article: NHL Offseason by the Numbers: Which Teams Have Improved the Most (and Least)?

[LindG1000]

And the average fan is meant to understand this without a degree in advanced mathematics? My fucking eyes glazed over just scrolling through this looking for the word "wrong."

 

I know you're a crotchety old man, but it doesn't take a mathematics degree to use the "find" feature in your browser.

 

I'll get off your lawn now.

[Keirik]

Great now I want to watch that Clint Eastwood movie

[Phil]

I know you're a crotchety old man, but it doesn't take a mathematics degree to use the "find" feature in your browser.

 

I'll get off your lawn now.

 

Literally just did. While trying to mentally map the teams he was most wrong about, what I found even more astonishing is the sheer volume of times he literally uses the word, including this gem of a line: "Essentially, I was continuously wrong during the season."

 

I'll go back to just dismissing it entirely, the same way that entire community dismisses anything that "can't be quantified."

[LindG1000]

Literally just did. While trying to mentally map the teams he was most wrong about, what I found even more astonishing is the sheer volume of times he literally uses the word, including this gem of a line: "Essentially, I was continuously wrong during the season."

 

I'll go back to just dismissing it entirely, the same way that entire community dismisses anything that "can't be quantified."

 

Much like data requires context, so too does the quote you're pulling:

 

They were a bad team and I realized that far too late, sinking way too much money backing them. Then I would lose more money betting against them. I can’t recall a more infuriating team that would lose so many games they shouldn’t have and then win when it’s not expected. Essentially, I was continuously wrong during the season.

 

#charted

[Phil]

This is the inherent problem with advanced metric statical models.

They skew based upon what the designer places value on in many instances, and other areas are hard to quantify statistically, so they get little to no representation and value.

 

It’s just nonsense that uses mathematics as it’s cloak.

 

Right, and the way the public digests it, they're effectively given negative value, which is why every single time a "grit" guy gets signed, it's met with endless bitching and moaning about "tuffness." Which makes me literally want to ask out loud — did you people not watch the fucking playoffs? Stop trying to build a Presidents Trophy-winner and start trying to build a Stanley Cup-winner.

[Phil]

Much like data requires context, so too does the quote you're pulling:

 

#charted

 

"That’s the beauty of trying to predict a game as chaotic as hockey — you have to be prepared to be wrong a lot."

 

Habs: "The biggest thing I was wrong about was doubting Carey Price in the playoffs."

 

Isles: "The Islanders indeed continued their 2020 playoff showing and looked like an analytics darling for much of the season, cruising to the playoffs as expected. The model was wrong, but thankfully I hedged after seeing the 2020 playoff run."

 

Oilers: "I was absolutely wrong about McDavid and was glad to be proven wrong that yes, he can play better defence. I was wrong about Mike Smith as well, but so was everyone else. I was wrong about Darnell Nurse, who played like a legit No. 1 defender this year."

 

Flames: "Tanev ended up playing like one of the 15 best defencemen in hockey thanks to an incredible defensive performance. I was not kind to that signing and turned out very wrong about it."

 

Blue Jackets: "The model, like everyone else, was very wrong about the Blue Jackets.

 

It overrated the impact Max Domi would have. It overrated how good Seth Jones would be (a truly hilarious sentence to type considering the preseason discourse surrounding Jones)."

 

#farted

[Sod16]

Between a commentator who thinks 35 year old scratch Ryan Reeves will be a plus for the Rangers and a commentator who thinks he's spent and not a positive addition, I'd have to say I respect the latter's opinion more.

 

I don't agree with everything this guy says, but I'm also reluctant to give much credence to opinions that gestate in a Rangers' bubble.

[LindG1000]

"That’s the beauty of trying to predict a game as chaotic as hockey — you have to be prepared to be wrong a lot."

 

Habs: "The biggest thing I was wrong about was doubting Carey Price in the playoffs."

 

Isles: "The Islanders indeed continued their 2020 playoff showing and looked like an analytics darling for much of the season, cruising to the playoffs as expected. The model was wrong, but thankfully I hedged after seeing the 2020 playoff run."

 

Oilers: "I was absolutely wrong about McDavid and was glad to be proven wrong that yes, he can play better defence. I was wrong about Mike Smith as well, but so was everyone else. I was wrong about Darnell Nurse, who played like a legit No. 1 defender this year."

 

Flames: "Tanev ended up playing like one of the 15 best defencemen in hockey thanks to an incredible defensive performance. I was not kind to that signing and turned out very wrong about it."

 

Blue Jackets: "The model, like everyone else, was very wrong about the Blue Jackets.

 

It overrated the impact Max Domi would have. It overrated how good Seth Jones would be (a truly hilarious sentence to type considering the preseason discourse surrounding Jones)."

 

#farted

 

And then you have....

 

Boston: Boston struggled relative to expectations to start the season but eventually ended up almost exactly where it was expected to be. The Bruins were projected for 71 points and finished with 73, though in a tough division that was only enough for the third seed.

 

Jets: When it comes to such an average team, opinions don’t have to be so extreme. I was probably more right than wrong about the Jets this year and they only ended up two points higher than their preseason projection of 61 points. That’s pretty damn close.

 

Toronto: The model has nothing in the way of measuring intangibles or killer instinct come playoff time — something the Leafs seemed to have fixed with all their moves to bring in grizzled veterans — so that will always be a blind spot. I’ve never been one to put too much stock into those things, but it’s hard not to after Toronto’s latest collapse. It inspired me to start working on changes to next year’s model.

 

Nashville: They were somewhere from good to great, salvaging what looked like a lost season to make the playoffs. In fourth. With 64 points. Exactly as expected.

 

Rangers: The Rangers were projected for 59.9 points and ended up with 60. Right on the money … except it’s worth noting that projection changed to about 61.6 points once lineups were finalized. Either way, extremely close. There were a few who believed the model underrated him (Zibanejad) after his red hot finish to the 2019-20 season, but it turned out to be pretty accurate in forecasting regression. He came on strong after a sluggish start, but still finished well below his pace from the previous season. Also, it’s worth noting the Adam Fox love started really early, with the model earmarking him as elite before the season started. His projected 2.4 wins was the 10th highest at the time so the numbers were clearly on to something there.

 

Anyone who can read can cherrypick statements that support their stances; at least give the guy some credit for acknowledging the weaknesses in his model, working to fix them, and holding himself accountable to constantly improving it over the past five years. I think getting the standings for a chaotic as fuck sport like hockey within 7.4 points average error is an exceptional feat.

[Ozzy]

That sounds like an awful lot of time and energy, G man!!! I'd rather just eyeball test it!! :rofl:

[Pete]

If only there were a long series of articles about the evolution of the model, what he got right and wrong and how he has adjusted his model in years past.

 

...you can guess where this is going.

 

It's almost like...When all the polling was wrong in 2016, and the pollsters who got it wrong "fixed it!" for 2020, and then got it wrong again in 2020...

[Phil]

It's almost like...When all the polling was wrong in 2016, and the pollsters who got it wrong "fixed it!" for 2020, and then got it wrong again in 2020...

 

This is... painfully accurate.

[RangersIn7]

Right, and the way the public digests it, they're effectively given negative value, which is why every single time a "grit" guy gets signed, it's met with endless bitching and moaning about "tuffness." Which makes me literally want to ask out loud — did you people not watch the fucking playoffs? Stop trying to build a Presidents Trophy-winner and start trying to build a Stanley Cup-winner.

 

And for the record, by nearly every single metric that evaluates defensive value/effectiveness, Goodrow is among the best defensive forwards in all of hockey.

[Phil]

And for the record, by nearly every single metric that evaluates defensive value/effectiveness, Goodrow is among the best defensive forwards in all of hockey.

 

Nah, blue bar good, red bar bad. He doesn't throw everything on net, so he sucks. Didn't you get that memo?

[RangersIn7]

Nah, blue bar good, red bar bad. He doesn't throw everything on net, so he sucks. Didn't you get that memo?

 

I did.

 

The one that says goals die when he’s on the ice.

I think some other people may have missed it though.

You know, the people who forget that keeping the puck out of your net helps win games too.

[Winter]

Rangers off-season plan was great but imo execution of it left a lot to be desired.

[Phil]

I did.

 

The one that says goals die when he’s on the ice.

I think some other people may have missed it though.

You know, the people who forget that keeping the puck out of your net helps win games too.

 

It's almost like hockey is played in two directions or something? Crazy.

[The Dude]

I've never been a fan of Dom Luszczyszyn as I'm more of the mindset that analytics are a tool in the box rather than a religion.. Needless to say ol' Dom not a big fan of the off season so far.

 

 

 

https://theathletic.com/2753099/2021/08/05/nhl-offseason-by-the-numbers-which-teams-have-improved-the-most-and-least-2/

 

You can't say he's wrong..

 

Well I can't. Because I totally agree.

[RangersIn7]

Rangers off-season plan was great but imo execution of it left a lot to be desired.

 

You say “was” and “left” as though they are done.

Probably not.

[RangersIn7]

It's almost like hockey is played in two directions or something? Crazy.

 

Insane that that gets missed.

Who wins a cup without a checking line?

Without getting viable minutes and contributions of all types from lines 3 & 4?

[The Dude]

Seriously. It's amazing these models continue to get the press they do like they're some kind of unflappable prognostication.

 

You know who else scored poorly with Dom's model? Two of the final four teams in the playoffs last season. Funny how that works.

 

Oh, and just for shits and gigs, a bunch of fucking lottery teams added "wins" according to his model last year, too. The Devils (4.8 wins), The Wings (5.9 wins), etc. all scored "well." Look what good it did them.

 

Meanwhile, Montreal (0.5 wins) and the NYI (-1.7 wins) laughed all the way to the "Conference Finals."

 

It's almost like calculating every single player/team by a universal role of "scorer" is a shitty way to "model" a successful team, eh? Who'd have thought.

 

I hope the Ranger snuff out every analytics darling this year and punch their way to the Cup like the Ducks did, just to watch all these blue bar good red bar bad robots melt down.

 

Not to nitpick, but isn't Drury supposed to be a guy who believes heavily in analytics? Didn't we have a back and forth on how Gallant doesn't dislike them (as I suggested) and that Drury kinda would get his own type of players (analytics darlings is what I figured) and Gallant would live with it?

 

Seems more like Gallant is hand picking a bunch of these moves. Which I have no problem with him doing. I trust his vision.

 

I myself think he'd rather make bigger changes to the top 6 and get harder to play against guys, who also play on the top lines. Which I think is what's really needed. I don't see how adding checking forwards for the bottom 6 is going to have an impact on other teams top lines. Having a line to play against the Isles 4th line is going to accomplish what exactly? It'll be entertaining as fuck, but in the grand scheme, is that REALLY harder to play against?

 

We are concentrating on Matt Martin and the likes? If Reaves is on the ice against Wilson then Washington is winning that matchup and likely scoring a goal or drawing a penalty. Harder to play against?

 

I dunno, I'm not seeing how these additions will add to wins. I hope I'm wrong and these guys can be matched up against anyone in the league and shut them down. I'm just used to seeing Ranger coaches match up their top lines against opponents top lines. Maybe this changes? THEN all this makes sense.

[Dunny]

Who just picks the skinny, fast guys when they play Ice Hockey on NES? That's what this guy would do.

[Bugg]

This premise is assuming that the only thing that matters when determining wins and losses is individual player stats... And, as we've seen, you can have three superstars put up a billion points and the team still stinks, so obviously there's more to it.

 

Does not consider that Kakko, Yukon and Krav are all expected to take the next step, such the loss of Buch won't be a big deal if they do.

[josh]

Who just picks the skinny, fast guys when they play Ice Hockey on NES? That's what this guy would do.

 

How can you make that decision without graphs and charts?

 

Hold on, let me check twitter

[Sod16]

I myself think he'd rather make bigger changes to the top 6 and get harder to play against guys, who also play on the top lines. Which I think is what's really needed. I don't see how adding checking forwards for the bottom 6 is going to have an impact on other teams top lines. Having a line to play against the Isles 4th line is going to accomplish what exactly? It'll be entertaining as fuck, but in the grand scheme, is that REALLY harder to play against?

 

We are concentrating on Matt Martin and the likes? If Reaves is on the ice against Wilson then Washington is winning that matchup and likely scoring a goal or drawing a penalty. Harder to play against?

 

Well taken points. What is needed is a tough forward with some skill to play on Zib's line, not Ryan Reeves to end up -2 for the night with a fight against the other team's bottom six. We need someone to do what Vic Hadfield did for Ratelle and Gilbert (who will now be Zib and LaF).

[jsm7302]

Blais played up the lineup in St.Louis. Can't we expect the same from Goodrow? These are players that are bottom six with potential to move up with the right linemates. I think Drury has made GREAT moves and I think this team undoubtedly adds wins in the W column.