Tyler Seguin the beauty tells it how it is

[Dr Pepper]

But the dude has a Stanley Cup ring and you don't. He's been coached by elite coaches and you haven't. He knows what is needed to win more than you do, based on his actual experience, even if he's not able to verbalize it. You can look at stats all you want and call him an idiot but you're on the outside looking in.

You're talking to a guy who still thinks Tyler Seguin actually tweeted that out, as opposed to someone else quoting him.

Not really worth the effort explaining it, IMO.

 

[Fixxer]

Whatever the guy is going to say, people will pick on him. Social media is such a plague for this. Hope you find your game Tyler!

 

[Dr Pepper]

Whatever the guy is going to say, people will pick on him. Social media is such a plague for this. Hope you find your game Tyler!

Pretty much.

Isn't groupthink fun?

 

[abo9]

Because numbers are more complex to understand and no one really knows exactly how to weigh their importance? To be clear, someone who doesn't know hockey very well but understands numbers might be able to evaluate guys better purely based on numbers, but if you're a scout worth your salt you don't really need the numbers. As I said, they can be a helpful tool to provide context to things you see on the ice, but they should never be a focal point of your evaluation, at least not at this time. Maybe if analytics improve in the future they can play a bigger role, but right now I don't see how it will help you in a significant way.

I think people in general are really overthinking it with the spreadsheets. Like for instance, yeah you can track certain statistics like zone entries, exits etc but if you don't watch on the ice every single one of those you miss vital context to those numbers. And if you're really going to watch every single play like that in close detail, what do you need the numbers for? You will get the full picture from the eye test and can just use a notepad to write down the trend you're seeing. Don't need a bunch of numbers in a spreadsheet to do that.

The thing is you can't possibly watch EVERY game for EVERY player on EVERY team, that goes for full time scouts, gm's, coaches and is even truer for fans who want to form an idea on a player they don't get to watch often.

I think that numbers on a spreadsheet can be a good way to capture a snapshot of a player over all games. You can go watch a player once or twice and be amazed, but then if he doesn't do anything the rest of the year how are you gonna know? If someone watched Gagner's 8 points game and never anything else, they might have given a lot to get him no? Ditto for Hertl and Eller.

Also, when talking about numbers, it's the sheer volume of them that makes them more powerful, and that's why it's very difficult to find meaningful measures in hockey, since you only get so many data points to look at. That's why it's easier to come up with measures for goalies too, they're on the ice for all shot against and all goals against. There is a lack of data points. In your example of zone entries, you could glean context from speed of entry, time of possession after entry, did the entry lead to a goal? Etc. Etc. Yes you do have to record all that when it happens, like any other data out there, but then you can put it all together in one place to get a snapshot of a player.

Now, I'm not saying analytics is the end-all be-all, it's extremely difficult to predict the result of a game on any given night, nevermind the Stanley Cup winner. But, really, who saw Dallas making the finals, analytics vs old-fashioned opinions? Who watched the games and said "Ahhh, I think Dallas is going far this year. In fact, it might be their year!". I don't think anyone predicted that lol

Edit: In fact, I just did a re-read of some predictions for the play-ins and 1st round, and most "experts" saw Dallas out IN THE FIRST ROUND.

 

[FirstRowUpperDeck]

What? A nice middle of the road, reasonable post on social media by abo9? 2020 really is a weird year.......

 

[SympathyForTheDevils]

But the dude has a Stanley Cup ring and you don't. He's been coached by elite coaches and you haven't. He knows what is needed to win more than you do, based on his actual experience, even if he's not able to verbalize it. You can look at stats all you want and call him an idiot but you're on the outside looking in.

I never said Seguin knew nothing about hockey. But I do think he knows nothing about analytics, given his response to that question.

Honestly, my post was probably unfair to Seguin. Statistical literacy is pretty poor in the general population; there's no reason to expect Seguin to do better than most people in that regard, especially since it's not exactly his job. I shouldn't have implied that he's dumb.

But this kind of take is so common now whenever a team that gets outshot wins, like it's somehow a refutation of analytics work. It's part of this weird narrative where "nerds" and their numbers are opposed to real-life hockey, which is nonsense, since these numbers are purely a reflection of real-life hockey in the first place. It's even weirder in the context of a question about goal-differential, a pretty straightforward stat that I haven't really seen used in analytics.

Normally I wouldn't mind, but I see this type of misunderstanding of statistics/probabilities often, in fields where it has bigger real-life consequences than just triggering a bunch of nerds. So it annoys me.

 

[MessierII]

You know what Tyler Seguin's least favorite analytic is?

Playoff goals scored.


In all seriousness, the reason advanced metrics like CF% and xGF% gained widespread use was because CURRENT goal-differential is not good predictor of FUTURE goal-differential.

CF% and xGF% both have higher correlation with future GF% and wins than current GF%.

But the correlation isn't 1:1. There are always going to be outliers. There are plenty of cinderella stories in sports history, and in a game as inherently random as hockey, sometimes worse teams are going to win because of a mix of hot goaltending, or the other team going cold, etc.

I’ve yet to see any compelling evidence that any analytic metric has future fortune telling capability.

 

[Butch 19]

Analytics serve a purpose but not all analytics tell the right story and not all are equally important.

I thought analytics were boring, and you make them sound even worse.

 

[serp]

I thought analytics were boring, and you make them sound even worse.

I mean a bunch of numbers are never very exciting unless you're a massive mathsnerd. Doesn't mean even if you're not you can't take something useful from them. Just don't take them as the be-all end-all .

 

[Paulinvancouver]

Agree 100%.
Proven analytics do though.
But we don’t have the sample size or the accuracy to prove them. If so, they’d be widely adopted stats on the NHL website and used widely.

 

[Boxscore]

Wow, I'm by no means a Seguin fan--but he's 100% correct.

 

[motopanekeku]

I assume most players look down on analytics especially the older ones. Think the newer generation might take them more serious but who knows.

I think every player wants to have good individual analytical numbers but yeah I'm sure some players (just like some posters here) are more drawn to that aspect of measuring success than others.

Most reasonable people find value in both the analytical numbers and the more immeasurable dynamics of on-ice performance. A team can be built using nothing but mathematics but I, and many others, believe that there is benefit to adding color that perhaps defies the numbers.

I think most reasonable, understanding people agree that analytics isn't everything. The people who believe solely in analytics should get jobs in that department so their specialized mathematical data can be utilized by the people who are actually building and developing rosters. The people who will put the analytical numbers into proper perspective.

Proper perspective.

 

[PoutineSp00nZ]

Pretty much.

Isn't groupthink fun?

Orwell called it. Except he thought we would need the thought police to enforce it. Turns out we just do it to ourselves

 

[DudeWhereIsMakar]

Probably misses his three dogs too much.

I couldn't own three dogs.

 

[ucanthanzalthetruth]

Statistics are meant to help complement, support or explain on-ice information. They are a supporting tool. Unfortunately, many people abuse them as the single source of an argument. Statistics without proper context and conferral with on-ice results are useless. This is why arbitrary stupid statements like "the numbers don't lie" is an ignorant approach to statistical usages.

Well said

 

[BenchBrawl]

He's right.

 

[haveandare]

Yesh they’re overrated by some. Imo it’s just a logical fact that having the puck more and taking more shots correlates to success. That doesn’t mean it’s a guarantee or the only thing that matters but obviously it’s important

 

[PaulD]

The thing is you can't possibly watch EVERY game for EVERY player on EVERY team, that goes for full time scouts, gm's, coaches and is even truer for fans who want to form an idea on a player they don't get to watch often.

I think that numbers on a spreadsheet can be a good way to capture a snapshot of a player over all games. You can go watch a player once or twice and be amazed, but then if he doesn't do anything the rest of the year how are you gonna know? If someone watched Gagner's 8 points game and never anything else, they might have given a lot to get him no? Ditto for Hertl and Eller.

Also, when talking about numbers, it's the sheer volume of them that makes them more powerful, and that's why it's very difficult to find meaningful measures in hockey, since you only get so many data points to look at. That's why it's easier to come up with measures for goalies too, they're on the ice for all shot against and all goals against. There is a lack of data points. In your example of zone entries, you could glean context from speed of entry, time of possession after entry, did the entry lead to a goal? Etc. Etc. Yes you do have to record all that when it happens, like any other data out there, but then you can put it all together in one place to get a snapshot of a player.

Now, I'm not saying analytics is the end-all be-all, it's extremely difficult to predict the result of a game on any given night, nevermind the Stanley Cup winner. But, really, who saw Dallas making the finals, analytics vs old-fashioned opinions? Who watched the games and said "Ahhh, I think Dallas is going far this year. In fact, it might be their year!". I don't think anyone predicted that lol

Edit: In fact, I just did a re-read of some predictions for the play-ins and 1st round, and most "experts" saw Dallas out IN THE FIRST ROUND.

 

[PaulD]

Wow, I'm by no means a Seguin fan--but he's 100% correct.

I agree with Tyler.

Now he needs to play better.

When was the last time he scored a goal?

 

[Sparty]

"Stats are for losers, the only stat that really matters is winning."

-Pat Narduzzi

 

[slapKing]

If Seguin is injured, he shouldn't be playing. Dude is doing nothing out there (exaggerating, but he's not doing good). If he isn't injured, he's a disgrace with his playoff play so far.

 

[Pip]

Seguin when the playoffs start

 

[Dr Pepper]

I agree with Tyler.

Now he needs to play better.

When was the last time he scored a goal?

Been too long.

Not even sure if he scored in the Vegas series.

 

[Hockeyholic]

He needs to step up. And quick.

 

[Sky04]

Tyler just said f*** stats altogether.