P/60 stats: How useful and reliable are they? Does production increase linearly with ice time?

[Nithoniniel]

In general, a better alternative than most. As so often when discussing stats, a large portion of the debate seem focused on how far from perfect it comes. I've always found that weird, because it doesn't need to be perfect. It just needs to be better than the alternative. The positives it has compared to the alternative just needs to outweigh the negatives. In this case, they do. That doesn't mean that you shouldn't consider such things on an individual level.

I'm a fan of exploratory statistical analysis. To not just look at some stats and draw some conclusions, but asking why the numbers are the way they are. McDavid is without doubt a premiere offensive performer in this league. So why isn't he higher in P/60? Well, it could be that he's used a ton in games where the Oilers are chasing. Bad teams chase more, when they do they want their best players on the ice more, these players are less likely to have produced already given that it's a losing effort.

I'm pretty sure that the stats support that when a player plays much more than his norm in a game, he has a worse pace than otherwise, and vice versa. It's the same reason why a lot of offensive superstars have a worse record in games they play a lot. It's not because the team is worse off with them on the ice. It's because if they've already had a great game, the team is less likely to have to chase more goals.

 

[Machinehead]

It's not an automatic "this guy will score more/less" but it is very useful.

I think it's funny that people criticize it as far as being used out of context when it literally IS context.

 

[zeke]

Unfortunately people on this site do not know how to properly use statistics in the right context or how to properly pair stats with other stats to give any sort of meaningful value.

funny, I see it used properly on this site all the time.

I find a much bigger problem on this site is that it is CRITICIZED improperly, tbh.

 

[zeke]

I think it's funny that people criticize it as far as being used out of context when it literally IS context.

well said.

 

[McDNicks17]

It(along with important contextual stats) can tell you how well a player is doing in their current role/context.

It has basically no predictive value on how well a player will do in a different role/context though.

 

[Pierce Hawthorne]

funny, I see it used properly on this site all the time.

I find a much bigger problem on this site is that it is CRITICIZED improperly, tbh.

That's not true. At all. 90% of the time it's used by itself and used to suggest that "player x plays less minutes but has a higher p/60 then player y, so give player x more minutes and he'd produce more"


This is absolutely not true at all by any means and is exactly why it gets used incorrectly.

 

[McFlyingV]

It's not an automatic "this guy will score more/less" but it is very useful.

I think it's funny that people criticize it as far as being used out of context when it literally IS context.

It is context in the sense that it shows how much players score per 60 minutes that they actually played. It does not, however, consider context of how much a player plays on a nightly basis, and that is a major flaw when you try to compare one player with another who plays considerably different minutes on a nightly basis. If they have similar usage, I don't really see how you can criticize P/60.

 

[McFlyingV]

That's not true. At all. 90% of the time it's used by itself and used to suggest that "player x plays less minutes but has a higher p/60 then player y, so give player x more minutes and he'd produce more"


This is absolutely not true at all by any means and is exactly why it gets used incorrectly.

This could not be more accurate.

 

[CantLoseWithMatthews]

That's not true. At all. 90% of the time it's used by itself and used to suggest that "player x plays less minutes but has a higher p/60 then player y, so give player x more minutes and he'd produce more"


This is absolutely not true at all by any means and is exactly why it gets used incorrectly.

i've never seen anyone suggest that

 

[zeke]

That's not true. At all. 90% of the time it's used by itself and used to suggest that "player x plays less minutes but has a higher p/60 then player y, so give player x more minutes and he'd produce more"


This is absolutely not true at all by any means and is exactly why it gets used incorrectly.

we must be involved in different conversations then.

most every discussion using it that I see is either comparing players in similar roles or goes out of its way to point if a player's impressive p60 might be due to sheltered cozy role.

 

[Pierce Hawthorne]

i've never seen anyone suggest that

Give me a break.

 

[McFlyingV]

i've never seen anyone suggest that

You've gotta be kidding me. I've seen you in threads where its been used in that exact way.

 

[McFlyingV]

we must be involved in different conversations then.

most every discussion using it that I see is either comparing players in similar roles or goes out of its way to point if a player's impressive p60 might be due to sheltered cozy role.

You've literally used it for players in different roles on more than one occasion.

 

[Pierce Hawthorne]

You've gotta be kidding me. I've seen you in threads where its been used in that exact way.

No kidding. He has to be trolling at this point.

 

[CantLoseWithMatthews]

You've gotta be kidding me. I've seen you in threads where its been used in that exact way.

okay well I'll give you the benefit of the doubt and agree that it's wrong to use it as a predictive stat. that doesn't take away its actual value though

 

[McFlyingV]

okay well I'll give you the benefit of the doubt and agree that it's wrong to use it as a predictive stat. that doesn't take away its actual value though

And I don't disagree. I think its a very good stat when you're looking at what players have done in their current role.

 

[zeke]

You've literally used it for players in different roles on more than one occasion.

I make every effort possible to compare players in similar roles, and point when they have different roles and exactly why they are different.

tbh, it's the "but POINTS" crowd that seems much less interested in using context or considering players' usage and roles. everyone who uses p60 seems to make every effort to decipher each player's role and it's effects.

 

[zeke]

You've gotta be kidding me. I've seen you in threads where its been used in that exact way.

you might want to go re-read those threads.

 

[Del Preston]

It is not perfect but I like looking at it to see how a player does with/without a specific teammate.

 

[McFlyingV]

I make every effort possible to compare players in similar roles, and point when they have different roles and exactly why they are different.

tbh, it's the "but POINTS" crowd that seems much less interested in using context or considering players' usage and roles. everyone who uses p60 seems to make every effort to decipher each player's role and it's effects.

That is simply not true. I've seen you make posts comparing P/60 of players with different roles, i.e. McDavid vs. Matthews threads. There's nothing wrong with P/60 as long as you recognize a player playing 21 minutes a night vs one playing 16 or 17 minutes can hardly be compared in a logical way.

For example, lets say player x has a 3.5P/60 playing 21 minutes a night while player y has a 3.4P/60 playing 16 minutes a night. You could say player y produced similarly in a lesser role, but you could not conclude that player y would produce similarly in player x's role, nor could you conclude that player y is close in calibre to player x.

 

[KevinRedkey]

i've never seen anyone suggest that

 

[Pierce Hawthorne]

P/60 does nothing to account for either A) The difficulty of minutes players are getting, B) The type of minutes a player is getting(In the form of PP, ES, or PK minutes), or C) The type of minutes a player is getting(In the form of extra minutes while trailing or less minutes with a lead, etc).



For example. Player X and Player Y may play the same 18 minutes a game. While Player X may have a 2.5 P/60 in said role, while Player Y has a 2.15 P/60 getting the same 18 minutes a game. Without looking at any other statistic it would be very easy to infer that Player X is creating more offense and producing better then what Player Y is. But it doesn't account for any other factor. Such as how that ice time is distributed. What if Player X is getting 15 minutes of ES Ice time a game, and 3 minutes of PP Ice time a game but 0 minutes of PK time. While Player Y is on the opposite of the spectrum getting 0 minutes of PP time and 3 minutes of PK time. P/60 by itself does not show those differentiations. Now obviously ES P/60 would eliminate the PP/PK differences to some extent, but it still doesn't do it entirely. It doesn't account for the difficulty of the extra minutes played and how it would effect a player playing at Even Strength.


P/60 also doesn't account for the quality of competition either. Player X in the example above might be getting 15/18 minutes a game feeding on other teams 2nd or 3rd pairing Defenders, while Player Y is spending 15/18 of his even strength minutes matching up against the other teams Top pairing. Not only does this have an effect on the two players current P/60 stats, but it would also have an applied effect to increased minutes as well. Since Player X getting increased minutes would certainly mean he matches up again top pairings more often in those extra minutes, likely driving his P/60 down. While Player Y in those extra minutes would get some softer minutes and the opportunity to actually improve his P/60 with more minutes.


There's other factors as well. Such as where shifts are being started. A player who gets 18 minutes of ice time but starts 60% of those shifts in the defensive zone is likely to have a lower P/60 then the same 18 minute player getting 60% of his shifts in the offensive zone.



The long story short is it can be a very useful tool when used properly. If its used in the right context and with the support of other stats I think it would be extremely capable of forecasting future expectations for players and the expected production of those players. But by itself I dont consider it to be very useful at all. A step up from +/- which is probably the single most useless stat in hockey, but still not one to use in arguments without context.

 

[CantLoseWithMatthews]

P/60 does nothing to account for either A) The difficulty of minutes players are getting, B) The type of minutes a player is getting(In the form of PP, ES, or PK minutes), or C) The type of minutes a player is getting(In the form of extra minutes while trailing or less minutes with a lead, etc).



For example. Player X and Player Y may play the same 18 minutes a game. While Player X may have a 2.5 P/60 in said role, while Player Y has a 2.15 P/60 getting the same 18 minutes a game. Without looking at any other statistic it would be very easy to infer that Player X is creating more offense and producing better then what Player Y is. But it doesn't account for any other factor. Such as how that ice time is distributed. What if Player X is getting 15 minutes of ES Ice time a game, and 3 minutes of PP Ice time a game but 0 minutes of PK time. While Player Y is on the opposite of the spectrum getting 0 minutes of PP time and 3 minutes of PK time. P/60 by itself does not show those differentiations. Now obviously ES P/60 would eliminate the PP/PK differences to some extent, but it still doesn't do it entirely. It doesn't account for the difficulty of the extra minutes played and how it would effect a player playing at Even Strength.


P/60 also doesn't account for the quality of competition either. Player X in the example above might be getting 15/18 minutes a game feeding on other teams 2nd or 3rd pairing Defenders, while Player Y is spending 15/18 of his even strength minutes matching up against the other teams Top pairing. Not only does this have an effect on the two players current P/60 stats, but it would also have an applied effect to increased minutes as well. Since Player X getting increased minutes would certainly mean he matches up again top pairings more often in those extra minutes, likely driving his P/60 down. While Player Y in those extra minutes would get some softer minutes and the opportunity to actually improve his P/60 with more minutes.


There's other factors as well. Such as where shifts are being started. A player who gets 18 minutes of ice time but starts 60% of those shifts in the defensive zone is likely to have a lower P/60 then the same 18 minute player getting 60% of his shifts in the offensive zone.



The long story short is it can be a very useful tool when used properly. If its used in the right context and with the support of other stats I think it would be extremely capable of forecasting future expectations for players and the expected production of those players. But by itself I dont consider it to be very useful at all. A step up from +/- which is probably the single most useless stat in hockey, but still not one to use in arguments without context.

wow, this sure is a whole lot of not understanding the stat

1) People can and do take QoC and QoT into account
2) People can and do break p/60 down by situation. it's people who look at raw point totals who tend not to actually
3) Zone starts don't really have an impact, but you can take that into account too if you really want

the irony is of course the people who cry for more context are dismissing a stat like P/60 that applies context to production. It's literally nothing like +/-, nor is it meant to be used to forecast future production. Weren't you just complaining about people doing that?

 

[zeke]

That is simply not true. I've seen you make posts comparing P/60 of players with different roles, i.e. McDavid vs. Matthews threads. There's nothing wrong with P/60 as long as you recognize a player playing 21 minutes a night vs one playing 16 or 17 minutes can hardly be compared in a logical way.

For example, lets say player x has a 3.5P/60 playing 21 minutes a night while player y has a 3.4P/60 playing 16 minutes a night. You could say player y produced similarly in a lesser role, but you could not conclude that player y would produce similarly in player x's role, nor could you conclude that player y is close in calibre to player x.

But Matthews and mcdavid have extremely similar roles, both being elite quality of competition #1 centers. Mcadavid has probably even had better linemates overall but close enough that I don't care.

the difference in their roles is purely ice time related, not toughness of minutes - in other words, this is the PERFECT CONTEXT in which to use p60. there couldn't be a more appropriate time to use it. that doesn't mean it's a perfect solution, but it is exactly the right time to use the stat.

 

[93LEAFS]

Like all stats, it needs to be nuanced. No stat is the be all and end all, especially in hockey. Although, P/60 and P1/60 catch some important stuff, but it isn't perfect. Inflated shooting percentages, where the player is getting less ice-time (for example, is someone getting full powerplay minutes and then sheltered at ES, etc). Since this appears to mostly be related to Matthews (and where NHL.com ranked him), lets acknowledge some things. One, Matthews over 2 seasons (not that small a sample size) has been elite in this metric. He isn't sheltered by zone-starts or by competition. He sees less PP time compared to most elite centers. His P/60 and p1/60 at 5v5 this year is probably out of whack, but his totals over 2 years are probably quite representative of the player he is (had a high-oiSH% this year, and a relatively low one for an elite player in 2016/17). Even if you want to remove rates, his 5v5 production in general terms since he entered the league is elite.