zeke's Official Top-20 Center Rankings

[zeke]

this is not true. Giroux switch to wing last year so not over half the time the past two years. He also took more draws than people on your list. No problem with excluding him but he was a about a 1/2 a center last year. and full time the year before.

giroux is such a weird case I'm happy not to stretch the definitions to include him. tbh as funny as it sounds he wouldn't even crack the top-30 in p1/60 over the last 2yrs, so he would have been excluded from my list even based on that. just such a bizarre turn for him last year that it's hard to know what to make of it, with or without stats.

 

[JetsWillFly4Ever]

Good work Zeke, I like the objectivity.

 

[tarheelhockey]

And by official, I mean mine, of course.

So after all the fighting on the other threads, i tried to crank out a stats-based Center ranking, but nothing too fancy. At the same time it's both complicated and simple.


NOTE: IF YOU HATE STATS JUST SKIP TO THE BOTTOM FOR THE FINAL RANKINGS

Here's how I approached it.

I looked at the last 2 seasons. This imo is the best combo of "recent" and "good sample size" to judge a player's current value. But it's not perfect, of course, especially when it comes to very young or old players. But it's a good place to start. Second, I put an absolute cutoff at a minimum of 14 ES minutes played per game.

So firstly, I used "Offense" as my most basic criteria for the list of "top centers". i.e. to be a top center, you have to score like one, and if you can't score, you're not a top center no matter how good you are defensively. To judge offense, I used the even strength p/60 and p1/60 stats - basically I ranked the top-30 centers in p/60, and the top-30 centers in p1/60, and then combined the ranks. I know some people might complain about ignoring PP production, but imo PP production is way too noisy and small sample and any offensive ability the players have should show up at even strength anyways. But that wasn't enough - I wanted to adjust this offense for quality of competition as well - so that guys with easier matchups didn't get extra credit. I used opponents' ice time (TOIqoc) to judge quality of competition, and combined it with the scoring ranks. Note that this used their league-wide TOIqoc ranks, of all centers who played minimum 14es mpg. Here's the list of Top Offensive Centers I came up with:

Top Offensive Centers Last 2yrs

1.McDavid 7
2.MacKinnon 22
3.Scheifele 25
4.Crosby 27
5.Matthews 31
6.Barkov 33
7.Malkin 34
8.Stamkos 37
9.Tavares 40
10.Getzlaf 41
11.Backstrom 46
11.Couturier 46
13.Kopitar 48
13.Barzal 48
15.Toews 49
15.Monahan 49
17.Kuznetsov 52
18.Seguin 54
19.Kadri 56
20.Staal 58
21.Bergeron 60
22.Eichel 60
23.Point 68
24.Duchene 72
25.Stastny 72
26.Zetterberg 73
27.Krejci 89
28.Horvat 92

Some surprises on this list when QOC is factored in - namely, that a guy like Toews is still quite good offensively, while guys like Malkin and Kuz really get hammered for going up against much easier competition than most of the others.


Secondly, I wanted to factor in Possession. This was a bit more complicated, but basically I averaged out zone/score adjusted CF% and xGF% and ranked them, and then averaged out zone/score adjusted CFrel and xGFrel and ranked them......and then combined their ranks. Now key here is that I kind of cheated - because I was using OFFENSE as the main determinant of "top center", I didn't look at league-wide possession, but I only looked at the 28 centers who made the Offense list and ranked their possession against each other. I think the basic principal makes sense, though of course my execution of it is pretty hacktastic. And then, again, I needed to adjust for Quality of Competition, so once again I combined the averaged possession ranks with the TOIqoc ranks, and ranked these 28 centers' possession against each other:

Last 2yrs best offensive centers in order of Possession:

1.Bergeron 10
2.McDavid 12
3.Crosby 18
4.Couturier 21
5.Kopitar 28
6.Toews 29
7.Barkov 30
8.Scheifele 32
9.Tavares 39
10.Backstrom 39
11.MacKinnon 41
12.Getzlaf 44
13.Stastny 49
14.Staal 51
15.Kadri 54
16.Stamkos 55
17.Matthews 57
18.Point 59
19.Zetterberg 60
20.Seguin 61
21.Barzal 61
22.Malkin 70
23.Monahan 71
24.Duchene 72
25.Eichel 74
26.Kuznetsov 82
27.Krejci 83
28.Horvat 87

this seems to pass the smell test to me. all the guys with the great defensive reputations are right at the top, while the guys at the bottom are the guys you expect.

So now with these two lists, I just combined the qoc-adjusted offense and the qoc-adjusted possession, and it spat out this list:

ZEKE'S PERFECT TOP-20 CENTER RANKINGS

1.McDavid 19
2.Crosby 45
3.Scheifele 57
4.Barkov 63
5.MacKinnon 63
6.Couturier 67
7.Bergeron 70
8.Kopitar 76
9.Toews 78
10.Tavares 79
11.Getzlaf 85
12.Backstrom 85
13.Matthews 88
14.Stamkos 92
15.Malkin 104
16.Staal 109
17.Barzal 109
18.Kadri 110
19.Seguin 115
20.Monahan 120
​

Some guys stick out in surprising ways here - Couturier comes out looking really, really good, while a guy like Malkin drops way down. Guys like Kuznetsov and Eichel don't even make the list at all. Thing is, I don't have any problem with any of those results. In the end, I love how the list rewards guys like Barkov and Bergeron while also acknowledging the question marks on guys like Malkin and Barzal.

Note, though, that this isn't a prediction of which centers are necessarily the best RIGHT NOW, heading into THIS SEASON. To do that I would want to add in some kind of Age adjustment, which would favor which would bump up kids like Eichel, Barzal, Matthews while knocking down guys like Backstrom and Staal a bit.



So what do you think? does my list beat the NHL's? or the famous Hockey Guy's?

or is it poop?

Glad to see Jordan Staal getting mad respect.

 

[WetcoastOrca]

no doubt I place great importance on QOC. thing is, Malkin is the actual 2nd line center that gets turned into a 1st line center by other stats. Couts, Toews, Monahan are all clearcut #1Cs on their teams, not sheltered 2nd line guys. And kadri over the last 2yrs started as the #1C as Matthews was eased into the league.



true, though again the difference you're thinking of is probably exaggerated by PP minutes. and I still think the PP is way too prone to small sample fluke and opporunity to get a good read from it. I'm not really sure why a player's superior true talent offense wouldn't show up at even strength.

and, of course, if we include PP we should probably include PK too.

But here's how these 28 rank in PP p/60 over the last 2yrs

1.Stamkos 7.51
2.Malkin 7.11
3.Backstrom 6.70
4.Seguin 6.45
5.Crosby 6.44
6.Kadri 6.33
7.Matthews 6.29
8.Eichel 6.12
9.Kuznetsov 6.00
10.McDavid 5.72
11.MacKinnon 5.71
12.Barzal 5.65
13.Kopitar 5.63
14.Bergeron 5.52
15.Getzlaf 5.32
16.Tavares 5.41
17.Staal 5.05
18.Krejci 4.71
19.Horvat 4.47
20.Scheifele 4.12
21.Zetterberg 4.06
22.Barkov 3.98
23.Point 3.92
24.Monahan 3.91
25.Toews 3.88
26.Stastny 3.44
27.Duchene 2.98
28.Couturier 2.56

and p1/60:

1.Stamkos 5.82
2.Kadri 5.28
3.Seguin 4.78
4.Crosby 4.65
5.Bergeron 4.65
6.Matthews 4.49
7.Malkin 4.43
8.MacKinnon 4.22
9.Backstrom 3.97
10.Tavares 3.94
11.Getzlaf 3.87
12.Kuznetsov 3.77
13.Staal 3.75
14.Horvat 3.72
15.Eichel 3.70
16.Krejci 3.49
17.Monahan 3.32
18.McDavid 3.23
19.Kopitar 3.14
20.Stastny 2.97
21.Zetterberg 2.84
22.Barzal 2.82
23.Duchene 2.54
24.Toews 2.44
25.Barkov 2.33
26.Scheifele 2.27
27.Point 2.05
28.Couturier 1.92

and then we combine them:

1.Stamkos 2
2.Seguin 7
3.Kadri 8
4.Crosby 9
5.Malkin 9
6.Backstrom 12
7.Matthews 13
8.MacKinnon 19
9.Bergeron 19
10.Kuznetsov 21
11.Eichel 23
12.Tavares 26
13.Getzlaf 26
14.McDavid 28
15.Staal 30
16.Kopitar 32
17.Horvat 33
18.Barzal 34
19.Krejci 34
20.Monahan 41
21.Zetterberg 42
22.Stastny 46
23.Scheifele 46
24.Barkov 47
25.Toews 49
26.Point 50
27.Duchene 50
28.Couturier 56

already I think you can see more anomalies pop up in the PP numbers, most of which is likely just small sample size issue. Or maybe scheifele really just sucks on the PP for some reason, I dunno.

Including the PP numbers would give a solid boost to Stamkos/Seguin/Kadri/Crosby/Malkin (2-3 spots maybe?) and would hurt Scheifele, Barkov, Toews, Couturier by a similar amount.

If I just eyeball it, maybe it would change my list into this?


1.McDavid 19
2.Crosby 45
3.MacKinnon 63
4.Bergeron 70
5.Scheifele 57
6.Barkov 63
7.Kopitar 76
8.Tavares 79
9.Couturier 67
10.Stamkos 92
11.Backstrom 85
12.Matthews 88
13.Getzlaf 85
14.Toews 78
15.Malkin 104
16.Kadri 110
17.Seguin 115
18.Staal 109
19.Barzal 109
20.Monahan 120

maybe something like that? does that look better?

I think you’ve done a great job.
But in the end there’s just a few that just don’t pass the smell test for me. For example I don’t think there’s a GM in the league who would take Kadri over Malkin, Seguin or Barzal. Toews for me is in the same boat although he has a pretty high pedigree. I also see Monhan as a lower end number one/ higher end number two.
But all in all I like how you’ve done this objectively.

 

[PenguinSpeed]

I think you’ve done a great job.
But in the end there’s just a few that just don’t pass the smell test for me. For example I don’t think there’s a GM in the league who would take Kadri over Malkin, Seguin or Barzal. Toews for me is in the same boat although he has a pretty high pedigree. I also see Monhan as a lower end number one/ higher end number two.
But all in all I like how you’ve done this objectively.

-Yea, Kadri with his fascinating .29 points per game in the playoffs. One of the top chokers in the entire NHL

 

[Martin Skoula]

I think you’ve done a great job.
But in the end there’s just a few that just don’t pass the smell test for me. For example I don’t think there’s a GM in the league who would take Kadri over Malkin, Seguin or Barzal. Toews for me is in the same boat although he has a pretty high pedigree. I also see Monhan as a lower end number one/ higher end number two.
But all in all I like how you’ve done this objectively.

It's not necessarily about taking Kadri over those other players if you were starting a team, it could just be that Kadri is in a situation suited to his exact skillset such that switching him out for Seguin wouldn't necessarily improve the team without changing everyone else's roles and how they play. That being said, not every team needs a guy to shadow the opponent's 1C while playing a very specific net crashing role offensively especially on the PP, we just have a distinct need for that because our #1C was very young and sheltered, and most of our wingers are playmakers who need the center to shoot and crash.

 

[zeke]

I think you’ve done a great job.
But in the end there’s just a few that just don’t pass the smell test for me. For example I don’t think there’s a GM in the league who would take Kadri over Malkin, Seguin or Barzal. Toews for me is in the same boat although he has a pretty high pedigree. I also see Monhan as a lower end number one/ higher end number two.
But all in all I like how you’ve done this objectively.

it's at least a bit funny that people's main objection to this ranking is a maple leaf, again.

here's one thing I'll add about Malkin (and it applies to Crosby too, probably) - these guys probably haven't been giving it their all in recent regular seasons. that might mean they're having lower impact than they are capable of giving....but intentionally lower impact, to rest their old bodies for the playoffs.

and I'm not sure that same comfort level exists for any of the other players on the list - not sure anyone else is in a position to be able to intentionally take it easy in the regular season for longterm gain. all the others seem to have been in positions where they really needed to give it their all just to get into good playoff position, or any playoff position.

 

[Kamus]

Pretty good list..
I like not adding PP points as this is a centre ranking not forward. Besides the face off there is no defined centre roll on the PP. many centres have different roles on the pp unlike even strength.

 

[WetcoastOrca]

it's at least a bit funny that people's main objection to this ranking is a maple leaf, again.

here's one thing I'll add about Malkin (and it applies to Crosby too, probably) - these guys probably haven't been giving it their all in recent regular seasons. that might mean they're having lower impact than they are capable of giving....but intentionally lower impact, to rest their old bodies for the playoffs.

and I'm not sure that same comfort level exists for any of the other players on the list - not sure anyone else is in a position to be able to intentionally take it easy in the regular season for longterm gain. all the others seem to have been in positions where they really needed to give it their all just to get into good playoff position, or any playoff position.

Actually I have a few objections which I listed. But Kadri is probably the most egregious along with Malkin.
When 31 of 31 GMs would all take Malkin and Seguin over Kadri I think the problem is with your formala. There’s nothing wrong with using stats but when they yield unsupportable conclusions it’s time to tweak the formula. I think that’s a fair statement which most stats people would agree.

 

[WetcoastOrca]

It's not necessarily about taking Kadri over those other players if you were starting a team, it could just be that Kadri is in a situation suited to his exact skillset such that switching him out for Seguin wouldn't necessarily improve the team without changing everyone else's roles and how they play. That being said, not every team needs a guy to shadow the opponent's 1C while playing a very specific net crashing role offensively especially on the PP, we just have a distinct need for that because our #1C was very young and sheltered, and most of our wingers are playmakers who need the center to shoot and crash.

I disagree. If the idea is to rank the top centers then if every GM in the league would take Malkin and Seguin over Kadri then I think that’s a good sign that the statistical formula needs to be tweaked. Stats are only as good as the formula that they’re used with.
It’s not about team need.
It’s about ranking the top centers.

 

[Martin Skoula]

I disagree. If the idea is to rank the top centers then if every GM in the league would take Malkin and Seguin over Kadri then I think that’s a good sign that the statistical formula needs to be tweaked. Stats are only as good as the formula that they’re used with.
It’s not about team need.
It’s about ranking the top centers.

I guess the list can better be described as the top 28 centers at excelling in their role, otherwise people are always going to have issues with the ranking due differences in weighing the competition faced and using /60 rates. Malkin and Seguin would not be performing as well in raw stats if they were tasked with shutting down the opponent's 1C and only playing half of every PP.

 

[zeke]

if every GM in the league would take Malkin and Seguin over Kadri

for the record, this appeal to authority isn't a great argument, imo.

 

[jetsforever]

Seems like a decent list and reasoning
Couturier seems really high though

 

[Sidney the Kidney]

but it shouldn't punish him - he should be able to dominate possesion in those easier minutes like, say, Barzal did last year.

He generally does, though? I know you're using different metrics for possession, but Malkin's career CF% and FF% is 52.9% and 53.2%, respectively. His career CFRel and FFRel are 2.6 and 1.7, respectively.

He might not be Patrice Bergeron when it comes to these metrics, but I don't see how he's so bad at them that it would cost him spots in rankings from where he should be based on raw production. The only two areas that seem to really, really hamper him are QoC and if anything PK or defensive responsibilities related is heavily involved.

 

[zeke]

Let's see if we can spot any trends in Malkin's career at even strength.

quick letter grades for quality of competition, then the average of cf/xgf%, and then the average of cfrel/xgfrel (all adjusted for score/zone).

Yr1: 79gms, A+ qoc, 50.5%, +3.8
Yr2: 82gms, A+ qoc, 47.8%, -1.9
Yr3: 66gms, A qoc, 52.1%, -1.6
----------------------------------
Yr4: 43gms, A qoc, 58.9%, +8.6
Yr5: 75gms, A+ qoc, 56.0%, +2.4
Yr6: 31gms, A+ qoc, 58.2%, +8.1
Yr7: 60gms, A qoc, 53.1%, +3.8
-----------------------------------
Yr8: 69gms, B+ qoc, 54.7%, +1.9
Yr9: 57gms, A- qoc, 54.8%, +2.4
-----------------------------------
Yr10: 62gms, B+ qoc, 52.2%, -0.1
Yr11: 78gms, B+ qoc, 51.7%, -0.2


hmm. despite his still great offense, I think we may be witnessing a serious deterioration in his game, tbh.

Yrs1-3: early career learning the ropes against top tier compeition
Yrs4-7: absolute dominance against top tier competition
Yrs8-9: a little less dominant against a little easier competition
Yrs10-11: just holding his own against middle tier competition.

and those last 2yrs are, of course, the years looked at in this thread.

there is no doubting that Malkin has been able to be absolutely dominant in all aspects of the game against the best competition during his career. But he's just not doing that anymore, despite the gaudy offensive totals.

Either Malkin is deteriorating, or he's taking it easy in the regular season to save some wear and tear on his oft-injured body.

 

[shtorm2005]

Probably you need to edit your title, it's 5v5 Top ratings, not overall.

 

[rintinw]

Getting the list by adding ranking positions of different stats together ...

First, there is problem with proportion when you are using rankings instead of raw numbers. Here is example situation:

Player 1: 50G, 200PT
Player 2: 55G, 100PT
Player 3: 54G, 95PT
Player 4: 53G, 90PT

Goal rank:
1. Player 2
2. Player 3
3. Player 4
4. Player 1

Point rank:
1. Player 1
2. Player 2
3. Player 3
4. Player 4

List based on rankings (and aggregating them):
Player 2: 3
Player 3: 5
Player 1: 5
Player 4: 7

And it wouldn't matter a bit that Player 1 is vastly superior to the other 3.

Second, while I do not agree with stats used (P/60, P1/60, QoC, CF% rel), the bigger problem is the proposed method completely ignores possible complex relations between those stats and replaces them with simple aggregation.

 

[zeke]

Getting the list by adding ranking positions of different stats together ...

First, there is problem with proportion when you are using rankings instead of raw numbers. Here is example situation:

Player 1: 50G, 200PT
Player 2: 55G, 100PT
Player 3: 54G, 95PT
Player 4: 53G, 90PT

Goal rank:
1. Player 2
2. Player 3
3. Player 4
4. Player 1

Point rank:
1. Player 1
2. Player 2
3. Player 3
4. Player 4

List based on rankings (and aggregating them):
Player 2: 3
Player 3: 5
Player 1: 5
Player 4: 7

And it wouldn't matter a bit that Player 1 is vastly superior to the other 3.

Second, while I do not agree with stats used (P/60, P1/60, QoC, CF% rel), the bigger problem is the proposed method completely ignores possible complex relations between those stats and replaces them with simple aggregation.

correct.

and intentional.

 

[Territory]

Nice list zeke!

 

[Garthinater]

Matthews isn't 5th in the league in points per 60 minutes, so uhhhh, you're wrong.

In an earlier discussion I said you were coming across silly but now your pushing that to new lengths lol just stop while your behind

 

[Volica]

LoLworthy list.
Too much work for a useless result.

 

[Garthinater]

Seeing a leafs fan made this I thought matthews was gonna end up top 7ish. Nice to see a neutral list without any bias for once. Of course this ranking isn't exact, no gm would take kadri over Malkin. Still it's very interesting

 

[Albi34]

To the people complaining, this is a WAY more thought out and objective list than most we see on here. Nice job Zeke

 

[Ratsreign]

Stats are great! You can pick and choose those which best fit your narrative....and downplay the relevance of the rest.

 

[Albi34]

Stats are great! You can pick and choose those which best fit your narrative....and downplay the relevance of the rest.

What is the narrative? The only one I can possibly see is that two-way forwards are maybe being ranked a little highly? I'm genuinely confused what your point is.