All-Encompassing Stats Thread

[devilsblood]

I find stat rejectors to be extremely arrogant. They think that their eyes, their ability to dissect the game by watching it, completely trumps any need to supplement the analysis with some hard data.

Like you said, myself, yourself and most stat guys are not 100% on the stats bandwagon. We all enjoy watching the game and making determinations from watching the game. The stats just offer a different, unbiased way of looking at the game.

Well I think you could do this IF: 1) you watch every game* by every team* in the league 2) you have a excellent understanding of what is happening in the game, 3) you remember every bit* of detail of the game.

* I think we can allow large majority in place of "every" in these cases.

 

[Hockey Sports Fan]

The human brain simply doesn't work that way. It's not designed to interpret vast amounts of boring data like that. It sifts through the boring parts and finds exciting things to latch on to. That's the whole crux of the "seen him bad" argument that Tyler Dellow would make when it came to player evaluations.

I'm with extensive training and practice you could probably get to a point where you could make somewhat-educated decisions using only your eyes, but like...why bother? Plus, you'd probably have to look through the stats to find out what's important to look for in the first place. That's the whole point of statistical analysis. It streamlines the process and eliminates any sort of error or biases that come part-and-parcel with the eye test.

 

[devilsblood]

The human brain simply doesn't work that way. It's not designed to interpret vast amounts of boring data like that. It sifts through the boring parts and finds exciting things to latch on to. That's the whole crux of the "seen him bad" argument that Tyler Dellow would make when it came to player evaluations.

I'm with extensive training and practice you could probably get to a point where you could make somewhat-educated decisions using only your eyes, but like...why bother? Plus, you'd probably have to look through the stats to find out what's important to look for in the first place. That's the whole point of statistical analysis. It streamlines the process and eliminates any sort of error or biases that come part-and-parcel with the eye test.

Well I was implying the extreme difficulty/unlikelyness of someone actually being able to do this.

But could you watch the game and see that Zajac is a good possession player? Yeah I think a Devils fan who watches all the time and really understands the game could see that. (though someone who watches all the time and does not understand the game so well might argue against Zajac's ability because of lack of points and/or big hit physicality/speed).

But it does have a major flaw in that it doesn't tell us how good he is relative to the rest of the league, sure he's good at it, but are there guys out there who are way better?

And the "eye exam" wouldn't help us much at all if we are talking about some guy on edmonton who we get to see a couple times a year?

Not to mention that "eye" people have nothing to back up their arguments besides opinion. And I think that's where Stray Cat gets upset. He's bringing discernable support to his argument, while others are trying to waive it away with non supported opinion.

 

[Zajacs Bowl Cut]

Well I was implying the extreme difficulty/unlikelyness of someone actually being able to do this.

But could you watch the game and see that Zajac is a good possession player? Yeah I think a Devils fan who watches all the time and really understands the game could see that. (though someone who watches all the time and does not understand the game so well might argue against Zajac's ability because of lack of points and/or big hit physicality/speed).

But it does have a major flaw in that it doesn't tell us how good he is relative to the rest of the league, sure he's good at it, but are there guys out there who are way better?

And the "eye exam" wouldn't help us much at all if we are talking about some guy on edmonton who we get to see a couple times a year?

Not to mention that "eye" people have nothing to back up their arguments besides opinion. And I think that's where Stray Cat gets upset. He's bringing discernable support to his argument, while others are trying to waive it away with non supported opinion.

yea but there are Devils fans that watch Zajac play that think he sucks. so no, I don't fully trust the eye test most of time.

but I do agree with your main point.

 

[devilsblood]

yea but there are Devils fans that watch Zajac play that think he sucks. so no, I don't fully trust the eye test most of time.

but I do agree with your main point.

My criteria for eye test effectiveness did include a high level understanding of the game.

and I'll admit, I wouldn't see Zajac's full effectiveness just by watching. Probably cause I go into pure fan mode(non analytical) and drink beers(forget half of what I'm watching sometimes). So these advanced stats do give me a greater appreciation for Zajac's game.

 

[manilaNJ]

New Jersey, Toronto and Edmonton all publicly hired advanced stats advocates into their front offices.

Los Angeles, Chicago and Pittsburgh have noted, for years now, that they've been stats friendly organizations.

At some point, I think, if you want to go deeper into following a team as a fan, having a general knowledge of how stats are applied will be somewhat beneficial.

There'll be a lot of GM moves and coaching decisions that might not align with the way we view the game, right now, but will make a lot more sense in the long run.

I saw a lot of people wondering why Stralman and Pouliot were so widely sought after and paid big this offseason.

A player like Pouliot doesn't seem like he's due for a payday, but he's the kind of depth player, possession driver that Edmonton has been lacking for years now. Overpaid by a bit, but not as outrageous as it appears to most people.

People often wonder why the game is "softening up" and the big hits and physicality is leaving. Well, the players who focus on this type of game have proved to be anchors in the modern NHL. Winning with these types of players on the ice, who lack the ability to score and make plays, is getting harder to do.

These questions become easier to answer when you start looking into the data.

There are still GMs and Coaches who will make their own oddball decisions and bad contract decisions... but more teams are slowly but surely using numbers to back up their moves.

 

[devilsblood]

I know this is a tangent, but I disagree that the game is becoming less physical, and we are seeing less big hits.

 

[Richer's Ghost]

New Jersey, Toronto and Edmonton all publicly hired advanced stats advocates into their front offices.

Los Angeles, Chicago and Pittsburgh have noted, for years now, that they've been stats friendly organizations.

At some point, I think, if you want to go deeper into following a team as a fan, having a general knowledge of how stats are applied will be somewhat beneficial.

There'll be a lot of GM moves and coaching decisions that might not align with the way we view the game, right now, but will make a lot more sense in the long run.

I saw a lot of people wondering why Stralman and Pouliot were so widely sought after and paid big this offseason.

A player like Pouliot doesn't seem like he's due for a payday, but he's the kind of depth player, possession driver that Edmonton has been lacking for years now. Overpaid by a bit, but not as outrageous as it appears to most people.

People often wonder why the game is "softening up" and the big hits and physicality is leaving. Well, the players who focus on this type of game have proved to be anchors in the modern NHL. Winning with these types of players on the ice, who lack the ability to score and make plays, is getting harder to do.

These questions become easier to answer when you start looking into the data.

There are still GMs and Coaches who will make their own oddball decisions and bad contract decisions... but more teams are slowly but surely using numbers to back up their moves.

Having the "advanced stats guy" position officially hired is much like the manufacturing world getting QS or ISO certification. It's not so much how much better it makes you, as much as the cost of not doing it.

Part form, part function, part being up to speed with how today's game is managed.

What happens when agents start negotiating contracts around individual metrics? Someone is going to be the first to have a CORSI bonus clause. I hope players (and coaches) don't focus on how they are playing the game to look better on paper the point it is detrimental to the team product.

 

[Hockey Sports Fan]

I was always skeptical of Pittsburgh and their analytics. I remember reading something (which I can't find anymore) about how they found some stat that said James Neal scored off a high percentage of shots from the scoring chance area (or something to that effect), so they thought he'd be a great fit for a good playmaker like Malkin.

But like...EVERYONE scores better from closer to the net, right? I really wish I could find this article, it made it seem like Pittsburgh was really misguidedly focused on shot quality.

On that note, Neal in Nashville is another case I'm really interested to see unfold next season.

 

[Zajacs Bowl Cut]

Having the "advanced stats guy" position officially hired is much like the manufacturing world getting QS or ISO certification. It's not so much how much better it makes you, as much as the cost of not doing it.

Part form, part function, part being up to speed with how today's game is managed.

What happens when agents start negotiating contracts around individual metrics? Someone is going to be the first to have a CORSI bonus clause. I hope players (and coaches) don't focus on how they are playing the game to look better on paper the point it is detrimental to the team product.

find me players with good advanced metrics that are detrimental to a team.

 

[Hockey Sports Fan]

find me players with good advanced metrics that are detrimental to a team.

If only Haggerty were here to tell you about how the Bruins lost because of Benoit Pouliot...

 

[Feed Me A Stray Cat]

Having the "advanced stats guy" position officially hired is much like the manufacturing world getting QS or ISO certification. It's not so much how much better it makes you, as much as the cost of not doing it.

Part form, part function, part being up to speed with how today's game is managed.

What happens when agents start negotiating contracts around individual metrics? Someone is going to be the first to have a CORSI bonus clause. I hope players (and coaches) don't focus on how they are playing the game to look better on paper the point it is detrimental to the team product.

I don't think you need to worry about that.

Most guys who know their **** realize CORSI is limited in what it can do. The more interesting work is coming with things like zone exits, zone entries, passing stats, etc.

 

[manilaNJ]

Having the "advanced stats guy" position officially hired is much like the manufacturing world getting QS or ISO certification. It's not so much how much better it makes you, as much as the cost of not doing it.

Part form, part function, part being up to speed with how today's game is managed.

What happens when agents start negotiating contracts around individual metrics? Someone is going to be the first to have a CORSI bonus clause. I hope players (and coaches) don't focus on how they are playing the game to look better on paper the point it is detrimental to the team product.

I don't think it's quite the slippery slope it seems to be.
I don't think any team will or should just sort by a players Corsi to start their summer shopping or trade deadline wishlist.

In terms of "new school" stats, these guys are being hired in advisory position, as they should be. Dubas is an AGM, Mehta reports directly to Lou who'll evaluate the information he's given from all his departments, Dellow will be one of many people reporting to Coach Eakins with his findings.

The current range of stats can help teams out in a few ways.

It can help player valuation by providing teams with notable value players and help put the brakes on a trade that might buy high and/or sell low.

It can help on ice evaluation by pinpointing situations where certain players fail/succeed and how to intelligently play your matchups.
It can also put up red flags when short term success will be heading for a big drop off. Think: stats guys yelling at Toronto this season that they were about to derail, but management and coaching decided to stay the course.

There's a lot that it can do, and it's important, like with all else, to not stretch the numbers beyond their limitations. If a team has a grasp on what stats can and can't tell them, they'll succeed. If they expect them to be an end all, be all, no context necessary guide.. they'll fail that way.

I've said it a few times, but to me, stats are a compass, not a map.

 

[manilaNJ]

I was always skeptical of Pittsburgh and their analytics. I remember reading something (which I can't find anymore) about how they found some stat that said James Neal scored off a high percentage of shots from the scoring chance area (or something to that effect), so they thought he'd be a great fit for a good playmaker like Malkin.

But like...EVERYONE scores better from closer to the net, right? I really wish I could find this article, it made it seem like Pittsburgh was really misguidedly focused on shot quality.

On that note, Neal in Nashville is another case I'm really interested to see unfold next season.

I was actually going to note in that post that it's funny that Pittsburgh has always been questioned about their statistical methods and that they've been one of the worst of applying whatever stats their using when it matters. But I felt like I was typing too much.... again.

I know what you're talking about. And go figure, I think there was another great Dellow article about Snake Oil Salesmen and Pittsburgh was a focal point.

 

[Richer's Ghost]

I don't think you need to worry about that.

Most guys who know their **** realize CORSI is limited in what it can do. The more interesting work is coming with things like zone exits, zone entries, passing stats, etc.

I think you skipped over my larger point but that's fine.

 

[Zajacs Bowl Cut]

I think you skipped over my larger point but that's fine.

I'll ask again- find me players who look good on paper (advanced metrics wise) that are detrimental to their teams.

 

[manilaNJ]

This is a great, albeit wordy, write up from Pension Plan Puppets.

Hockey Analytics: Why They Help And What's Coming Next

I think the interesting part is the last section on the future of what type of information will be accessible and debatable.

But here are a couple of snippets related to the discussion we're having here:

On why they're becoming more relevant:

The key point of this shift is to improve decision making. We now have lots of information we didn't have before and we can sift through it and find answers to questions that we haven't been asking. Often we stumble across things in the information that we didn't even know were happening, and this leads us to other areas of interesting analysis which furthers our understanding of what's happening.

And on using advanced stats to further player evaluation:

A single NHL game generally has around 80-90 shot attempts at 5v5, 40-50 shots on goal at 5v5, 30-40 total scoring chances, and only 4 or 5 goals at 5v5. What you can logically infer from this information is that shot attempts occur more frequently than shots on goal and scoring chances, and vastly more frequently than goals.

Scoring chances and goals are excellent to track, but a problem arises when we consider what happens in between these events. The amount of randomness and noise present in individual series of plays in hockey extends to individual games and beyond - it interferes with accurate assessment beyond individual seasons worth of play. Consider that for an entire NHL team - skill doesn't overtake luck in terms of it's impact on the NHL standings until the season is almost 90% complete. For individual players (goaltenders or skaters) the best assessments we have indicate that we require over 3 complete seasons of play to be relatively confident in assessments of their true NHL talent level.

 

[Richer's Ghost]

I'll ask again- find me players who look good on paper (advanced metrics wise) that are detrimental to their teams.

You'd have to do that - I don't bother with advanced stats in general because I think it gets too granular and can be misleading when you dismiss 8 stats and focus on just the 2 that say what you want them to say. But that also wasn't the point - the point was creating a system where advanced metrics analysis leads to playing to pad their individual metrics over the team's benefit because there's a financial incentive to do so (especially if their team is not a playoff one).

It's about creating an unintentional conflict where goals are not aligned for the individual and the organization. They may both point in general direction, not necessarily in opposition, but not harmonious either.

I think NFL contracts are farther down this road than the NHL but I see the potential for performance based incentives to allow this to creep in as advanced metrics are becoming more and more common. Baseball seems to be the grand master of advanced metrics but it's unique in that there's only 1 offensive player in play most of the (baserunners aside) time so it lends itself to a more individual analysis than football with set plays or most of all, hockey which is so much read and react by both offense and defense simultaneously and is continuous in flow without a whistle every 6-12 seconds to reset.

Not to say this is happening now or even in the next few years but it's a progression that begins with labels and rankings. Those lead to comparisons and comparisons are where contracts are negotiated. Some slick agent will eventually connect his client to being a "top 10" of something outside goals and assists (or blocked shots and hits) and the more advanced stats become the norm, the quicker that will happen.

 

[Feed Me A Stray Cat]

I think you skipped over my larger point but that's fine.

The larger point is that if teams begin accepting analytics/hiring analytical guys, they will see through the BS if agents try to spin stats a certain way. Incentive based contracts are actually quite smart and far more efficient (provided the incentives are good), because it directly aligns team and player interests.

The result will certainly be no worse than agents inflating player value because of crap like game winning goals.

 

[Richer's Ghost]

I've said it a few times, but to me, stats are a compass, not a map.

I like that.

Kind of like a trip to the eye doctor - which is better - A or B. Move and adjust relative to the results.

 

[Richer's Ghost]

The larger point is that if teams begin accepting analytics/hiring analytical guys, they will see through the BS if agents try to spin stats a certain way.

The result will certainly be no worse than agents inflating player value because of crap like game winning goals.

For every voice of reason though - there's a GM that ignores one. Do you think no GM or owner has been duped by that GWG or similar argument before? I think we know which GMs in recent years tend to hallucinate when it comes to tangible talent vs. agent's puffery. And it only takes a couple of bad contracts to set the market value and ruin the economy of free agents.

 

[manilaNJ]

You'd have to do that - I don't bother with advanced stats in general because I think it gets too granular and can be misleading when you dismiss 8 stats and focus on just the 2 that say what you want them to say. But that also wasn't the point - the point was creating a system where advanced metrics analysis leads to playing to pad their individual metrics over the team's benefit because there's a financial incentive to do so (especially if their team is not a playoff one).

It's about creating an unintentional conflict where goals are not aligned for the individual and the organization. They may both point in general direction, not necessarily in opposition, but not harmonious either.

I think NFL contracts are farther down this road than the NHL but I see the potential for performance based incentives to allow this to creep in as advanced metrics are becoming more and more common. Baseball seems to be the grand master of advanced metrics but it's unique in that there's only 1 offensive player in play most of the (baserunners aside) time so it lends itself to a more individual analysis than football with set plays or most of all, hockey which is so much read and react by both offense and defense simultaneously and is continuous in flow without a whistle every 6-12 seconds to reset.

Not to say this is happening now or even in the next few years but it's a progression that begins with labels and rankings. Those lead to comparisons and comparisons are where contracts are negotiated. Some slick agent will eventually connect his client to being a "top 10" of something outside goals and assists (or blocked shots and hits) and the more advanced stats become the norm, the quicker that will happen.

There can be a lot of value to looking beyond goals and hits - when applied properly and with context. This is where we start to hit a blurry line, but the more public teams make their stats hirings, the less likely they'll be using some of the shadier, simple solution philosophies.

If you had metrics that could help the team find undervalued free agents for cheap? Maybe a guy who was on the verge of breaking out offensively, but his numbers didn't seem to be eye popping?

Or if you had metrics that stopped your team from signing a player who lucked into a bunch of goals and was due to fall on his face?

Or maybe opened up the possibility of trading a player, who may look like he's contributing on ice, but in reality is hurting the team in other ways?

Having these advantages can honestly help the team to be better, again, if you're not using them beyond their limitations.
On the flip side of having a slick agent from saying, "This is a Corsi superstar, sign him to the same contract you would as your 40 goal scorer" -- it's smart to have someone on the other side of the conversation, in the front offices, who can provide worthy weight and context to all the numbers going back and forth.

In most cases, you won't find someone whose advanced stats over value as opposed to under value (unless you're looking at really small sample sizes, and most stats oriented people warn against that).

 

[Feed Me A Stray Cat]

For every voice of reason though - there's a GM that ignores one. Do you think no GM or owner has been duped by that GWG or similar argument before? I think we know which GMs in recent years tend to hallucinate when it comes to tangible talent vs. agent's puffery. And it only takes a couple of bad contracts to set the market value and ruin the economy of free agents.

Market value is not set by a handful of GMs. Market value is set by the salary cap. You might have one off-season where a select group of players are overpaid, however the result will be another group of players being underpaid. In totality, the teams are not worse off in regards to spending.

Agents will always twist numbers in a way to benefit their client. They currently twist things like hits, blocked shots and game winning goals, which have little value. I'm sure some are already twisting Corsi. That's a reality that's always been and will never go away.

 

[Richer's Ghost]

You're telling me market value isn't set by what someone is will to pay? A handful of GMs most certainly set the market rate each year when they sign the next "highest paid" player at his position because only a handful of those players exist and you can't un-ring that bell. And those contracts redefine what the league average is. The salary cap doesn't decide that - the GM's do by how much they will pay each player. The cap just limits how much they can spend in total. Just like the housing market - comparables and the most recent deals set the expectation. The final number may be negotiated from there but it sets the range.

 

[Feed Me A Stray Cat]

You're telling me market value isn't set by what someone is will to pay? A handful of GMs most certainly set the market rate each year when they sign the next "highest paid" player at his position because only a handful of those players exist and you can't un-ring that bell. And those contracts redefine what the league average is. The salary cap doesn't decide that - the GM's do by how much they will pay each player. The cap just limits how much they can spend in total. Just like the housing market - comparables and the most recent deals set the expectation. The final number may be negotiated from there but it sets the range.

The bolded statement is not true. The league average salary is defined by the salary cap, which is driven by revenue. If the salary cap stays the same, the league average salary will more or less stay the same, regardless if some GMs get crazy one summer. The only real inflation would come from certain lower revenue owners feeling like they need to spend beyond their means to be competitive.

However, if league revenue were to decrease, the salary cap would decrease, any precedent set by gyms would be immediately trumped by the lower salary cap, and salaries would go down.

There is a finite amount of money that can be spent on players in the NHL. This is not the case for the housing market, as investors can freely shift billions from one investment to another.