Analytics should be the only way to judge players. Eye tests are faulty because all humans are subject to biases. This is proven in science and doesn't apply to only hockey: it applies to everything.

It's also one of the first things you learn in any intro to Psych class.

All coaches will adapt or slowly lose their jobs to people who do them better. Vigneault has gotten favorable results even though he's deployed Girardi poorly. A professional can still make mistakes.
- James_Tanner

So sad, but not unexpected. Thankfully, the hockey world, and NHL coaches and GM's know better. What you obviously don't understand about analytics, is that they can only tell you what happened, not why it happened. Analytics require analysis, and are also subjected greatly to human biases.
Your blog today proved that. It is literally impossible for analytics to be the sole source of judging players.

Because team success is based on individual aptitude collectively.
- rangerdanger94

So sad, but not unexpected. Thankfully, the hockey world, and NHL coaches and GM's know better. What you obviously don't understand about analytics, is that they can only tell you what happened, not why it happened. Analytics require analysis, and are also subjected greatly to human biases.
Your blog today proved that.
- MJL

What? How so? There was literally zero bias in that blog at all...

- jtommyt

A team is a collection of individuals. Each individual must perform well for the team to be successful.

What? How so? There was literally zero bias in that blog at all...
- jtommyt

I disagree. His blog showed that analytics require analysis. I think he got some of the analysis incorrect, probably due to his bias against Girardi.

I don't think that at all, but it's pretty irrelevant to the conversation. If Girardi is a legitimate top pairing defenseman, he's going to face the same level of competition that every other top pairing defenseman in the league faces.

His performance can therefore be judged apples to apples with other top pairing defensemen. When you do that, it becomes obvious that Girardi under performs in comparison to his peers.
- jtommyt

But you base it mostly on possession and shots against. I don't care if Crosby spends his entire shift in the offensive zone, if he has no scoring chances or a bad angle shot I consider it a job well done by Girardi. To me that's where the analytics fail.

A team is a collection of individuals. Each individual must perform well for the team to be successful.
- rangerdanger94

That's not actually true.

Enough individuals need to perform well enough to beat the opposition. There's plenty of room for some individuals to suck while being part of a successful team.

That's not actually true.

Enough individuals need to perform well enough to beat the opposition. There's plenty of room for some individuals to suck while being part of a successful team.
- jtommyt

Well I wont disagree with this. We did have Glass play most of the year.

But you base it mostly on possession and shots against. I don't care if Crosby spends his entire shift in the offensive zone, if he has no scoring chances or a bad angle shot I consider it a job well done by Girardi. To me that's where the analytics fail.
- mdw7413

You don't think Crosby can score from a bad angle?

Would you not prefer to spend the entirety of that same shift in the Penguins zone instead? There's a much lower chance that he'll score from his own end, no?

Well I wont disagree with this. We did have Glass play most of the year.
- mdw7413

A decision made by the same coach that chooses to deploy Girardi on the top pairing.

I disagree. His blog showed that analytics require analysis. I think he got some of the analysis incorrect, probably due to his bias against Girardi.
- MJL

He didn't even make a judgement call. He simply pointed out the discrepancy that occurs when you look at the same set of stats in 2 different circumstances. When McDonagh and Girardi play together, and when they play apart.

It's math... you can't argue it.

You don't think Crosby can score from a bad angle?

Would you not prefer to spend the entirety of that same shift in the Penguins zone instead? There's a much lower chance that he'll score from his own end, no?
- jtommyt

Well of course I would, but realistically the best forward in the NHL is going to be in your zone a lot more than you will be in his when he is on the ice. Doesn't matter who is out there against him.

But you base it mostly on possession and shots against. I don't care if Crosby spends his entire shift in the offensive zone, if he has no scoring chances or a bad angle shot I consider it a job well done by Girardi. To me that's where the analytics fail.
- mdw7413

It's all about the data collected. +/- used to be a popular stat, it has it's flaws which is why Corsi was developed, to improve upon that, but using shots, because it's a more frequent event in a hockey game, compared to goals. Ironically, Corsi has the same flaws when used to look at individual players. Years ago, Roger Nielsen used video to come up with an actual +/- for a player. Only giving a + or a - if he actually contributed to the event, rather than just simply being on the ice. It's the same for NHL teams. They are so far past corsi and the stats that are available to the public, because they collect better data. As you alluded to, not all shots are created equal.

A decision made by the same coach that chooses to deploy Girardi on the top pairing.
- jtommyt

In AV's defense we really didnt have anyone for the 4th line anyway. This year most likely a different story.

That's not actually true.

Enough individuals need to perform well enough to beat the opposition. There's plenty of room for some individuals to suck while being part of a successful team.
- jtommyt

But if it's a top pairing defenseman who goes against the other teams top opposition for 25 minutes a game, doesn't that person's play impact the outcome of the game much more than Tanner Glass's?

I mean, Alex Ovechkin scored 2 goals in the first 2 games of the 2nd round against us in the playoffs. Guess who he victimized on both of those goals? Dan Boyle. And when I say victimized, I mean absolutely victimized (not that Dan Boyle was just on the ice). He scored 1 goal in the remaining 5 games with Dan Girardi on his ass.

He didn't even make a judgement call. He simply pointed out the discrepancy that occurs when you look at the same set of stats in 2 different circumstances. When McDonagh and Girardi play together, and when they play apart.

It's math... you can't argue it.
- jtommyt

TJ Brodie's CF% is 45.3. Therefore he is an ECHLer.

But if it's a top pairing defenseman who goes against the other teams top opposition for 25 minutes a game, doesn't that person's play impact the outcome of the game much more than Tanner Glass's?

I mean, Alex Ovechkin scored 2 goals in the first 2 games of the 2nd round against us in the playoffs. Guess who he victimized on both of those goals? Dan Boyle. And when I say victimized, I mean absolutely victimized (not that Dan Boyle was just on the ice). He scored 1 goal in the remaining 5 games with Dan Girardi on his ass.
- rangerdanger94

To piggy back off of this, why do other coaches try to get their top line away from Girardi & McDonagh? hmmmm

He didn't even make a judgement call. He simply pointed out the discrepancy that occurs when you look at the same set of stats in 2 different circumstances. When McDonagh and Girardi play together, and when they play apart.

It's math... you can't argue it.
- jtommyt

I absolutely can argue it. It's far more than just math.

"Ergo: Girardi doesn't have bad stats because he's always playing tougher minutes, he actually forces himself to play tougher minutes by being bad."

I don't think that at all, but it's pretty irrelevant to the conversation. If Girardi is a legitimate top pairing defenseman, he's going to face the same level of competition that every other top pairing defenseman in the league faces.

His performance can therefore be judged apples to apples with other top pairing defensemen. When you do that, it becomes obvious that Girardi under performs in comparison to his peers.
- jtommyt

Under performs in what category? Corsi is not the be all that ends all in player evaluation. It's also never apples to apples. Far too many variables in the game of hockey.

So sad, but not unexpected. Thankfully, the hockey world, and NHL coaches and GM's know better. What you obviously don't understand about analytics, is that they can only tell you what happened, not why it happened. Analytics require analysis, and are also subjected greatly to human biases.
Your blog today proved that. It is literally impossible for analytics to be the sole source of judging players.
- MJL

Yeah, everyone who has a different opinion than you do just doesn't get it.

I would trust a mathematician who has only looked at data and never seen a game over the world's most trusted scout.

To piggy back off of this, why do other coaches try to get their top line away from Girardi & McDonagh? hmmmm
- mdw7413

Giordano's CF% is 48.9%. Dan Boyle is a much better defenseman than Giordano.

You don't think Crosby can score from a bad angle?

Would you not prefer to spend the entirety of that same shift in the Penguins zone instead? There's a much lower chance that he'll score from his own end, no?
- jtommyt

Would you rather have Crosby take a shot from a bad angle, or from directly in the slot? Crosby is going to get his shots in most games, no matter what. If a defenseman can limit the quality of those shots, he's done his job. There's a reason why NHL teams chart quality of shots given up.

I disagree. His blog showed that analytics require analysis. I think he got some of the analysis incorrect, probably due to his bias against Girardi.
- MJL

First of all, everything requires analysis. Literally everything.

Second, if something is incorrect, or you think it is, let's discuss, that's what were here for.

Third, what I am giving you here is theory. I didn't use hardly any stats to prove anything, I'm simply discussing what observations I make off other people's work. I'm not trying to prove anything either, just discussing some of the implications of the data we have access to. (Someone had to compile that data on Stats.HockeyAnalysis.com)

Yeah, everyone who has a different opinion than you do just doesn't get it.

I would trust a mathematician who has only looked at data and never seen a game over the world's most trusted scout.
- James_Tanner

Thats crazy talk bro.

But you base it mostly on possession and shots against. I don't care if Crosby spends his entire shift in the offensive zone, if he has no scoring chances or a bad angle shot I consider it a job well done by Girardi. To me that's where the analytics fail.
- mdw7413

But if he didn't have any shot attempts, then Girardi would have good stats.