Player Discussion Ethan Bear: It's a Bear Market

[PerformanceMcOil]

Because they score more often?!?!

You don't think it's apparent to the casual fan that the best players in the world are better than average players simply by watching them?

Yeah and this is more bias - the stats work cause they identify the really good players, so if they identify a player as being bad, it must be so.

The McDavid's of the world are always going to stand out, no matter what test is applied. He's the definition of an outlier. It's the players near the median that are harder to peg. I've asked this many times before, and never had it answered: For any given stat, what is the distribution? How big of a difference between two players does their need to be in order to say that it is significant? If you have an xGF% (or whatever) of 55% - what range of results does that predict for a player randomly sampled from the pool of 55 xGF% players? How does that compare to the 50% xGF players? Etc. etc.

The advanced stats scene in hockey claims to be scientific, but more often than not, seems bizarrely cultish to me. Surely there is some value to be had, but there are far too many conclusions being made and far too few questions being asked.

 

[CycloneSweep]

I get that but if nothings happens in the other end when he is on the ice, that is not good. I don't care how well he has kept the puck out of your own net, if he ends up as a net negative. 17-18 he was +/-0, 16-17 was the last time he had positive goal share 5on5. You don't need to spend 4+ mil to get those results. Useful player in 3rd pairing and heavy usage in PK. At the price for 3rd pairing guys. That's my opinion.

Okay and how much of that is on the group of forwards in front of him?
Pinning GF numbers on defenders is hard as they rarely drive the offense. If you look at teams that have defenders that have good GF% on their defenders, they are teams that are very good and deep at forward.

Look at Brayden Coburn, guy isn't a good player but last year in Tampa, he had a positive goal for differential. Was it cause he is good at moving the puck and was good offensively? No. He isn't, but the forwards on that team were great offensively.

Let's look at this for example. Over the past 3 seasons we have a total of 9 players with 50%+ goal differential.

Yamamoto, Benning, Draisaitl, McDavid, Enis, Barrie, Sekera, Nuge and Russell. 5 forwards total.

The offense in this team outside of a few forwards is non existent so why are we knocking Larsson for that? Doesn't matter how good an offensive threat a defender is if the forwards suck.
Let's look at purely defensive stats when Larsson is on the ice compared to fellow regular defenders in the league (minimum 300 minutes)
*These are his rankings*

Corsi Against 127
Fenwick Against 31
Shots Against 43
Goals Against 79
xGA 43
Scoring Chances 155
High Danger Chances 74
High Danger Goals Against 44

So looking at defensive numbers solely, Larsson is doing extremely well.

His offensive numbers aren't great but let's look at the guys he plays with most...
This season Larsson has played 745 minutes 5v5
He has played 383 without any of Nuge, McDavid and Draisaitl on the ice. He spends half of his ice time with our shitty players, and that's when McDavid and Drai often play almost half the game. That's massive

Away from the trio he has a 1.28 xGF and a 1.82 xGA. 41.30% xGF, 34.78% GF

Ethan Bear for example, who you are championing as one of our best defenders? 1.36xGF, 2.08 xGA without any of the trio. 39.57% xGF, 46.15% GF

Darnell Nurse? 1.48 xGF, 2.74x GA 35.09% xGF, 33.33% GF

Nurse and Bear COMBINE FOR 6:34 a game on average away from Nuge, McDavid and Draisaitl, Larsson does 8:30 on his own.

So even looking at advanced stats, Larsson's inability to push offense is really almost no different than any other defender on this team when he plays with 50% of this teams forwards. He just gets far less McDrai time.

TLDR: Remove our top 3 guys from all our defenders play and Larsson is right there with them in GF%. Larsson plays the most minutes per game away from our stars' and still manages to put up very solid defensive stats and like ALL our defenders, struggles to put up good offensive ones with forwards who aren't good offensively. Who knew. Goals for stats are heavily influenced by the forwards and goals against heavily influenced by defenders for the most part. Larsson is a very very very solid defensive player this year, period

 

[PerformanceMcOil]

TLDR: Remove our top 3 guys from all our defenders play and Larsson is right there with them in GF%. Larsson plays the most minutes per game away from our stars' and still manages to put up very solid defensive stats and like ALL our defenders, struggles to put up good offensive ones with forwards who aren't good offensively. Who knew. Goals for stats are heavily influenced by the forwards and goals against heavily influenced by defenders for the most part. Larsson is a very very very solid defensive player this year, period

Again highlighting the fact that isolating individual performances is no easy task.

Nevermind that hockey players are human and being destroyed every time you come down Larsson's side might cause some hesitation. Certainly, Draisaitl didn't look like someone who was enjoying his day job yesterday. However, we are told such tangibles can't be measured and are lost in the noise. Coincidentally, any time numbers differ from expectations, it is luck.

 

[Little Fury]

My point is when a player goes against the stats, it's absolutely ridiculous to just blame it on luck rather than acknowledge it as an outlier.

A player with good advanced stats isn't necessarily a good player and a player with bad isn't necessarily bad.

You can't hand waive away things that break the advanced stat narrative as purely luck as it's kind of ridiculous

Dismissing the huge role luck or randomness plays altogether, using outliers to reject such statistical analysis altogether and defaulting to the massively flawed eye test is even more ridiculous IMO and yet plenty of people do just that.

 

[CycloneSweep]

Again highlighting the fact that isolating individual performances is no easy task.

Nevermind that hockey players are human and being destroyed every time you come down Larsson's side might cause some hesitation. Certainly, Draisaitl didn't look like someone who was enjoying his day job yesterday. However, we are told such tangibles can't be measured and are lost in the noise. Coincidentally, any time numbers differ from expectations, it is luck.

A huge part of playoff hockey is the physicality. You don't need to be a big heavy team, but all playoff runs have guys who can deal out some solid punishment and wear them out. A guy like Larsson crushing your stars constantly is going to wear them down and make them a little less effective, or have to change their game and play a bit to avoid those hits.

I have time for advanced stats but when people use them on the surface to make proclamations about players.. it's hard.

For example if you often look at top pairing guys their numbers defensively often look a little worse than guys lower in the lineups, which makes sense as they are often playing against much better players

 

[Soundwave]

Somewhat predictable, young D-Man gets overly excited about a good game and then follows it up with a bad one trying to do too much.

 

[Ol' Jase]

Dismissing the huge role luck or randomness plays altogether, using outliers to reject such statistical analysis altogether and defaulting to the massively flawed eye test is even more ridiculous IMO and yet plenty of people do just that.

Dismissing the role luck and randomness has altogether is the crux of the Expected Goals metric when applied to an individual's performance, so that really doesn't hold a lot of weight here.

No one is "rejecting" advanced stats, it's more questioning the concept that determining an individual's worth no longer requires actually seeing him play. That's just folly.

 

[Little Fury]

Dismissing the role luck and randomness has altogether is the crux of the Expected Goals metric when applied to an individual's performance, so that really doesn't hold a lot of weight here.

Not really. All xG and other unit stats do is give you a picture of what happens when a guy is on the ice. To see how much they might have to do with you need to look at how other teammates do without him and them maybe you can tease out the individual impact while bearing in mind that randomness can account for differences one way or another especially in small sample sizes.

No one is "rejecting" advanced stats, it's more questioning the concept that determining an individual's worth no longer requires actually seeing him play. That's just folly.

Once again, you're questioning a concept that no one is advancing.

 

[Ol' Jase]

Once again, you're questioning a concept that no one is advancing.

That’s crap, and you know it.

This forum is riddled with multiple assertions that players don’t have to be seen to determine their worth.

Not really. All xG and other unit stats do is give you a picture of what happens when a guy is on the ice. To see how much they might have to do with you need to look at how other teammates do without him and them maybe you can tease out the individual impact while bearing in mind that randomness can account for differences one way or another especially in small sample sizes.

The giant missing variable here is that the event is weighted equally among ten skaters. There is no accountability for the actual action (or actions) a player may or may not take that facilitate the event to occur and the outcome of said event.

 

[Forgot About Drai]

That’s crap, and you know it.

This forum is riddled with multiple assertions that players don’t have to be seen to determine their worth.

I think thats a stretch to be honest. A lot of people here do lean heavily towards analytics but anyone who says you dont need to watch a player is either a troll or so far gone you should never take their analysis legitimately.

They are tons of smart analytics posters on here that should not be lumped in with these "posters". It delegitimizes the work and some of the great evidence they have put together to lump them in with that group.

 

[Forgot About Drai]

That’s crap, and you know it.

This forum is riddled with multiple assertions that players don’t have to be seen to determine their worth.



The giant missing variable here is that the event is weighted equally among ten skaters. There is no accountability for the actual action (or actions) a player may or may not take that facilitate the event to occur and the outcome of said event.

But isnt that exactly the same for a 2nd assist or a drop off to mcdavid in the defensive zone? Not like conventional stats cant be manipulated/misrepresented either.

 

[Ol' Jase]

I think thats a stretch to be honest. A lot of people here do lean heavily towards analytics but anyone who says you dont need to watch a player is either a troll or so far gone you should never take their analysis legitimately.

They are tons of smart analytics posters on here that should not be lumped in with these "posters". It delegitimizes the work and some of the great evidence they have put together to lump them in with that group.

There was a poster in the “Fire Holland thread” who made it very clear that the stats told him all he needed to know about the team to the extent that he claims to have actually stopped watching the games.

The overhyping and under appreciating of players actual contributions are constantly skewed under the guise of advanced metrics. This thread is a perfect example.

 

[Ol' Jase]

But isnt that exactly the same for a 2nd assist or a drop off to mcdavid in the defensive zone? Not like conventional stats cant be manipulated/misrepresented either.

I’m not sure what you’re trying to say here.

 

[Forgot About Drai]

There was a poster in the “Fire Holland thread” who made it very clear that the stats told him all he needed to know about the team to the extent that he claims to have actually stopped watching the games.

The overhyping and under appreciating of players actual contributions are constantly skewed under the guise of advanced metrics. This thread is a perfect example.

I frankly feel the exact opposite. Advanced analytics have told me to pay more attention to Ethan Bear and I noticed all these small things he does that make him such a good dman projecting forward.

Before that I'd notice his 1-2 big mistakes a game and think he sucked. Now I notice a lot more nuance in his game.

 

[Forgot About Drai]

I’m not sure what you’re trying to say here.

There is no accountability for the actual action (or actions) a player may or may not take that facilitate the event to occur and the outcome of said event.

I am responding to the above. You said there is no accountability or action that facilitated the event. I argue that that the exact same thing can be said for some secondary assists and something like a drop off to Mcdavid in our own zone before he dangles 3 people and scores.

 

[Ol' Jase]

I frankly feel the exact opposite. Advanced analytics have told me to pay more attention to Ethan Bear and I noticed all these small things he does that make him such a good dman projecting forward.

Before that I'd notice his 1-2 big mistakes a game and think he sucked. Now I notice a lot more nuance in his game.

In statistics, that scenario has a very specific and well defined term...

 

[Ol' Jase]

There is no accountability for the actual action (or actions) a player may or may not take that facilitate the event to occur and the outcome of said event.

I am responding to the above. You said there is no accountability or action that facilitated the event. I argue that that the exact same thing can be said for some secondary assists and something like a drop off to Mcdavid in our own zone before he dangles 3 people and scores.

But all 10 players on the ice when that happens aren’t attributed a secondary assist. That’s the difference, and it’s pretty significant.

 

[Forgot About Drai]

But all 10 players aren’t attributed a secondary assist. That’s the difference, and it’s pretty significant.

I see what you are saying. Fair enough. Same reason +/- is garbage stat.

As for the confirmation bias piece, that was a gotcha moment. But that stand up at the blue line and those slick breakout passes arent fake and are very real. No matter what thats going to show me this guy has potential going forward.

 

[Little Fury]

That’s crap, and you know it.

This forum is riddled with multiple assertions that players don’t have to be seen to determine their worth.

OK, show me.

The giant missing variable here is that the event is weighted equally among ten skaters. There is no accountability for the actual action (or actions) a player may or may not take that facilitate the event to occur and the outcome of said event.

A stat that shows what happens when a player is on the ice need not explain why that event occurred and to cite its failure to do something it is not designed to do is ridiculous.

 

[Little Fury]

The overhyping and under appreciating of players actual contributions are constantly skewed under the guise of advanced metrics. This thread is a perfect example.

The overhyping and under appreciating of players actual contributions are constantly skewed under the guise of the eye test. This thread is a perfect example.

 

[Ol' Jase]

I see what you are saying. Fair enough. Same reason +/- is garbage stat.

As for the confirmation bias piece, that was a gotcha moment. But that stand up at the blue line and those slick breakout passes arent fake and are very real. No matter what thats going to show me this guy has potential going forward.

I’m not trying to show you up or anything, I have been a proponent of Advance stats since they started being tracked and they absolutely serve an important purpose.

They just severely lack being held up to scrutiny when it comes to individual assessment for more than one reason.

Bear in mind that one of the biggest pushes for advanced stats being used in player assessment was the correlation between TEAM success and possession metrics. Team success. There are so many variables not at play when team assessment is done compared to player assessment that it makes “objective analysis” much more credible.

 

[Ol' Jase]

The overhyping and under appreciating of players actual contributions are constantly skewed under the guise of the eye test. This thread is a perfect example.

Strong rebuttal.

This is almost as strong as “you just don’t understand advanced stats”...

I’m going to go out on a limb and say that Darnell Nurse is far more valuable to this team right now than Ethan Bear is. It doesn’t mean Bear isn’t an up and comer and someone I hope the Oilers keep, but it’s really not close to who is actually more valuable to the team at this moment.

 

[Forgot About Drai]

I’m not trying to show you up or anything, I have been a proponent of Advance stats since they started being tracked and they absolutely serve an important purpose.

They just severely lack being held up to scrutiny when it comes to individual assessment for more than one reason.

Bear in mind that one of the biggest pushes for advanced stats being used in player assessment what the correlation between TEAM success and possession metrics. Team success. There are so many variables not at play when team assessment is done compared to player assessment that it makes “objective analysis” much more credible.

Absolutely man, no hard feelings. We arguing about f***ing hockey analytics haha

I love these convos. Im just starting to dive into analytics deeper so these debates are great for everyone becoming more objective and knowledgeable.

 

[Ol' Jase]

A stat that shows what happens when a player is on the ice need not explain why that event occurred and to cite its failure to do something it is not designed to do is ridiculous.

Then the statistic doesn’t account for the variables surrounding its occurrence and therefore cannot be claimed to be objective on any level.

This is literally the fundamental rule surrounding statistics. The less variables identified and accounted for, the less reliable that it is to answering a proposed outcome.

 

[Aceboogie]

Wont dive into the analytics debate too much. But I will pose the question about why its only "advanced stats" that need scrutiny and questioning. Why do people just assume their "eye test" is infallible?

-Even if you watch every Oiler game, there are 10 skaters on the ice at one time, all with a fast pace of play and multiple variables going on. If you watch any play back multiple times, you can come away with a completely different evaluation of the play/players. Who can say they actually will rewatch games multiple times to do this? we all watch games in real time, we are all subject to the same limiations

-Also even if you watch every game, every person is biased in many ways. Some people love rough and tumble players, so if a player makes a big hit- they will come away loving that player, even if it means overlooking (subconsciously) many not-so-good plays. The biggest thing is probably many people overweigh point totals (even if they do not like advanced stats). It is always interesting watching a player have a extremely high s% stretch, and then looking at HF or twitter and watching all the fans fall in love with this player and essentially invent narratives why hes so good and will keep this up

-Lastly- the casual fans probably watches only their team for the full game, plus maybe bits and pieces of other games each night. Yet, despite probably watching 5% of the total NHL games per season, will have strong opinions about players on other teams

To me, the "eye test" has gone far too long without being questioned. It shouldn't be "analytics are wrong, its doesnt agree to my eye". It should be "what I am I possibly missing in my evaluations"