Player Discussion Jack Campbell

[5 Mins 4 Ftg]

If JC takes a leave we should get that cap space back I would think but I am positive Bettman will decline the request because reasons.

 

[Aerchon]

Goal # 3 here today. Unsavable and Nurse not where he should be, no pressure, no gap control.

Goalies not the problem. We'll. Not the main problem. Lol. Clarify.

Nurse benched at the end of that game?

 

[OilfaninTO]

.

 

[bucks_oil]

In fact, I see it as exactly the opposite of what you are asking. It's the noise that matters. Ideally you really need to do is to count the errors on every play and then determine the relative rate of goals, not just on the ones that go in.

Part of the problem I have with what you are saying is that it almost seems to support the premise that every goalie is relatively equal and that the core distinguishing characteristic between saves or no save is simply the nature of the shot. I know you don't really think this but read on the surface this could be taken as an extreme corollary of your approach. Is it really true that very two on one would have the same result regardless of who was in net? Similarly with every shot from the slot. In fact, I see it as exactly the opposite of what you are asking. The noise matters or rather in this case the non-goals which fade from memory much faster than goals and evade perception much more readily. Ideally you really need to do is to count the errors on every play and then determine the relative rate of goals.

Looking at individual plays in isolation how can both be revelatory and also misleading. Baseball is not a perfect analog of hockey for sure but there are some at least plausible parallels here. Two batters could face a guy with a nasty fastball at 98MPH with movement. Both guys could swing and miss and it would be easy to say that this is to be expected. The way you distinguish who is the best hitter is not to look at outcomes after 500 or 1000 or pick your poison number of at bats. In those 500 at bats the pitches the two hitters face may be quite different. But in the end if one guy gets hits 5% more more often it is generally expected that that player is the better hitter unless there is compelling evidence otherwise.

Two goalies could face the same shot or the same 2 on one and the outcome could go either way regardless of who was in net. But if you could repeat the same event hundreds of times one would expect the better goalie to come out ahead. Unfortunately, shots are not repeatable experiments with controlled conditions. So you have to introduce some smoothing. That is what HDSA are, smoothing factors. The algorithms for xG's take this to a much higher level. Again this is far from perfect, but I contend that over a long enough span and even with some context added they can be quite useful.

I'll repeat, this is far from an exact science. Randomness is part of hockey, much more so than in baseball. So it is through many events that one has a chance to filter out the noise. And of course, minimizing errors is absolutely important. But so is reducing the number of attempts that have a historically high probability of resulting in a goal. These two things are not impendent of course.

Thanks for the response... I'll spend a bit more time with it and reply more fully, but big broad strokes here:

To the 2nd paragraph: no I'm not saying that at all and I don't think my system would show that all goalies are equal either.

What I am saying though is:
1) that errors matter way more than reflexes and what you'll find is that if you count up goals, there are going to be goalies who make a whole lot of their own errors and there are going to be goalies who are much closer to perfect. Isn't that what we are interested in? Which goalies minimize errors and tend not to contribute to their own goals against?
2) the physical side, while I will still argue matters less when comparing goalies to each other... there are limits of course to human reaction time... that would still get counted/rewarded in my error model. To your point, if a goalie makes an unexpected save (on a shot from the slot or 2:1 save where we would not normally charge an error) then he'd get a + in his error account.

Putting the two together, you'd be able to accurately judge the better goalies in the league as being guys who are always in position, rarely make errors and make saves that are unexpected... you may end up with a Vezina-quality goalie being +25 on the year or something like that meaning he saved more sure goals than he caused.

And yes, you are right... we've have to at least score all high danger shots (plus any goals from non-high danger places) rather than just goals. I intended that, I just wasn't clear earlier.

 

[Soundwave]

Goal # 3 here today. Unsavable and Nurse not where he should be, no pressure, no gap control.

Goalies not the problem. We'll. Not the main problem. Lol. Clarify.

Nurse benched at the end of that game?

Skinner got over on that goal about as fast as the Titanic trying to avoid ice bergs though.

Neither Campbell or Skinner are legit starters and this is a team that makes goalies of that tier look real bad (see also: Mikko Koskinen).