Don Granato Interim Coach for Sabres/Ellis and Girardi to the Bench. Steve Smith Fired

[Thorner]

3 minutes ago, Sabre fan said:

Skinner did however seem to really mesh with Jack (when he was healthy) and hopefully once we get Jack back (probably not til next year) we can put Skinner back with Jack and Sam and get the kind of production we got a few years back (which was great!). Obviously we desperately need Skinner to contribute at his salary...

I can't believe we are on the hook for 9 mil for another 6 years. I hope he does work with Jack (and he should play with him, if he does - where else can we put him?) or else they seriously need to find out if there's some combination of things we can attach to him to move him to Seattle, or somewhere. It'd probably be worth it just to convert that cap space into something. 

If he continues producing zero like he is now, it's probably worth more than a 1st round pick to get him out.

What an albatross

[JoeSchmoe]

On 4/7/2021 at 11:31 AM, LGR4GM said:

On 4/7/2021 at 11:31 AM, JoeSchmoe said:

The XGF% is lower under Granato than Krueger?!?! 

We couldn't get the puck to the slot under Krueger, nor did the team actually go there. 

That's why the HDCF is higher. Pucks to the slot.

I know you'll probably disagree, but I tend to think that once you start using other advanced stats to explain why another advanced stat isn't "working", you are in effect making the case why the eye test is the best stat of them all.

Again, I'll harken back to how humans can beat some of the world's most powerful super computers at chess. Moreover, the game of hockey is many orders more dynamic than chess, so using numeric methods to evaluate hockey players and teams would be even more difficult than creating a chessmaster.

The stats do have greater value when GMs are playing the game of comparing large numbers of players and are looking to mine for value picks and seeing how players compare using isolated metrics. This is related to how machines are better at processing large amounts of data than humans. If you're going to play the Moneyball, you're going to be these stats! But a smart GM might want to eyeball his picks before throwing down the cash for a contract.

 

Edited April 8, 2021 by JoeSchmoe

[Doohicksie]

The problem is the eye test is too variable.  I think Okposo looks great lately.  Weave thinks he sucks.  Who's right?  As a self-proclaimed numbers guy, you should see that looking at fancy stats could inform the discussion.

Stats in general are an attempt to quantifiably model a physical system.  Fancy stats are simply more complicated models.  There's nothing wrong with discussing the validity of those models or understanding their limitations.  They eye test didn't get us to the moon, keep aircraft flying, design safer cars; it was modeling of all those systems that got the job done.  Without modeling we'd still be killing pilots who broke the sound barrier.

 

 

 

 

(I'm one of those engineers too.)

[Weave]

Who the what now?

[JoeSchmoe]

20 minutes ago, Doohickie said:

The problem is the eye test is too variable

Variable isn't necessarily a bad thing. It's what you value when assessing a player.

Remember, the widely established XGF% shows that we should be worse under Granato than Krueger, which I think we can all universally agree is not close to being the case. But someone that favours high danger chances would say we're better under Granato.

No different than someone thinking Okposo has really gained speed, but his friend thinking he still doesn't create enough scoring chances.