Top Offensive Players Since 2005, Using an Objective Scoring System

[Trocity]

Crosby's better.

 

[666]

I only am willing to take this goal v assist thing so far. That said, last season 2 players had 50 goals. 32 players had 50 assists.

3 players with 50 primary assists.

Goals 51,50,47
Primary assists 56,53,52

2017-18. 0 players with 50 goals. 29 players with 50 assists.

zero players with primary assists.

Goals 49,45,43
Primary assists 48,41,40

 

[Muikea Bulju]

Bud, you're not using the word "objective" correctly. You're giving us your subjective opinion on an arbitrary system.

Well, you are not using it correctly.

The current points-system is not objective, not even close.

It is a completely subjective valuation

If people want "points" to come even close of portraying someone offensive output, different goals and different assists need to be given different values.

 

[Muikea Bulju]

The correlation between 5v5 individual primary assist rate and on-ice 5v5 team GF/60 is higher than that same correlation between goal rate and team GF/60.

So you want to give primary assists the most weight?

No. I never said that any stat should be given a fixed value / weight

The best option would be to just have ja panel of judges deciding on different "points values" different players get on each goal

This, of course, would be in addition to the current system, which would be maintained, in order to compare modern guys to old players.

 

[Regal]

This occurred to me as I was reviewing the numbers. I read up on identical twins. The assumption that the Sedins must be of equal athletic prowess is simply flawed - there are lots of cases of identical twins where one becomes the better athlete despite identical training.

No shortage of research on the subject:

BBC Sport - Are sporting twins made equal?

Take Jose and Ozzie Canseco for example. Jose was one of the best players in the world. Ozzie couldn't break out of the minors.

And anyone who watched them regularly will tell you that they were near equals, and if anything, Henrik was the more valuable player

 

[KBobs]

The problem with the logic in each of these posts is this: I didn't create the formula. A Mario Lemieux fan did. I merely applied it to players from 2005 to present, and I've given you everything you need to check the math and logic. It's not my objectivity that is on the line here.

Surely some of you can do better than ad hominems:

You ran with it knowing full well what the implications would be. If Ovechkin was down around, say #50, you'd dismiss it entirely and this thread would cease to exist. Simply put, there is no financial model that is entirely objective, as all the inputs/assumptions are entirely subjective. It's why all these fancy models that come up with a numerical value to assess players are a complete joke. Anyone who swears by something this silly has no idea what they're talking about with regards to financial modelling yet probably thinks they're more clever than the field.

But continue to state how you have found no valid arguments against it...

 

[Midnight Judges]

And anyone who watched them regularly will tell you that they were near equals, and if anything, Henrik was the more valuable player

What are the chances anyone formed that opinion without being heavily influenced by the traditional scoring system?

I'd say near zero.

 

[Midnight Judges]

You ran with it knowing full well what the implications would be.

That people would disagree? That some people would dislike the results? Sure, I knew that.

As a matter of principle I don't alter opinions or facts or data or methodology in an effort to appeal to the masses. Certainly I would be more popular if I did.

If Ovechkin was down around, say #50, you'd dismiss it entirely and this thread would cease to exist.

Or maybe I'd still post it precisely the same way simply because I think there is a more equitable way to account for scoring in the NHL.

Simply put, there is no financial model that is entirely objective, as all the inputs/assumptions are entirely subjective. It's why all these fancy models that come up with a numerical value to assess players are a complete joke. Anyone who swears by something this silly has no idea what they're talking about with regards to financial modelling yet probably thinks they're more clever than the field.

But continue to state how you have found no valid arguments against it..

This isn't akin to financial modeling where there are a bazillion unknown significant factors and the choices for inputs are tantamount to assuming the outputs. The link in the OP makes this clear.

 

[txpd]

3 players with 50 primary assists.

Goals 51,50,47
Primary assists 56,53,52





zero players with primary assists.

Goals 49,45,43
Primary assists 48,41,40

And? Unless I am wrong a secondary assists counts the same as a primary. Right?

 

[Regal]

What are the chances anyone formed that opinion without being heavily influenced by the traditional scoring system?

I'd say near zero.

No it's painfully obvious even if you had no idea what their points were

 

[Regal]

The idea of the traditional point system being flawed is fair. The systems used to determine the value of each are all bunk though. Tying value to repeatability continues to be inherently flawed. The totals are already too subjective to variance because of how low they are, and since goals will always be a higher number than either primary or secondary assists, they will always be more repeatable. There's also no context to star players vs. complimentary players or who someone plays with. Defensemen and complimentary forwards are going to be more prone to variance due to the players they play with than elite star players. Frankly, I don't know how anyone can be a hockey fan and believe that primary assists are worth roughly half of a goal for elite playmakers. It doesn't pass the smell test.

 

[BoredBrandonPridham]

If we were comparing top tier players to scrubs who never see PP time, you'd have a stronger point

I think you’re vastly underestimating the impact PP production has on overall production at the scale of PP time variability brought in solely by game management effects and team/coaching effects.

It’s not just no PP time vs PP time, it’s like 200 mins of PP1 time per season vs 450mins of PP1 time per season across teams and eras.

 

[Midnight Judges]

The systems used to determine the value of each are all bunk though. Tying value to repeatability continues to be inherently flawed. The totals are already too subjective to variance because of how low they are, and since goals will always be a higher number than either primary or secondary assists, they will always be more repeatable.

If the sample data set is large enough for goals but insufficient for primary assists or secondary assist, then I think you'd have a point. But if the data sets are large enough, the impacts on persistence due to the quantity of data ought to be minimal.

 

[Midnight Judges]

I think you’re vastly underestimating the impact PP production has on overall production at the scale of PP time variability brought in solely by game management effects and team/coaching effects.

It’s not just no PP time vs PP time, it’s like 200 mins of PP1 time per season vs 450mins of PP1 time per season across teams and eras.

I don't deny that PP time is significant, but this thread is about why new valuations of 1/.54/.1 are superior to simply assuming 1-1-1.

The traditional 1-1-1 system does nothing to address your point either.

 

[n00bxQb]

Most Canucks fans will tell you Henrik is a (slightly) better player than Daniel, so something is very wrong if Henrik is dropping 22 spots while Daniel is flat.

My assumption for this are:
- Daniel missed more time with injuries and the Sedins are generally less effective when separated
- On goals in which the Sedins both had an assist, Henrik is probably more likely to have the secondary assist (since he was usually passing it to Daniel as his primary option)

Neither is really indicative that Daniel is a significantly better offensive player (Scorer, yes. But the entire Canucks offense ran through Henrik).

 

[Midnight Judges]

There's also no context to star players vs. complimentary players or who someone plays with. Defensemen and complimentary forwards are going to be more prone to variance due to the players they play with than elite star players.

The traditional scoring system was also created, tallied, recorded without that context. I don't see how this is worse in that regard.

Frankly, I don't know how anyone can be a hockey fan and believe that primary assists are worth roughly half of a goal for elite playmakers. It doesn't pass the smell test.

It doesn't actually change things as much as you think. Assists may still total up to 65% of the goal. Look at the charts in the OP. Joe Thornton is still top 9 in per game stats post lockout despite being 40 and among the lowest in terms of goals to assists ratio. Of the 8 players who ranked higher in regular season per game stats, who doesn't belong?

 

[BoredBrandonPridham]

I don't deny that PP time is significant, but this thread is about why new valuations of 1/.54/.1 are superior to simply assuming 1-1-1.

The traditional 1-1-1 system does nothing to address your point either.

Thats not what the thread is about. The thread outlines an effort to submit a list ranked by these valuations and apply it to the “smell test” to try and conclude if it is superior.

But as we concluded here, there is such an overpowering stench of mixing PP production with 5v5 production that the subtleties of the point valuations cannot even be reasonably detected.

It’s like trying to figure out what grapes smell like by sniffing wine.

 

[Gurglesons]

Glad someone finally put this together.

As a Pens fan I’ve been searching for more objective ways to view Crosby.

 

[Tobias Kahun]

Objective? Don’t think you know what that means

 

[Bowski]

Crosby is the best, already a Canadian/Hockey legend to be revered.
Cry into any Ovy fail pillow you need to until the end of time.

 

[Plural]

The problem with the logic in each of these posts is this: I didn't create the formula. A Mario Lemieux fan did. I merely applied it to players from 2005 to present, and I've given you everything you need to check the math and logic. It's not my objectivity that is on the line here.

Surely some of you can do better than ad hominems:

Yeah...it's not about the method. It's about you proclaiming it as objective when in fact it's arbitrary just like any scoring system is. I know you prefer it, because it props up OV. But you liking it doesn't make the system objective. Claiming so just highlights your bias and people are calling you on that.

Nobody here needs a lesson in what ad hominem is. But it's ironic as hell to see a person with zero ability to self-reflection giving back handed jabs on other posters motives.

Seems to me that we've found the holy trinity of Jagr fan, Sid fan and OV fan. Jags6868, daver and MJ. Everyone else posting here can see the similarities in agenda driven posting with blinders on, the arrogance about your cause is identical in all but the player you're championing. And most funny of all, you all probably dislike each other for the exact qualities you all have in common.

So yeah, feel free to ignore all the criticism as ad hominems and continue to pretend you're actually objective. Cheers.

 

[Khomutov]

Ovechkin>Crosby

 

[bobholly39]

Ovechkin>Crosby

Yes but not by enough.

Lets raise the value of a goal to 3 points, lets lower the value of a primary assist to 0.1 points and lets give a negative 10 to any secondary assists.

This should better - objectively - calculate the true separation between the 2.

 

[Midnight Judges]

Thats not what the thread is about.

Dude. It's in the very first sentence (which I wrote):

The purpose of this thread is to look at the top 30 offensive players (all forwards) since 2005 using an objective scoring system, compare how those results differ from the traditional point system, and discuss which metric grades better against the smell test.

The entire purpose of this thread is to compare 1-1-1 to 1-.54-.1. Few threads have such a clearly stated objective.

But as we concluded here, there is such an overpowering stench of mixing PP production with 5v5 production that the subtleties of the point valuations cannot even be reasonably detected.

Again, the traditional system - which you are defending by default - also does absolutely nothing to address situational stats.

Conversely, the valuations suggested in this thread do nothing to prevent folks from looking at situational stats.

 

[Midnight Judges]

Yes but not by enough.

Lets raise the value of a goal to 3 points, lets lower the value of a primary assist to 0.1 points and lets give a negative 10 to any secondary assists.

This should better - objectively - calculate the true separation between the 2.

I'd be interested in seeing you run the same analysis outlined in the link and coming to that result.

You seem to think it's possible. I'm absolutely sure it is not. How am I wrong?