Top Forwards by WAR 2007-2020

[wetcoast]

I've said on many occasions that Crosby has a better/more consistent prime (1 of the best ever) so I'm not surprised by that. But OV clearly has the better peak seasons (yes Crosby getting injured robbed us of that debate but that's unfortunate).

Sure fair enough but I have always been a huge proponent of rate of play and while Crosby was injured for what could have been an incredible peak his consistency also makes the likelihood of that peak very likely as well.

I'm not a Caps or Pens fan, heck not even really a Canucks fan but more a fan of the game so I like to use the same standards for all players as much as I can.

The first graph which shows Crosby with an absolute lead, even with the games missed just confirms what I have seen.

I don't know what EP has to do with anything, maybe you were replying to someone else? But yeah he's a great player.

No the EP reference was just something I missed the first time and just injected as even as good as he has been I was surprised to see him anywhere on the list.

 

[BlueBaron]

Have to thank the OP for the delightful thread. It has been great fun watching people short circuit when the hard data doesn't match their biases.

What happened with Eichel?

 

[kugelbahn]

yeah, those negative passing years give no "grace period" slack to OV,
he probably f*ed up his passing model by eg having 3 (three) ES primary assists in 79 games in 15/16
and ended up somewhere in between Cody McLeod -67.3 passing and Phil Kessel +68.6 passing

I am in general interested in regression models with such a zoo of different input parameters, especially with stats like Corsi.
It works surprisingly well I must say with few outliers.

 

[grapes55]

I find it interesting seeing how Drai has the highest WAR this year while other parts of the advanced stats community have him ranked lower for his performance

 

[Midnight Judges]

This needs repeating,
It's pretty ironic that MJ doesn't like a metric that breaks down points like this

Even strength

Goals= 1.0
Primary assist =0.88
Secondary assist =0 .69

Power Play

Goals= 0.7
Primary assist = 0.61
Secondary assist = 0.44

There's absolutely no irony at all, if you comprehend my opinions.

There is no rational basis for a powerplay goal equaling an ES secondary assist.

 

[wetcoast]

There's absolutely no irony at all, if you comprehend my opinions.

There is no rational basis for a powerplay goal equaling an ES secondary assist.

Well you and I might agree that all points should be the same then eh?


The War breakdown makes sense you just need to take your Ovi-Ray sunglasses off to see it.

 

[goldenbladz1]

I love that Kopi is quietly sitting there at 5th despite not having any 7+ WAR seasons over that span. That's the most classic Kopitar move, it cracks me up. Not too high, not too low, just right.

Same with Toews, people can say what they want but the guy has been gold throughout his career literally.

 

[NHL WAR]

There's absolutely no irony at all, if you comprehend my opinions.

There is no rational basis for a powerplay goal equaling an ES secondary assist.

Well they better cancel the Art Ross

 

[wetcoast]

Well they better cancel the Art Ross

If you asked MJ, he will tell you only one Art Ross actually counts since the lockout.

 

[member 311261]

I think too much weight is given to WAR statistics not just in hockey, but baseball too. The issue is if you change one criterion just slightly, it can alter the results significantly. Same with QB ratings.

I think natural for people to want to try to quantify players into static lists like these. It's fun. Same with all the other advanced stats out there.

The stats are useful, but they aren't the be all, end all. There are simply too many variables at play to make even the most advanced stats a 100% accurate way to evaluate a player vs player. It's never apples to apples... but I do see that people will try to view them that way. Again human nature.

Do they have a use... Absolutely. But they need to be viewed with a additional context to truly have meaning. That's where the science becomes art.

 

[BringTheReign]

Same with Toews, people can say what they want but the guy has been gold throughout his career literally.

It was funny seeing Kopitar sandwiched between Kane and Toews given they spent 2012-2015 trading off Cups. I've always thought that Kopitar should've won the Conn Smythe in 2014. Williams is a gamer, but Kopitar was the engine that drove those great Kings teams on the front end. At least he's won Selkes, I guess?

 

[Midnight Judges]

Have to thank the OP for the delightful thread. It has been great fun watching people short circuit when the hard data doesn't match their biases.

What happened with Eichel?

If you think this is hard data, then you aren't comprehending it.

WAR is, by definition, a rough approximation - and that's assuming it's done properly, which it isn't in this case.

 

[Midnight Judges]

And LOL @ "WAR per 82"

 

[Oneiro]

What's interesting about the first list is that many of those players are good with some elite seasons as opposed to being consistently elite (routine top 5, top 10 finishes each year).

So I could see the Per 82 list change quite a bit by 2030 and it's probably irrelevant.

 

[StonebeatsMcdavid]

Bottom 20 ( all at least a couple hundred games):

player	GP	TOI		SHOOTING	PASSING	DGS	FO	PD	REP		GAR	WAR		PER 82
CODY.MCLEOD	776	7764.9		-19.5	-67.3	-10.2	-0.3	-20.8	66.7		-51.4	-6.8		-0.7
COLTON.ORR	388	2502		-24.9	-31.6	0.6	-0.1	-16.7	21.6		-51.2	-6.8		-1.4
TANNER.GLASS	527	5405.5		-43.4	-38.4	-7.8	-0.7	-5.4	46.5		-49.4	-6.6		-1
ZENON.KONOPKA	317	2720.4		-27.7	-29.9	0.1	7.2	-10.4	23.4		-37.3	-5		-1.3
ZAC.RINALDO	370	3145.2		-21.8	-27.8	-3.1	-0.1	-9.4	26.9		-35.3	-4.7		-1
MIKE.BROWN	407	3350.2		-19	-39.1	-0.3	-0.2	-4.9	28.8		-34.6	-4.6		-0.9
JERRED.SMITHSON	445	5740.7		-43	-44.7	5.3	6.1	-6.2	49.5		-33	-4.4		-0.8
JOHN.SCOTT	266	1845		-21.6	-20.5	3.2	0	-10.3	15.9		-33.4	-4.4		-1.4
RAITIS.IVANANS	212	1350.3		-11	-20.8	-0.6	0	-9.1	11.7		-29.9	-4		-1.5
KRYS.BARCH	354	2300.1		-24.4	-20.1	2.7	-0.1	-7.6	19.8		-29.6	-3.9		-0.9
NICOLAS.DESLAURIERS	376	3947.1		-16.6	-37.6	-5.1	0	-3.1	33.7		-28.7	-3.8		-0.8
ERIC.BOULTON	386	2860.6		-17.9	-20.5	-4.9	-0.1	-8.6	24.7		-27.3	-3.6		-0.8
ETHAN.MOREAU	243	3432.5		-8.8	-23	-15.1	-0.3	-9.7	29.6		-27.3	-3.6		-1.2
JARED.BOLL	578	4371.3		-29.6	-35.7	2.5	-0.2	-1.3	37.6		-26.7	-3.6		-0.5
TRAVIS.MOEN	544	7330.7		-47.8	-41.4	-1.4	-0.4	1	63.2		-26.6	-3.5		-0.5
BJ.CROMBEEN	445	4988.9		-23.2	-31.4	0.3	-0.8	-12.4	42.9		-24.5	-3.3		-0.6
BRANDON.BOLLIG	241	2183		-15.7	-22.5	2.9	0	-8	18.7		-24.5	-3.3		-1.1
JACOB.DE LA ROSE	229	2701.3		-21.2	-17.5	-0.4	-2.2	-3.9	23		-22.2	-3		-1.1
SHAWN.THORNTON	625	5603.8		-23.4	-44.8	6.3	-0.4	-8	48.2		-22.1	-2.9		-0.4
TODD.MARCHANT	304	4439.9		-27.9	-25.3	-1.9	0.3	-5	38.4		-21.5	-2.9		-0.8

[TBODY] [/TBODY]

Mostly eradicated goons

Hello, I have a couple questions. I know this is from a while ago, hope that's alright. So to start off with, out of curiosity what is the lowest WAR in a particular season. Also, what is considered to be/what level is the replaceable player. Lastly, I feel like I was reading an article where it showed that Panarin had the highest WAR this season with like 4.4 or something.

 

[Devil Dancer]

WAR, possibly the most useless stat in hockey, since it's completely made up and doesn't actually measure anything.

 

[JoVel]

Is WAR any good on defensemen?

 

[Tkachuk Norris]

Another stat showing Gaudreau is one of the leagues best.... weird. I’m actually shocked that one of the most productive forwards since he entered the league is considered one of the most valuable

 

[NHL WAR]

Hello, I have a couple questions. I know this is from a while ago, hope that's alright. So to start off with, out of curiosity what is the lowest WAR in a particular season. Also, what is considered to be/what level is the replaceable player. Lastly, I feel like I was reading an article where it showed that Panarin had the highest WAR this season with like 4.4 or something.

The lowest WAR for a forward in one season is -1.8 for Ryan Kesler in 2018/19. Perhaps worse than that, though, is the -1.4 and-1.5 Eric Boulton and John Scott once put up in less than 400 minutes TOI.

For forwards, I had replacement level at anyone making near league minimum that was 5 years post draft and played between 50 and 250 minutes.

Yes, you likely are remembering the article correctly. There have been multiple WAR constructs, with varying results.

 

[StonebeatsMcdavid]

The lowest WAR for a forward in one season is -1.8 for Ryan Kesler in 2018/19. Perhaps worse than that, though, is the -1.4 and-1.5 Eric Boulton and John Scott once put up in less than 400 minutes TOI.

For forwards, I had replacement level at anyone making near league minimum that was 5 years post draft and played between 50 and 250 minutes.

Yes, you likely are remembering the article correctly. There have been multiple WAR constructs, with varying results.

If you don't mind me asking, what would create the varying results?

 

[NHL WAR]

Is WAR any good on defensemen?

I've been working on a model for defenseman, as well as goaltenders, that I'm not quite ready to publish. I can say, though, that for this season Carlson and Josi have the two highest, with Hedman and Slavin only separated by 0.002 for 3rd. That seems to line up with public perception.

 

[NHL WAR]

If you don't mind me asking, what would create the varying results?

I can't say for sure, as I haven't taken a good luck under the hood of other WAR models, but I think they rely almost entirely on expected goals and the like. This somewhat muddles the effectiveness of WAR for the current season, as scoring actual goals is what leads to wins. But it is more designed as a predictor of future value.

Whereas, I tried to keep mine viable as a descriptor of the current season as well. This leads to it being less conservative than the other models as well. They tend to consider a great player to be worth about 4 wins over a replacement player, but 7 wins in more the mark of an MVP in mine.

It's just the nature of the beast. Even in baseball, where WAR has been on the back of cards for at least 5 years, there are two main formulas that have slightly different results each year.

 

[StonebeatsMcdavid]

I can't say for sure, as I haven't taken a good luck under the hood of other WAR models, but I think they rely almost entirely on expected goals and the like. This somewhat muddles the effectiveness of WAR for the current season, as scoring actual goals is what leads to wins. But it is more designed as a predictor of future value.

Whereas, I tried to keep mine viable as a descriptor of the current season as well. This leads to it being less conservative than the other models as well. They tend to consider a great player to be worth about 4 wins over a replacement player, but 7 wins in more the mark of an MVP in mine.

It's just the nature of the beast. Even in baseball, where WAR has been on the back of cards for at least 5 years, there are two main formulas that have slightly different results each year.

Awesome, thanks for responses!

 

[BlueBaron]

Yes, there are post histories. That's why it's funny to see you "recall" instead of actually finding proof.

Matthews greatest claim to fame is coming 2nd in goals as a rookie. He hasn't been closer since then and his point totals are not even close to the premier producers.

Clearly you are not paying attention to my posts as well as you seem to think.

The rest of your message doesnt even deserve a reply.

It is interesting how a simple "surprised to see Matthews so high" post managed to trigger you so much. Might want to look into that.

Are you sure you know who Matthews is?

 

[YoSoyLalo]

WAR, possibly the most useless stat in hockey, since it's completely made up and doesn't actually measure anything.

Good contribution. How about you dedicate your free time working on a statistic to try and help people understand a sport they love, then have someone tell you it’s worthless and useless.