Morgan Rielly

[Advanced stats]

What goes into calculating defensive point shares?

Calculating Point Shares | Hockey-Reference.com
And click on "crediting defensive point shares to players."

 

[Dreakmur]

Calculating Point Shares | Hockey-Reference.com
And click on "crediting defensive point shares to players."

Yep. I was just hoping that by asking the question, it would cause people to look it up, and that would cause them to realize what a useless stat it is.

 

[meefer]

Yep. I was just hoping that by asking the question, it would cause people to look it up, and that would cause them to realize what a useless stat it is.

I'm stats illiterate. Please fill me in on why it's a useless stat, as to make that comment, you must be the expert on the process.

 

[Dreakmur]

I'm stats illiterate. Please fill me in on why it's a useless stat, as to make that comment, you must be the expert on the process.

If you can’t even understand how a stat is calculated, that’s probably a good hint.

 

[meefer]

If you can’t even understand how a stat is calculated, that’s probably a good hint.

That you're talking out of your a**? Or that you can't complete a full sentence? Or that you're trying to construct a series of posts to support an idea that you haven't researched and are thus incapable of offering reasoned opinions. Suggesting that you 'laid a trap' to get other people to recognize what you opine as being "useless" is insufferable. Insinuating that a person's claimed inability to understand how a stat is calculated is insulting. Try to do better, I'm sure you can.

 

[zeke]

If you can’t even understand how a stat is calculated, that’s probably a good hint.

no it's not. quite the opposite.

point shares aren't an awful stat. they do a decent job of adjusting goals against for a player's ice time and team quality. much better than simple plus minus stats.

 

[Dreakmur]

no it's not. quite the opposite.

point shares aren't an awful stat. they do a decent job of adjusting goals against for a player's ice time and team quality. much better than simple plus minus stats.

Plus minus is a big part of these calculations. Then add some ice time and look at some team averages. It does not measure defensive ability. At all.

 

[Dreakmur]

That you're talking out of your a**? Or that you can't complete a full sentence? Or that you're trying to construct a series of posts to support an idea that you haven't researched and are thus incapable of offering reasoned opinions. Suggesting that you 'laid a trap' to get other people to recognize what you opine as being "useless" is insufferable. Insinuating that a person's claimed inability to understand how a stat is calculated is insulting. Try to do better, I'm sure you can.

Look at the calculations. You can’t honestly say that you really think it measure defensive ability....

 

[BoredBrandonPridham]

What % of his starts are in the offensive zone? Feels like a lot. Giordano goes up against the other team's best forwards and is still putting up great offensive numbers, I'd say he has the edge at this point

Feels like a lot huh? You just get the feels on it to bring it up? ZS% is identical between the 2. Do you have "the feels" that Rielly doesn't go up against top forwards too?

 

[BoredBrandonPridham]

Point share comparisons on hockey reference puts Rielly right up there with the great D-men of our generation such as Olli Maatta

 

[BoredBrandonPridham]

no it's not. quite the opposite.

point shares aren't an awful stat. they do a decent job of adjusting goals against for a player's ice time and team quality. much better than simple plus minus stats.

Better than plus-minus, sure. But it is yet another stat being used to line up players against one another without taking into consideration qot, qoc, zs, pp/pk/5v5 or score effects. Probably does more harm than good for the creators of the stat to start using it to compare players with each other such as what hockey-reference does on their player pages.

I actually can't think of a question I could ask that this stat would effectively answer -- can you? Some non-contextual stats can answer a specific stat for either a team or a player, so long as you can bucket the player into a role and say, for example, "does this player excel in this area at his role?". Some on ice stats can answer questions like these, but point shares would be penalizing a player fulfilling a role on the ice if the players on their team sitting on the bench while they're on the ice are superstars.

What kind of useful questions does this stat really answer?

 

[Joey Hoser]

If you can’t even understand how a stat is calculated, that’s probably a good hint.

So if you're too dumb to understand something, that means that thing is useless?



Yeah sure bud. Smart people are so stupid.

 

[zeke]

Better than plus-minus, sure. But it is yet another stat being used to line up players against one another without taking into consideration qot, qoc, zs, pp/pk/5v5 or score effects. Probably does more harm than good for the creators of the stat to start using it to compare players with each other such as what hockey-reference does on their player pages.

true.

but in Rielly's case, adding in all those factors would only make him look even better.

 

[BoredBrandonPridham]

true.

but in Rielly's case, adding in all those factors would only make him look even better.

Probably, I think so yea. But we've concluded that it's wrong for a player like Rielly. Can we not agree that this stat is pretty much useless then? I've updated my original post that you've quoted (sorry -- thought you might be sleeping ).

 

[zeke]

Probably, I think so yea. But we've concluded that it's wrong for a player like Rielly. Can we not agree that this stat is pretty much useless then? I've updated my original post that you've quoted (sorry -- thought you might be sleeping ).

the issue with the stat is that it uses goals against to measure defense, which can be very misleading due to sample size luck and goalie performance.

but for those who hate the shot-attempt-based analytics, these point shares are at least better than the usual plus/minus that they use.

the truth is Mo has been on the ice for very few goals this year.

 

[BoredBrandonPridham]

the issue with the stat is that it uses goals against to measure defense, which can be very misleading due to sample size luck and goalie performance.

but for those who hate the shot-attempt-based analytics, these point shares are at least better than the usual plus/minus that they use.

the truth is Mo has been on the ice for very few goals this year.

That's the least of this stats problems. Any goal-based stats for both offense and defense suffer from sample size, luck and goalie performance (on your team or against). This stat at least uses career data which puts it in a better spot when it comes to sample size.

The problem is no context of player usage. No specification of how it should be used, aside from just sorting a list of players without any useful insight.

 

[zeke]

That's the least of this stats problems. Any goal-based stats for both offense and defense suffer from sample size, luck and goalie performance (on your team or against). This stat at least uses career data which puts it in a better spot when it comes to sample size.

The problem is no context of player usage. No specification of how it should be used, aside from just sorting a list of players without any useful insight.

but that's a problem with every stat.

this stat at least uses some usage context - I.e. ice time and team quality.

 

[unitedstars87]

He is a legit top pairing dman. He hasn’t had the luxury of playing with another top pairing partner yet has thrived. So I guess that makes you a franchise dman?

 

[BoredBrandonPridham]

but that's a problem with every stat.

this stat at least uses some usage context - I.e. ice time and team quality.

Many stats adjust for the aforementioned factors. This one does not “at least” do anything. What value is overall team quality — including teammates that are on the bench when you are on the ice.

You haven’t offered up the question you would ask that this stat answers yet. If there is no question it could answer, it is useless. What is the question? Refer back to a previous post for the examples of questions that other stats that take that context into consideration can “at least” answer, and which types of questions that this one cannot. You haven’t really addressed any of that.

 

[DoobieDubas]

the issue with the stat is that it uses goals against to measure defense, which can be very misleading due to sample size luck and goalie performance.

but for those who hate the shot-attempt-based analytics, these point shares are at least better than the usual plus/minus that they use.

the truth is Mo has been on the ice for very few goals this year.

is that true? i feel like hes been risky at certain times looking to get a point

 

[zeke]

Many stats adjust for the aforementioned factors. This one does not “at least” do anything. What value is overall team quality — including teammates that are on the bench when you are on the ice.

You haven’t offered up the question you would ask that this stat answers yet. If there is no question it could answer, it is useless. What is the question? Refer back to a previous post for the examples of questions that other stats that take that context into consideration can “at least” answer, and which types of questions that this one cannot. You haven’t really addressed any of that.

it tells us the amount of goals for and against he's been on the ice for adjusted for icetime and team quality.

not a great stat, but better than many that get used more.

 

[BoredBrandonPridham]

it tells us the amount of goals for and against he's been on the ice for adjusted for icetime and team quality.

not a great stat, but better than many that get used more.

That’s not an interesting question. It’s a description of the formula.

Less useless can still be useless.

I think we’ve settled this conversation up. Cheers.

 

[LeafFever]

He is a legit top pairing dman. He hasn’t had the luxury of playing with another top pairing partner yet has thrived. So I guess that makes you a franchise dman?

Legit Top-Paring? I guess that leaves room for you to name him a #2 D then?

Why are people still debating he's not a #1?

 

[ZippityDooDa]

50 point defenseman playing 22 minutes a night is "dead weight", and I'm the delusional one?

Gardiner is 13th in the league amongst defensemen in defensive point shares btw, which is darn good.

You're seriously unrealistic and ridiculous if you think Zaits and gards are dead weight

Lmao Gardiner cheerleaders. Cute.

 

[ZippityDooDa]

Keep proving my points Gards! He's still playing better than Nylander though so that's something I guess...