News Article: Mid Season Goal Assessment

[wgknestrick]

I've been following these guy's game reviews for a while where they try to apply +- goal "contributions" to players. Basically not every player on ice gets a + or a negative - when a goal is scored. Only the players responsible for it get marks.

I think it is pretty much in line with most of our thoughts about the players, but it's nice to quantify them for comparison.

Spoiler alert: Adams sucks.

http://network.yardbarker.com/nhl/a...publisher_the_pensnation_module_head_15531263

Click to read more...

 

[Winger for Hire]

I feel dirty seeing Tanner Glass given some credit for every goal he was on the ice for, but only credit for 2 out of 15 goals against. I'm not saying it's wrong, but just doesn't feel right.

 

[Cardboard Orpik]

HFBoards hive mind isn't going to like that Glass stat line.

 

[Valarukar]

I wonder if he counts a failed zone exit as a factor for goals against. Also I highly doubt he counts getting hemmed into your own zone for long periods of time as a factor.

 

[WhatsaMaatta]

This is the problem with just relying on stats in hockey. Unlike baseball, you can't just isolate each player and each event. One play is the result of another, and each player is affected by every other player out on the ice. There are different qualities of players, scoring chances, etc. the reaction isn't going to be just "HFHivemind." It will be a warranted response to a flawed measurement.

 

[wgknestrick]

This is the problem with just relying on stats in hockey. Unlike baseball, you can't just isolate each player and each event. One play is the result of another, and each player is affected by every other player out on the ice. There are different qualities of players, scoring chances, etc. the reaction isn't going to be just "HFHivemind." It will be a warranted response to a flawed measurement.

**** on anything that isn't perfect over the internet.....

Does your car have 0 value as soon as it gets a dent in it?

 

[Crafton]

This is the problem with just relying on stats in hockey. Unlike baseball, you can't just isolate each player and each event. One play is the result of another, and each player is affected by every other player out on the ice. There are different qualities of players, scoring chances, etc. the reaction isn't going to be just "HFHivemind." It will be a warranted response to a flawed measurement.

i was under the impression that the exercise was an attempt to do exactly what you've just outlined? examine a series of plays and determine what actions by what players ultimately contributed to each goal for and against. i'm not sure your criticism of stats as a whole is justified in this context? this seems to me more a catalog of subjective determination.

 

[Big McLargehuge]

This is the problem with just relying on stats in hockey. Unlike baseball, you can't just isolate each player and each event. One play is the result of another, and each player is affected by every other player out on the ice. There are different qualities of players, scoring chances, etc. the reaction isn't going to be just "HFHivemind." It will be a warranted response to a flawed measurement.

Hockey isn't a zero-sum game, which baseball is, all stats in hockey are inherently flawed.


But that doesn't mean that the stats are meaningless, there's a lot of truth to be found in them...just not absolute truth.

 

[WhatsaMaatta]

My response was geared towards Cardboard Orpik's idea that the negative view of Glass is just HFBoards "hive mind". I myself do value statistical analysis, although as Big said, a lot of said analysis is flawed in hockey.

 

[DoktorZaius]

As per the author's discussion on Adams, I think most fans would say that Glass has been much improved relative to last year (not necessarily saying that much), and that Adams has clearly lost even yet another step, making him (rightfully) the new whipping boy this year. It irritates me when people make it sound like the fans just need a player to be mad at, but I don't think that's the case. Glass was terrible last year, and Adams is terrible this year, and there's nothing wrong with saying so.

 

[IcedCapp]

HFBoards hive mind isn't going to like that Glass stat line.

ya, we all hate Glass because we can't think for ourselves. You caught us.

 

[IcedCapp]

As per the author's discussion on Adams, I think most fans would say that Glass has been much improved relative to last year (not necessarily saying that much), and that Adams has clearly lost even yet another step, making him (rightfully) the new whipping boy this year. It irritates me when people make it sound like the fans just need a player to be mad at, but I don't think that's the case. Glass was terrible last year, and Adams is terrible this year, and there's nothing wrong with saying so.

People who have to use the terms whipping boy and hive mind in discussions surrounding players like Tanner Glass and Craig Adams are doing nothing more than trying to prop up their own self-worth by making it seem like they are the only ones who see a religious sign in a pile of poop.

 

[DoktorZaius]

My response was geared towards Cardboard Orpik's idea that the negative view of Glass is just HFBoards "hive mind". I myself do value statistical analysis, although as Big said, a lot of said analysis is flawed in hockey.

Last year, I would have argued vociferously that Glass wasn't merely worthless, but a major detriment to the team, and people who claimed it was "hive mind" were completely full of ****. This year, while he's still not a "good" 4th liner, he's at least playing up to a baseline NHL level, which is significantly more than last year.

 

[vikingGoalie]

i dunno, i'll duck the pitchforks but I think that these stats have a lot more value then some I've seen.

The big factor, imo, is the column indicating the diff/60. Glass is .83 on here.
The only players lower then him are.
Sutter , Conner, Zolnierczyk, Ebbett, Sill, Pyat, Adams

Sutter is demonstrably better and this chart doesn't seem to factor in things like zone start/end position and quality of opposition.

But I really don't have a big qualm saying that Glass is overpaid for what he brings,but is worlds better then last year. I also don't have a huge issue if someone wants to argue that most of the list of players above are more of a drag on the team then Glass.

It does nicely show in a very small sample size what most of us see with Bennett, that his contributions are such that he is top 6 on the team.

Stats are not the end all, they are part of the evaluation. But you can't completely disregard statistics like this either. Sure you can adjust these numbers a tad due to variances not accounted for in the formulation. But it's hard to argue argue with it showing that Adams and Pyatt are the worst players on the team. Though Pyatt deserves a little bit of a pass until he is acclimated to his new team.

 

[sf expat71]

That was fairly elucidating. It does seem like Niskanen has improved a lot this year, the physical playoff hockey concerns some of us have not withstanding. Orpik "feels" much worse than the stat suggests though. The author wasn't fully clear as to how he determined whether a goal against was applied to certain players. Does a chip shot clear to a covered winger at the blue line resulting in a goal off a turnover count against you? Because I feel like that is one of Brooks' specialties...

 

[vodeni]

People who have to use the terms whipping boy and hive mind in discussions surrounding players like Tanner Glass and Craig Adams are doing nothing more than trying to prop up their own self-worth by making it seem like they are the only ones who see a religious sign in a pile of poop.

well said my friend, well said...

 

[Winger for Hire]

I dream of a day when everyone learns to use stats as evaluation tools, not as the end-all-be-all truth. There is never going to be a magical stat (in any sport) that is going to be a perfect way of telling you how good a player is. New stats are nice and can help open your mind to different ways of thinking, but it's still only part of the equation.

The best evaluations are the ones that combine the stats (new and old) and actually seeing what is happening (watching games, reading other opinions, etc). I think this is where a lot of people get hung up, the melding of the two.

Every time a statistical evaluation is brought up, it seems there are two main groups that form: "Stats can tell you everything if you understand them" and "You have to see them play; numbers lie".

Maybe not the best place to throw this out, but better than starting a new thread I guess.

 

[IcedCapp]

I dream of a day when everyone learns to use stats as evaluation tools, not as the end-all-be-all truth. There is never going to be a magical stat (in any sport) that is going to be a perfect way of telling you how good a player is. New stats are nice and can help open your mind to different ways of thinking, but it's still only part of the equation.

The best evaluations are the ones that combine the stats (new and old) and actually seeing what is happening (watching games, reading other opinions, etc). I think this is where a lot of people get hung up, the melding of the two.

Every time a statistical evaluation is brought up, it seems there are two main groups that form: "Stats can tell you everything if you understand them" and "You have to see them play; numbers lie".

Maybe not the best place to throw this out, but better than starting a new thread I guess.

I have not/cannot read the OP, but here's what I find interesting, and maybe I'm wrong:

the person who made this thread was using Corsi to justify Pascal Dupuis' existence. Like, vehemently using Corsi as a justification for Pascal Dupuis not being a horrible top-line player this year.

But in this thread, I'm guessing he's not regaling us with songs of Cosi, because if he were, this thread would be about how Tanner Glass is one of the worst players on the team.

If you want to use advanced stats, fine, but don't cherry pick which advanced stats you used based on what side of the bed you woke up on and which bad player you want to defend against "THE HIVE MIND GROUP THINK SHEEP"

 

[Speaking Moistly]

What people need to learn about statistics of any kind is that they're used to support an argument, they're not supposed to be the argument. You also don't say X because the number, because you're damn argument shouldn't be based on one thing. Context is what really matters, stats are meant to be analyzed too; you don't just blurt out a number as gospel.

Example:

Glass is good because this chart says so.
Right, what about other factors? Who he plays with, when he's played, what he's doing that wouldn't be included in the calculation, is there a bias or flaw in the system used, etc. His GA/60-GF/60 is also horrible, Adams is the only regular roster player with worse.


87 popping up with Crosby again.

 

[Crafton]

i find the larger conversation about the nature of stats misguided. the article in the OP is merely an attempt to create a subjective version of plus/minus - the article writer watched the goals for/against and merely tried to figure out why players were responsible. counting those positive and negative plays is hardly the stuff of 'advanced statistics' - in the false dichotomy of stats vs. observation this article falls far closer onto the observation side of things.

 

[Ragamuffin Gunner]

This is all too confusing to me.

Can you guys please figure out who I don't like and let me know so I can go on a not like him? Maybe we can have a poll so we can find out who the whole group decides who to not like? I don't want to be disliking the wrong players.

Thanks

 

[AjaxTelamon]

Well, stats like this are interesting, but another way of looking at this is how much TOI players have vs. how many statistics they produce, and with what matchups and zone start %'s.

Tanner Glass and Adams have a massive amount of TOI, and very little statistics, minor positive in Glass's case, with significant negative in Adams case. Considering the toughest minutes on the team are played by the Crosby line, you can't consider that Glass or Adams have been used in a shutdown role (except when it really matters, at the end of games). And Glass has lower Defensive Zone start percentages from last year. Last year Glass had a majority of his starts in the Dzone (37%), this year a majority in the neutral zone. So he's naturally not directly involved with as many goals against. He just takes up space and TOI without producing much of anything.

So another interpretation of these stats is that Glass and Adams eat up far too much TOI for their contributions, which are a black hole offensively, and are not "shut down" minutes either. Adams starts in the Dzone [too] often, but against awful competition. And even so, his stats suck. So Adams is being used as a shutdown player against crap competition, and still can't stop goals or move the play to the O zone.

So bottom line after you review these stats, our 4th line sucks. But you could tell that by watching one hockey game.

The other thing that sticks out to me when looking at these stats and considering QoC and FO Start Zone is how Crosby is used against the toughest competition and more often in the Dzone start than O, and he still piles up such big positive numbers. Malkin's minutes are more sheltered, and more mixed.

Crosby is a machine:

http://imgur.com/y6Kd7YT

 

[Sidney the Kidney]

If you want to use advanced stats, fine, but don't cherry pick which advanced stats you used based on what side of the bed you woke up on and which bad player you want to defend against "THE HIVE MIND GROUP THINK SHEEP"

This is my issue with a lot of folks who throw out the advanced stats arguments. They pick and choose stats that fit their argument, but ignore/don't mention the advanced stats that don't.

Poster 1: "Who is the best defenseman on the team?"

Poster 2: "Based on GA ON/60 minutes, it's quite clear that the guy getting 12 minutes a night on the bottom pairing is the best! See? CORSI says so."

Poster 1: "Yeah, but what about his quality of competition? What about his offensive and defensive start times? What about a whole host of other advanced stats?"

Poster 2: " ...."

Poster 1: "Well?"

Poster 2: "Based on GA/60 minutes, it's quite clear ..."

 

[ss53mech]

Try this one on if it interests you:

(poster who doesn't like statistics as a means of performance analysis): Player X isn't very good.

(poster who thinks stats mean more than the visual interpretation of the game): Player X is doing fine this year, you're wrong (because only on the internet are opinions wrong)

(eye test poster): Do you even watch the games?
(stats guy): yes but look at his (insert quantification of play)
(eye test poster): Stop hiding behind numbers.
(stats guy): but he's fine the numbers prove it.

.... and so on.


Or maybe we cold actually just have opinions (based on statistics, visual analysis, a combination, a crystal ball, media input, et cetera) and discuss them. A tactic I would implore all involved in this discussion to utilize is address the problem not the person. People say things that I disagree with all the time, I normally try to refute their point by offering factual/physical disagreement to their offering or try to offer a different perspective on the same point so that we can properly evaluate each others offerings to the discussion.

Why can't we do the same things here folks?

 

[ss53mech]

People who have to use the terms whipping boy and hive mind in discussions surrounding players like Tanner Glass and Craig Adams are doing nothing more than trying to prop up their own self-worth by making it seem like they are the only ones who see a religious sign in a pile of poop.

I'll be honest I don't even know what you are trying to say here. Want to help me out?