Points vs ppg vs points per 60 mins toi

[BoredBrandonPridham]

5v5 p/60 & PP p/60 matches up better with the eye test. it's literally how effective the player produces offence when they are deployed by their coach in the situations they are deployed in, the teammates they are deployed with, and the competition they are deployed against.

no one watches the game and honestly thinks "wow <player> is invisible out there, not very good" when he is injured, or when he is on the bench between shifts. when the player is not on the ice, they are not being evaluated, and p/60 reflects that.

whether you want to use points or p/60 really depends on what the question is. if the question is which player has been most effectively offensively in the situations that the coach has deployed them in, then that is p/60. if you want to ask "who won the art ross?", then you want to look at points. not sure what other question you'd want to look at raw points for, though, unless it's an era where /60 stats aren't available.

if you're a team looking for effective offensive players to acquire or sign who may be undervalued by their team, you could do a lot worse than triaging your pro scouting efforts by sorting natural stat trick by 5v5 p/60 and look for some 3rd liners with exceptionally high rates of production.

 

[Dekes For Days]

If you want to change context of the original post, fine,
But you should say, yes that’s right but what about these points.
Instead of just replying your wrong

The context of the original post involved that individual attempting to suggest that Matthews had a better 5v5 season previously, according to per-60 metrics, when that was false. This while incorrectly addressing his own original question, "Which version of Matthews was better?". If his "higher in other seasons" comment was referring to exclusively rank, then I guess we could say "extremely misleading" instead.

 

[Golden_Jet]

5v5 p/60 & PP p/60 matches up better with the eye test. it's literally how effective the player produces offence when they are deployed by their coach in the situations they are deployed in, the teammates they are deployed with, and the competition they are deployed against.

no one watches the game and honestly thinks "wow <player> is invisible out there, not very good" when he is injured, or when he is on the bench between shifts. when the player is not on the ice, they are not being evaluated, and p/60 reflects that.

whether you want to use points or p/60 really depends on what the question is. if the question is which player has been most effectively offensively in the situations that the coach has deployed them in, then that is p/60. if you want to ask "who won the art ross?", then you want to look at points. not sure what other question you'd want to look at raw points for, though, unless it's an era where /60 stats aren't available.

if you're a team looking for effective offensive players to acquire or sign who may be undervalued by their team, you could do a lot worse than triaging your pro scouting efforts by sorting natural stat trick by 5v5 p/60 and look for some 3rd liners with exceptionally high rates of production.

Teams don’t use natural stat trick, last I read over a year ago was 30 of 31 teams pay use sportlogiq.com as do a lot of Pro/Amateur teams in hockey, football and International football (soccer).
They could use other platforms as well.

 

[tucker3434]

Points. The others are used to to project hopeful future points. If they never get there with actual production, it doesn't matter. Doesn't mean the others aren't useful though.

 

[Sidney the Kidney]

whether you want to use points or p/60 really depends on what the question is. if the question is which player has been most effectively offensively in the situations that the coach has deployed them in, then that is p/60. if you want to ask "who won the art ross?", then you want to look at points. not sure what other question you'd want to look at raw points for, though, unless it's an era where /60 stats aren't available.

So would you say that Jason Robertson has been "more effective offensively" than Auston Matthews so far this season? And Spezza has been the Leafs' true MVP because he's been "more effectively offensively"?

The problem with what you're doing above, dismissing the value of raw points in a discussion of how well a player is performing offensively, in favor of per 60 rates is you'd end up drawing the above conclusions.

And I don't think anyone would argue that either Robertson or Spezza have been better than Matthews this season.

if you're a team looking for effective offensive players to acquire or sign who may be undervalued by their team, you could do a lot worse than triaging your pro scouting efforts by sorting natural stat trick by 5v5 p/60 and look for some 3rd liners with exceptionally high rates of production.

This I agree with. P/60 is a useful tool to find diamonds in the rough or under-utilized players who may have more potential if only given a bit more ice time. I just don't think it's as cut-and-dried as a useful tool when comparing two players who are already getting the same sort of usage.

 

[Yackiberg8]

I think the problem with this approach is it's assuming one just chooses *one* of the options without then evaluating further after making that choice to see if there are extenuating circumstances that lead to those total points not accurately detailing who is "better".

For instance, I'll give myself as an example. Of the three stats listed in the OP I'd favor total points. However, my evaluation doesn't just stop there. It's more of a step by step process, whereby the first "step" would be me comparing two players by first comparing them by point totals. But after that, I'd look at the "why". Is there a reason one player has less points than the other (ie. injured half the year, anomalous down year with zero puck luck, etc.)?

So in your above example, even though I favor total points, before I'd settle on a final answer for who is better, I'd see that the reason Cooke has more points than Crosby is because Crosby missed most of the season. I'd see that the reason Domi outscored McDavid is because McDavid missed significant time with injury.

The situations in which I'd favor total production is when the two players being compared are in similar situations (ie. their games played are similar, their usage by their team is similar, etc.). But that doesn't mean I'd apply that to situations where there are other factors for why raw points are different.

That's a reasonable approach and I agree that none of the ways listed are perfect on their own. Just pointing out to the "you don't get credit for imaginary games" crowd how their view is not smart.

I prefer Pts/GP as it is easy to compute and gives you a quick idea of a players production. Of course there are many factors that go into it but I don't get the "only total points matter" crowd.

 

[Yackiberg8]

How about Lupul with 46 points in 56 games versus Gallagher with 47 in 82 or Taylor Hall with 37 points in 33 games vs Sean Couturier with 37 points in 82 games?

Not sure what you're getting at here.

My examples were from the same season, you have cherry picked a Joffery Lupul season when Brendan Gallagher was 3 years away from even being drafted and a Taylor Hall 2018-19 season vs a Sean Couturier 2014-15 season?

 

[BoredBrandonPridham]

So would you say that Jason Robertson has been "more effective offensively" than Auston Matthews so far this season? And Spezza has been the Leafs' true MVP because he's been "more effectively offensively"?

The problem with what you're doing above, dismissing the value of raw points in a discussion of how well a player is performing offensively, in favor of per 60 rates is you'd end up drawing the above conclusions.

And I don't think anyone would argue that either Robertson or Spezza have been better than Matthews this season.



This I agree with. P/60 is a useful tool to find diamonds in the rough or under-utilized players who may have more potential if only given a bit more ice time. I just don't think it's as cut-and-dried as a useful tool when comparing two players who are already getting the same sort of usage.

I haven’t looked, but if their 5v5 p/60 is higher than someone like matthews, then yes they have been more effective offensively in the role they were put in. It’s the literal definition. If a player is deployed for one minute and happens to help produce a point, then sees no more ice time, then that player player was quite effective in the role in which he was deployed.

Don’t confuse that with who the better offensive player is though. If you want to compare 2 players to see who is the better offensive player, you need to take a lot of other things into account.

Total points is ok to try and combine a lot of things, like in what manner the coach deploys the player, what kind of ice time they get, what type of competition they play against, it’s a decent proxy, but just like p/60 you’ll have outliers. p/60 just removes a lot of noise from production, which is more useful than a noisy catch-all of factors.

 

[c3z4r]

Not sure what you're getting at here.

My examples were from the same season, you have cherry picked a Joffery Lupul season when Brendan Gallagher was 3 years away from even being drafted and a Taylor Hall 2018-19 season vs a Sean Couturier 2014-15 season?

I was unaware of your same season criteria. So here's same season examples:


- Mats Zucarello 40 points in 48 games versus Alex Killorn 40 points in 82 games

- Taylor Hall 37 in 33 versus Andrei Svechnikov 37 in 82

- Jack Eichel 57 in 61 versus Giroux 58 in 82 or Ryan Kesler 58 in 82

- Conor Sheary 53 points in 61 games versus Patrice Bergeron 53 in 79 or Nathan Mackinnon 53 in 82

 

[Yackiberg8]

I was unaware of your same season criteria. So here's same season examples:


- Mats Zucarello 40 points in 48 games versus Alex Killorn 40 points in 82 games

- Taylor Hall 37 in 33 versus Andrei Svechnikov 37 in 82

- Jack Eichel 57 in 61 versus Giroux 58 in 82 or Ryan Kesler 58 in 82

- Conor Sheary 53 points in 61 games versus Patrice Bergeron 53 in 79 or Nathan Mackinnon 53 in 82

I’m not sure what point you’re trying to make.

 

[c3z4r]

I’m not sure what point you’re trying to make.

Points over PPG

 

[Sidney the Kidney]

I haven’t looked, but if their 5v5 p/60 is higher than someone like matthews, then yes they have been more effective offensively in the role they were put in. It’s the literal definition. If a player is deployed for one minute and happens to help produce a point, then sees no more ice time, then that player player was quite effective in the role in which he was deployed.

Yes, they are ahead of Matthews (or were prior to tonight. Not sure if his 2 goal game changes that). That was why I mentioned both of them.

Don’t confuse that with who the better offensive player is though. If you want to compare 2 players to see who is the better offensive player, you need to take a lot of other things into account.

But that's the problem. Often times that's exactly what people who use P/60 are trying to argue. That if you compare two players, P/60 does a better job of telling you who the "better offensive player" is between the two compared to raw points.

IMO, that's flawed and I think the Matthews/Spezza/Robertson example is the perfect example of this. If one were to use P/60 to determine who was the best offensive player of those three, one would conclude Matthews was dead last. But if one were to use raw points to determine who has been the best offensive player of those three, one would conclude it's Matthews by a mile. And I'd suggest that's exactly how reality has been.

 

[BoredBrandonPridham]

Yes, they are ahead of Matthews (or were prior to tonight. Not sure if his 2 goal game changes that). That was why I mentioned both of them.



But that's the problem. Often times that's exactly what people who use P/60 are trying to argue. That if you compare two players, P/60 does a better job of telling you who the "better offensive player" is between the two compared to raw points.

IMO, that's flawed and I think the Matthews/Spezza/Robertson example is the perfect example of this. If one were to use P/60 to determine who was the best offensive player of those three, one would conclude Matthews was dead last. But if one were to use raw points to determine who has been the best offensive player of those three, one would conclude it's Matthews by a mile. And I'd suggest that's exactly how reality has been.

Using p/60 as part of why a player is a better offensive player is not worse than using points. Both siffer from outliers.

Just like when you claimed OV peaked as a goal scorer in his first 5 seasons in the league, you were obviously wrong, and 5v5/PP g/60 makes it very clear.

 

[BoredBrandonPridham]

Teams don’t use natural stat trick, last I read over a year ago was 30 of 31 teams pay use sportlogiq.com as do a lot of Pro/Amateur teams in hockey, football and International football (soccer).
They could use other platforms as well.

 

[Golden_Jet]

Lol, no idea who that is, or what it even means, or if anyone cares.

 

[trentmccleary]

PPG.

Points are used when somebody wants to convince me that a soft, one-dimensional, iron man was better than a PPG player who gets hurt playing defense, blocking shots, working the boards, etc.

Points per 60 is what is used when somebody wants to convince me that their 29 yo, 13th forward is actually a crouching tiger, ready to take the league by storm.

 

[DitchMarner]

They're different stats that have different meanings and values. I'm not sure you can say one is generally "better" or more reflective of a player's offensive ability than the others.

A player who scores a lot of points may benefit from a ton of PP TOI and have strong ES linemates compared to similarly talented players.

A player with a very high PPG in 35 games may have an unsustainable shooting percentage that wouldn't hold up over 82 games.

A player who scores at a very high clip per minute at ES or five on five may benefit from having teammates on other lines drawing the toughest defensive assignments and may not be able to score at a similar clip playing higher in the lineup and averaging more TOI.


None of these stats is perfect and each one generally requires more context.

 

[Hockey Outsider]

In terms of accuracy from year to year, I'd suggest points per game seems to have less variance than points per 60, at least for players who maintain the same role from season to season (ie. a 1st liner is always a 1st liner).

I looked at ES production for forwards over a span of 10 seasons. I had something like 3,000 player-seasons in the population. The data was clear - points per game has a stronger correlation year-over-year compared to points per 60.

Neither correlation is overly strong (meaning that there's a lot of randomness in hockey - not surprising) but points per game was the better of the two. I'll have to dig up the file so I can post exactly what the parameters were.

 

[Sidney the Kidney]

I looked at ES production for forwards over a span of 10 seasons. I had something like 3,000 player-seasons in the population. The data was clear - points per game has a stronger correlation year-over-year compared to points per 60.

Neither correlation is overly strong (meaning that there's a lot of randomness in hockey - not surprising) but points per game was the better of the two. I'll have to dig up the file so I can post exactly what the parameters were.

Imagine that.

 

[Dekes For Days]

If one were to use P/60 to determine who was the best offensive player of those three, one would conclude Matthews was dead last.

Using per-60 metrics in the proper ways they have always been used, nobody would conclude that, based on your small sample result in one type of production, for wildly different players in wildly different roles with wildly different underlying metrics, in one isolated game state.

As I'm sure you're aware, you're just using this statistic horribly wrong, despite the fact that proper usage has been shown and explained to you literally dozens of times, only to fall on deaf ears.

 

[Dekes For Days]

So would you say that Jason Robertson has been "more effective offensively" than Auston Matthews so far this season?

If one were to use P/60 to determine who was the best offensive player of those three, one would conclude Matthews was dead last.

Jason Robertson is actually below Matthews in 5v5 P/60, but using this same logic with the stat form you like to use, P/GP...

Jason Robertson (your chosen player) has a higher 5v5 (your chosen game state, used in isolation, like you did) P/GP (your chosen stat form) than...

Malkin
Pastrnak
Kane
Ovechkin
Barkov
Barzal
Huberdeau
Pettersson
Tavares
Couturier
Point
Aho
Zibanejad
Scheifele
...Etc...

To be consistent with what you said above, if we were to "use P/GP to determine who was the best offensive player" of all of these, would we conclude that Robertson was the best? Is that conclusion true, or, using the same logic, is P/GP also bad according to you?

Should we do raw points next? Or... crazy thought I know, but it's almost like all statistics, regardless of form, require some form of context, and we shouldn't use any stat in one game state in a small isolated sample to determine something like "best offensive player"... Hmm...

 

[TopShelfYzerman]

I discount p/60 metrics for the elite offensive forwards, Full Stop.

Because that would ruin your whole Crosby is best narrative..

 

[TopShelfYzerman]

So Matthews is 9th in p/60 this season NHL Stats

He was 3rd in a previous season.

Which version of Matthews was better?

What season?

 

[ijuka]

Points per 60mins is a useless trash statistic. The less ice time you get, the easier it is to have a high stat, but it's not going to increase linearly, aka you can't just give a 10min/game 4th liner 20min of ice time and have their scoring double. Their scoring might even decrease due to them having less energy etc. Shame on anyone voting for it.

I voted for PPG over points, but that's situational too. Playing 10 games and scoring 15 points isn't as impressive as playing 82 and scoring 110. Overall, though, it's the better indicator.

 

[Rants Mulliniks]

Yes, they are ahead of Matthews (or were prior to tonight. Not sure if his 2 goal game changes that). That was why I mentioned both of them.



But that's the problem. Often times that's exactly what people who use P/60 are trying to argue. That if you compare two players, P/60 does a better job of telling you who the "better offensive player" is between the two compared to raw points.

IMO, that's flawed and I think the Matthews/Spezza/Robertson example is the perfect example of this. If one were to use P/60 to determine who was the best offensive player of those three, one would conclude Matthews was dead last. But if one were to use raw points to determine who has been the best offensive player of those three, one would conclude it's Matthews by a mile. And I'd suggest that's exactly how reality has been.

That's actually a terrible example. It's comparing apples to oranges. Most people referencing P60 are comparing apples to apples and suggesting similar usage will produce similar raw results. Then people are amazed when it happens.