No, that's not what i said.. Umberger and Nash BOTH increased their GWGs which was my point.
That's great, but it still has Umberger over Nash. The numbers are simply smaller.
This version of the stat is meaningless for the same reason the other version is meaningless. Umberger scores to make it 1-0 five minutes into the game and the Jackets go on to win 5-4, and he is more clutch than Nash scoring a hat trick in a game they lost 4-3 at the last second?
That's great, but it still has Umberger over Nash. The numbers are simply smaller.
This version of the stat is meaningless for the same reason the other version is meaningless. Umberger scores to make it 1-0 five minutes into the game and the Jackets go on to win 5-4, and he is more clutch than Nash scoring a hat trick in a game they lost 4-3 at the last second?
I don't know about measuring clutch factor... but it's no secret that GWINNINGGs are noted only in victories and only accounts for the goal which leads to a team winning.
Umberger scoring in a victory vs Nash scoring 3 in a loss? I'm not claiming one being more 'impressive' or more 'clutch'... but it's definite Nash's goals don't factor in the equation of GWGs at all... which is what we're talking about.
I wouldn't buy that theory yet...
In the small sample size of 2 teams alone above, we see the 'stars' for the most part gain GWGs with more of the lesser players losing theirs... Burrows aside.
There's a reason for that, and it has nothing to do with "clutch".
When a team has the lead, they don't play their star players as much. Thus, star players are more likely to score the goal that puts the team ahead, because 1) they're better at scoring goals and 2) they get more ice time when the game is tied.
Imagine a game that is 3-3 with 10 minutes left. It's not unrealistic to suggest that Daniel and Henrik Sedin will play at least 5 minutes of the game. What if the score is 3-1 Vancouver? Vigneault is going to ride Higgins/Hansen/Kesler/Malhotra/Lapierre.
Quote:
I'm leaning the other way than you right now...
And goals are a random event? not exactly sure where you're going there... unless you are simply discounting the value of GWGs all together, which is also fine.
The present method allows for many GWGs to be awarded in a random sort of way when players have them fall to em late on meaningless goals... my method would at least diminish that factor giving it a little more meaning... at least imo.
Goals are random. They're unpredictable. If a player scores 5 goals in a game, what are the odds that he scores in his next game? If an elite 40-goal scorer doesn't score for 5 games in a row, what are the odds that he scores in the next game? Totally incalculable. There's not enough of them to have any predictive power - let alone when you're filtering down to 5 or 10 goals.
If a number doesn't have any predictive power - there's no correlation between month-to-month, season-to-season, etc - then it is purely descriptive. Players can score GWGs, sure, but players don't possess the ability to score GWGs.
Put it this way: Player can't repeat total goal totals year to year because shooting percentage is so volatile and out of a players' control. How can they possibly maintain an ability to score goals at a faster rate when the game is tied? Do they play below their ability when the game is in hand? Do they possess superhuman abilities when the team needs it most?
I do not think this stat is useless at all. It would be much more reliable if we took in to consideration what percentage of a player's goals are gwg. If a player goes a few seasons with a significantly higher than average percentage of gwg than goals I would be confident in saying that is a clutch player. It is all about sample size really. I just feel this stat would be much much better represented as a ratio of gwg to total goals.
I don't know about measuring clutch factor... but it's no secret that GWINNINGGs are noted only in victories and only accounts for the goal which leads to a team winning.
Right, so there's a selection bias right off the bat. That bias continues to exist regardless of how we tweak the data.
Quote:
Umberger scoring in a victory vs Nash scoring 3 in a loss? I'm not claiming one being more 'impressive' or more 'clutch'... but it's definite Nash's goals don't factor in the equation of GWGs at all... which is what we're talking about.
And you don't think there's a problem with a stat that rewards the 1-0 goal in a 6-0 win, but ignores a hat trick in a 4-3 OTL?
Right, so there's a selection bias right off the bat. That bias continues to exist regardless of how we tweak the data.
And you don't think there's a problem with a stat that rewards the 1-0 goal in a 6-0 win, but ignores a hat trick in a 4-3 OTL?
Dude... I don't know where you got that I somehow love the GWG stat or are advocating for it... I've stated that quite clearly in past posts.
So I don't know why you are trying to argue how crappy the stat is to me.
I don't care if the stat is ever referenced again.
All I proposed is what I think is better than the present version of the stat.
Me: I like this better than GWG
You: GWG is flawed and bias
Me: yeah. It's not the best stat. But here's what I think is an improvement..
You: But GWG suck.
Me: ok. But at least this may be an improvment on it.
You: but can't you see how biased and flawed the stat is?
Me: I don't care! *bangs head against wall.
There's a reason for that, and it has nothing to do with "clutch".
never claimed it did.
Quote:
When a team has the lead, they don't play their star players as much. Thus, star players are more likely to score the goal that puts the team ahead, because 1) they're better at scoring goals and 2) they get more ice time when the game is tied.
Imagine a game that is 3o-3 with 10 minutes left. It's not unrealistic to suggest that Daniel and Henrik Sedin will play at least 5 minutes of the game. What if the score is 3-1 Vancouver? Vigneault is going to ride Higgins/Hansen/Kesler/Malhotra/Lapier
Exactly! Agreed.
What's the whole point of GWG stat?
To somehow give extra recognition to some goals as being more important than others...
So I proposed a way to better determine which players are breaking ties and swinging the momentum of games a little better than the present model does.
Basically shifting slightly which goals would be defined as GWGs I think the more appropriate goal is recognized more often.
Dude... I don't know where you got that I somehow love the GWG stat or are advocating for it... I've stated that quite clearly in past posts.
So I don't know why you are trying to argue how crappy the stat is to me.
I don't care if the stat is ever referenced again.
All I proposed is what I think is better than the present version of the stat.
Me: I like this better than GWG
You: GWG is flawed and bias
Me: yeah. It's not the best stat. But here's what I think is an improvement..
You: But GWG suck.
Me: ok. But at least this may be an improvment on it.
You: but can't you see how biased and flawed the stat is?
Me: I don't care! *bangs head against wall.
If you're going to build a strawman, you might as well build a big one.
You are claiming to have taken what is essentially a random indicator, and refined it into something not-random (ie, "cream rising" and similar comments).
You: I know GWG is completely descriptive and doesn't really relate important information, but here's an improvement!
Me: But that stat is just as descriptive and irrelevant as the old one.
If you're going to build a strawman, you might as well build a big one.
You are claiming to have taken what is essentially a random indicator, and refined it into something not-random (ie, "cream rising" and similar comments).
You: I know GWG is completely descriptive and doesn't really relate important information, but here's an improvement!
Me: But that stat is just as descriptive and irrelevant as the old one.
ok.. fair enough... altho i would state 'less random' would be more accurate than 'not-random'. (not that i believe the stat is random).
we'll agree to disagree, since i think my modification is less irrelevant than the original and you don't.
It can put things in perspective, for instance:
Dave Andreychuk,,,,been a discussion about him and the HoF. Another argument against could be:
5th - total games
14th - total goals
1st - PP goals
T29th - GWG (tied with Guerin)
Among his contemporaries:
Luc Robitaille
16th - total games
10th - total goals
5th - PP goals
17th - GWG
Obviously from this stat it shows that Robitaille scores are far morworthy than Andreychuk, who seemed to be more of a PP specialist that probably added to leads....which in itself is worthy to a degree because it can also deflate an opponent from mounting a comeback.
Luc in the HoF. Andreychuk probably won't make it, (maybe an old timers committee)
How about this argument:
Pierre Turgeon
31st - total games
36th - total goals
18th - PP goals
T18th - GWG (tied with Sakic)
Seems like Turgeon, another contemporary of both Robitaille and Andreychuk, was a more clutch, go-to guy, seemingly at even strength then either of them.....yet the consensus is he probably won't make the HoF.
Not having a breakdown split as to how or when the GWGs occured is a missing factor in the argument.
Sergei Federov
40th - total games
48th - total goals
47th - PP goals
T9th - GWG (with Nieuwendyk)
Somewhere down the list in comparison to those above, yet he's 9th overall in GWGs.....and there are some who feel he may not be worthy for the HoF, (check any HoF nominee threads to confirm, mostly because of his tail off late in his career)
GWG is a flawed stat for sure, but still relative enough for argument if put in the right context.
Put almost no weight into it. Varies considerably from season to season because of how random it is. And, surprise, players that score more goals in general will usually have more GWG.
Gosh, I really can't figure out why Fedorov went from a G:GWG ratio of 1:4 with the Red Wings to a ratio of 1:9 with the Blue Jackets. Maybe he got less clutch as he aged
