team A: 20W - 9L
Team B: 17W - 8L

Team B is 1 GBL Team A. Team A can never unlose a game if Team B runs the schedule.

 

[morehockeystats]

Not aware of OTL or ties in MLB or the NBA.

Percentage does not convey games played or ties or post regulation loses generating points.

Example.

Team A: 20W - 9L
Team B: 17W - 8L

Team B is 1 GBL Team A. Team A can never unlose a game if Team B runs the schedule.

Team A cannot unwin a game, and Team B has to win 3 out of 4.

 

[Nenikoj]

It's already been determined that the presence of ties (or OTLs) is not the reason, as an OTL is the net equivalent of half a win and half a regulation loss.

Not aware of OTL or ties in MLB or the NBA.

My point is if you take the OTLs of any team and divvy them up as half a win and half a regulation loss, you'll get the same Games above 500, the same points percentage and the same points in the same # of games played. Example: Detroit is 17-13-7 (or 41 pts in 37 games with a .554 PPCT). Take the 7 OTLs and make each half a win and half a regulation loss and you get 20.5-16.5-0, which is also 41 pts in 37 games with a .554 PPCT). Make sense?

Percentage does not convey games played or ties or post regulation loses generating points.

I find it somewhat ironic that one who supports the NHL's points system would criticize the percentage method for not conveying games played, when the NHL's points system also doesn't convey games played.

Points percentage takes into consideration 100% of a teams played games (see the 17-13-7 example above). Points earned ignores losses.

Example.

Team A: 20W - 9L
Team B: 17W - 8L

Team B is 1 GBL Team A. Team A can never unlose a game if Team B runs the schedule.

No team can "unlose" a game. In your example, Team A is one game ahead of Team B. Similar thing would be true if you reversed the records around to have 9W-20L and 8L-17W. 8L-17W would be ahead by one game.

 

[morehockeystats]

My point is if you take the OTLs of any team and divvy them up as half a win and half a regulation loss, you'll get the same Games above 500, the same points percentage and the same points in the same # of games played. Example: Detroit is 17-13-7 (or 41 pts in 37 games with a .554 PPCT). Take the 7 OTLs and make each half a win and half a regulation loss and you get 20.5-16.5-0, which is also 41 pts in 37 games with a .554 PPCT). Make sense?

Bear in mind that in the NHL .500 is usually not par. .5625 is the approximate par, since about 1/4 of the games go to OT, where the extra point is handed out.

 

[Nenikoj]

Bear in mind that in the NHL .500 is usually not par. .5625 is the approximate par, since about 1/4 of the games go to OT, where the extra point is handed out.

That doesn't impact my argument one iota.

There is a great article on that at http://www3.telus.net/dmarchak/can500.htm

 

[morehockeystats]

That doesn't impact my argument one iota.

There is a great article on that at http://www3.telus.net/dmarchak/can500.htm

Ranking teams by P% rather than by actual points is a matter of taste. You can see the latter one also widely used, for example, in European soccer leagues (and it was used there when the point system was 2-1-0), or in chess tournaments.

Another example is European basketball, where 1 point is awarded for loss, and the teams are ranked exclusively by points.

I like the points ranking more. Any point gained is worth more than a million "up there for grabs".

 

[eternalbedhead]

If so, then why sort the teams during the season using a method in which a loss is equivalent to not playing when there are established methods in which a loss is factored in?

Because it's not really that big of a deal, and hockey, at least right now, has the pity point to be factored in.

 

[Nenikoj]

Ranking teams by P% rather than by actual points is a matter of taste. You can see the latter one also widely used, for example, in European soccer leagues (and it was used there when the point system was 2-1-0), or in chess tournaments.

Another example is European basketball, where 1 point is awarded for loss, and the teams are ranked exclusively by points.

I like the points ranking more. Any point gained is worth more than a million "up there for grabs".

How do you feel about the fact that when using a points system to sort the teams, a regulation loss is treated precisely the same as not playing. And since there are quite a lot of regulation losses, that's an awfully large number of game results which aren't factored in.

 

[Spade]

Looks like a thread griping about a problem that doesn't exist.

At the end of the year, the totals will even out. It's convenient for websites and newspapers with an established table at the beginning of the year to not have to reformat the table at the end (either designing a new table in the sports section or coding a new graphic on a website), which is why the standings you see midseason are organized the way it is. Even then, most people are able to think well enough to understand the concept of points percentage enough to understand that in a hypothetical scenario, a team that is 23-24-0 with 46 points is not performing better than a team that is 22-21-0 with 44 points.

With the MLB and NBA, the lack of OT losses and a wins-based system means that the media sites that follow those sports would have a "games back" table instead. Those are easily understood by their fanbases as well because the win and losses are clearly defined. You only have 2 possible results from a game.

Having "halfwins" and "halflosses" would be asinine and not intuitively understood to most people and just overcomplicates a very easy to follow table. Not to mention it's a lot more difficult to figure out how many actual wins a team had, which is even more important thanks to ROW being a major tiebreaker in the NHL. Does a team with 18 wins actually have 18 or did they lose 4 times in OT?

Solution to a problem that doesn't exist.

 

[Canadiens1958]

Loss

If so, then why sort the teams during the season using a method in which a loss is equivalent to not playing when there are established methods in which a loss is factored in?

Not playing is neutral, no other team involved. A loss is not neutral, another team was involved and enjoyed the benefit of a win.

 

[Canadiens1958]

Rational

Team A cannot unwin a game, and Team B has to win 3 out of 4.

In what rational sport would a participant, team or individual ever want to unwin? No perpetual mulligans allowed for the ideal win.

 

[Canadiens1958]

Point

My point is if you take the OTLs of any team and divvy them up as half a win and half a regulation loss, you'll get the same Games above 500, the same points percentage and the same points in the same # of games played. Example: Detroit is 17-13-7 (or 41 pts in 37 games with a .554 PPCT). Take the 7 OTLs and make each half a win and half a regulation loss and you get 20.5-16.5-0, which is also 41 pts in 37 games with a .554 PPCT). Make sense?


I find it somewhat ironic that one who supports the NHL's points system would criticize the percentage method for not conveying games played, when the NHL's points system also doesn't convey games played.

Points percentage takes into consideration 100% of a teams played games (see the 17-13-7 example above). Points earned ignores losses.


No team can "unlose" a game. In your example, Team A is one game ahead of Team B. Similar thing would be true if you reversed the records around to have 9W-20L and 8L-17W. 8L-17W would be ahead by one game.

The standings have to reflect what actually happened not similarities. What you suggest did not actually happen in the arena. So it is fantasy based.

 

[morehockeystats]

How do you feel about the fact that when using a points system to sort the teams, a regulation loss is treated precisely the same as not playing. And since there are quite a lot of regulation losses, that's an awfully large number of game results which aren't factored in.

I don't feel about it.
If two teams have the same number of points, let the tie-breaks be the number of games played.
If they have a different number of points, it's more important to have the points in hand, not the games in hand - in my opinion.

 

[Canadiens1958]

Interesting

I don't feel about it.
If two teams have the same number of points, let the tie-breaks be the number of games played.
If they have a different number of points, it's more important to have the points in hand, not the games in hand - in my opinion.

Interesting observation.

Still comes down to the situation. Which alternative allows a team to control its own destiny without relying on a third party to win.

 

[Nenikoj]

Looks like a thread griping about a problem that doesn't exist.

At the end of the year, the totals will even out.

Given that, why would it be a problem for the NHL to use a percentage and/or games behind method? Remember, as a tie (or OTL) is the net equivalent of half a win and half a regulation loss, the presence of ties (or OTLs) cannot be the reason a percentage and/or games behind method for sorting the teams wouldn't work.

It's convenient for websites and newspapers with an established table at the beginning of the year to not have to reformat the table at the end (either designing a new table in the sports section or coding a new graphic on a website), which is why the standings you see midseason are organized the way it is. Even then, most people are able to think well enough to understand the concept of points percentage enough to understand that in a hypothetical scenario, a team that is 23-24-0 with 46 points is not performing better than a team that is 22-21-0 with 44 points.

This is true if you're looking at all elements of the standings. But if all one has time for is "who trails who", then the fan will be misled into thinking the 23-24-0 team is better off than the 22-21-0 team.

With the MLB and NBA, the lack of OT losses and a wins-based system means that the media sites that follow those sports would have a "games back" table instead. Those are easily understood by their fanbases as well because the win and losses are clearly defined. You only have 2 possible results from a game.

The MLB and NBA look at 100% of the games played to sort the teams. The NHL looks at roughly 40% of results, ignoring the regulation losses. With respect to sorting teams in the standings, regulation losses are treated precisely the same as not playing. Is that something you like?

Having "halfwins" and "halflosses" would be asinine and not intuitively understood to most people and just overcomplicates a very easy to follow table. Not to mention it's a lot more difficult to figure out how many actual wins a team had, which is even more important thanks to ROW being a major tiebreaker in the NHL. Does a team with 18 wins actually have 18 or did they lose 4 times in OT?

Who said anything about displaying the standings showing half wins or half losses? My suggestion is to display what they currently display, but get rid of the points column and replace it with a points percentage column and games behind column and sort by the PPCT column.

Solution to a problem that doesn't exist.

Standings sorts in which nearly half the results aren't considered is not a problem? Please explain.

 

[Nenikoj]

Not playing is neutral, no other team involved. A loss is not neutral, another team was involved and enjoyed the benefit of a win.

Correct it is not neutral. But with respect to the primary method of sorting the teams in the standings, a regulation loss is no different from not playing. Thus, some 40% of the results of teams are completely disregarded when it comes to sorting the teams. Do you like it that way? If so, why?

Example:
Detroit is 17-13-7 (or 41 pts). They then lose five in a row to become 17-18-7 (or 41 pts). Note that their pts total - or what is used to sort them in the standings - hasn't changed one iota. Are they any worse off at 17-18-7 than they were at 17-13-7? If so, why wouldn't you want that reflected in the way teams are sorted in the standings?

 

[Nenikoj]

The standings have to reflect what actually happened not similarities. What you suggest did not actually happen in the arena. So it is fantasy based.

My proposal of using PPCT to sort teams DOES reflect what actually happens.

 

[Canadiens1958]

Point Totals

Correct it is not neutral. But with respect to the primary method of sorting the teams in the standings, a regulation loss is no different from not playing. Thus, some 40% of the results of teams are completely disregarded when it comes to sorting the teams. Do you like it that way? If so, why?

Example:
Detroit is 17-13-7 (or 41 pts). They then lose five in a row to become 17-18-7 (or 41 pts). Note that their pts total - or what is used to sort them in the standings - hasn't changed one iota. Are they any worse off at 17-18-7 than they were at 17-13-7? If so, why wouldn't you want that reflected in the way teams are sorted in the standings?

Point totals are not limited to one team. So five losses resulted in wins for other teams. Plus you have the results over the same time span in the rest of the league that also produced wins.

So winning teams would move ahead of or approach the losing teams in a point based system. Trackable across conferences, divisions and within. Reflective of league play as a whole.