Stars @ Leafs - Nov. 1st, 7:00pm - TSN4, TSN 1050

[herman]

Shots are probabilistic events. Dallas got more than a little lucky. The Leafs are working and generating without two of their top scorers and play drivers. Relax.

All events are probabilistic.  What do you mean here?

Some people don?t understand puck luck.

It was a serious question.

That was a pretty serious and succinct answer; my post history is riddled with long-ass write ups about this.

Evaluating a team's performance based solely on pucks that go in (or not) is what leads to the 2013-14 Leafs; they had a nice shooting percentage bender that resulted in a playoff berth in a shortened season the year before and that had management convinced the team was better than its true performance (see 2016-17 Ottawa Senators). The underlying numbers showed that they were structurally turd muffins that had some hot shooting.

A good portion of what determines if a puck goes in or not is outside the control of the players. Take Tavares' chances tonight for example. One was on target but apparently hit the crossbar. The margin there is less than the height of the puck. He had several pucks hop over his stick in the process of a scoring rush thanks to SBA ice; does that mean Tavares is trash? Marner whipped a cross-ice pass to Tavares, but it was leaning a bit on one edge and thus hopped off his stick blade, nullifying the window of opportunity to score; what if it was leaning the other way and cradled the curve of his stick as intended and he rifled it short-side shelf?

This is one of the first things Kyle Dubas explained to Shanahan and the MLSE ownership. PDO* is the 'stat' that sort of encapsulates this effect, taking the Save Percentage and Shooting Percentage of the team and adding them together. Over time, hockey history shows that teams will regress to the average, which is a PDO of 1.000 (usually a 0.915 sv% and an 8.5% shooting%). Sometimes a team can go on a run with a hot goalie (Fleury last year), or disgusting shooting (Leafs first 8 games), but it never lasts. So if you're evaluating your team in the long run, yes the score matters on a game by game basis, but coaches and managers need to keep an eye on the structure, chance generation/suppression, and basically underlying reasons why teams either score or get scored on.

* PDO is named for the internet user that first developed the stat and explained its effect. In and of itself the name is meaningless; one of the analytics team dubbed it 'probability percentage-driven outcome' to help the execs understand it.

 

[herman]

https://twitter.com/DomGalamini/status/1058173295311372289

Take a look at where the Leafs were able to shoot from. You don't get to those areas without doing something very right.

Look how often they were doing it. You don't get to take that many chances without doing those right things fairly consistently.

https://twitter.com/kristen_shilton/status/1058177595169157120

Guess what the coach's message is to the team after a night like this? Stick with the process.

 

[herman]

Connor Brown with a shift that Babcock will point to in subsequent practices as something emulate.

https://twitter.com/TheLeafsIMO/status/1058162854270922753

Look at how Brown hustles to get on top of the puck and into the hands of the defenders to take away their ability to play the puck. And then track all the way back to get under the stick of Gemel Smith to break up his chance.

 

[herman]

Just going to pile this on here because Babcock gets puck luck and understands that it's the process that eventually leads to sustainable results:
https://mapleleafshotstove.com/2018/11/02/mike-babcock-post-game-stars-2-vs-leafs-1-i-thought-our-team-was-engaged-i-thought-we-worked-hard-in-the-end-youve-got-to-score/

I thought our power play was dangerous. I thought we had a lot of looks, but we didn?t shoot it in the empty net. When you think about it, we had it on Mitchy?s stick and we had it on John Tavares? stick a number of times with real good looks and it didn?t go in the net.

Now, I don?t know? if I?m picking the guys, that?s who I want on. We had those guys on a 2-on-1 and they lost the puck. I?ll take those guys all day long in those situations. Over a period of time, it?ll go in.

 

[Kin]

I don't know if my hypothesis will be backed up by the lab work I'm doing later but as is I have a hunch that when your team is built on not great defense and goaltending but really top flight group of elite forwards and you're missing two of the four forwards your team will be....

(crunches numbers)

(embarks on 14 month trip to the Galapagos onboard HMS Beagle)

(electrocutes elephant to discredit Nikola Tesla)

...not good.

 

[ZBBM]

herman, thanks for another long-ass answer but I am talking about real probability, not metaphorical hockey probability.  What I mean is that every potential event in the world can be assigned a probability, so saying "shots are probabilistic events" doesn't really tell you anything.

Then, when you write things like

A good portion of what determines if a puck goes in or not is outside the control of the players.

I have to assume, again, that you mean this in some kind of metaphorical. not physical, sense.  Players shooting pucks are 100% in control of what's happening (since we can assume the laws of physics are uniform for all players on the ice at the time of the shot event, given the scale we are talking about).  The reason Tavares clanked it, and Marleau too, is because they and they alone caused the puck to be XXX mm too far to the outside of the post.

I think advanced stats, such as those you helpfully explained again, are just fine.  But it's easy to mix metaphor into reality when describing them, and that's confusing IMO.  The term "puck luck" being a prime example.  There is no such thing.

 

[herman]

herman, thanks for another long-ass answer but I am talking about real probability, not metaphorical hockey probability.  What I mean is that every potential event in the world can be assigned a probability, so saying "shots are probabilistic events" doesn't really tell you anything.

Then, when you write things like

A good portion of what determines if a puck goes in or not is outside the control of the players.

I have to assume, again, that you mean this in some kind of metaphorical. not physical, sense.  Players shooting pucks are 100% in control of what's happening (since we can assume the laws of physics are uniform for all players on the ice at the time of the shot event, given the scale we are talking about).  The reason Tavares clanked it, and Marleau too, is because they and they alone caused the puck to be XXX mm too far to the outside of the post.

I think advanced stats, such as those you helpfully explained again, are just fine.  But it's easy to mix metaphor into reality when describing them, and that's confusing IMO.  The term "puck luck" being a prime example.  There is no such thing.

What sort of precision can you possibly guarantee with a biomechanical movement? How many times can you throw a paper ball into the trash can from a certain distance? How many times can you do it while running laterally? Diving to the side?

I think the confusion lies in the belief that players are 100% in control of what they do with the puck, when there are so many external and internal variables at play in a game of inches at high speeds. How in control of the ice surface are the players? How in control of the skate angle of a defenseman is a shooter? How in control are they of the sticks hooking their sweaters?

It is unreasonable to believe these players, even the elitest of players, are infallibly accurate, let alone deliberately missing as you've suggested. Physics itself has probabilistic events encoded into its very fabric at the quantum level, so yeah it's uniformly going to have variance.

 

[CarltonTheBear]

Zaitsev played a career-low 15:55 last night.

That includes just 3:12, or 4 shifts, in the 3rd while the Leafs were pushing for a goal. Gardiner took shifts with both Rielly and Dermott in that period.

Time to try:

Rielly-Hainsey
Gardiner-Dermott
Marincin-Zaitsev

 

[herman]

Zaitsev played a career-low 15:55 last night.

That includes just 3:12, or 4 shifts, in the 3rd period while the Leafs were pushing for a goal. Gardiner took shifts with both Rielly and Dermott in that period.

Gauthier was also benched for most of the latter half of the game, not because of anything he did wrong (he played well), but because similar to Zaitsev, he was not the right tool for the task at hand. Zaitsev played well too.

 

[CarltonTheBear]

Gauthier was also benched for most of the latter half of the game, not because of anything he did wrong (he played well), but because similar to Zaitsev, he was not the right tool for the task at hand. Zaitsev played well too.

Gauthier's SF% was 66.7%! His HDCF% was 100%!

 

[herman]

herman, thanks for another long-ass answer but I am talking about real probability, not metaphorical hockey probability.  What I mean is that every potential event in the world can be assigned a probability, so saying "shots are probabilistic events" doesn't really tell you anything.

Then, when you write things like

A good portion of what determines if a puck goes in or not is outside the control of the players.

I have to assume, again, that you mean this in some kind of metaphorical. not physical, sense.  Players shooting pucks are 100% in control of what's happening (since we can assume the laws of physics are uniform for all players on the ice at the time of the shot event, given the scale we are talking about).  The reason Tavares clanked it, and Marleau too, is because they and they alone caused the puck to be XXX mm too far to the outside of the post.

I think advanced stats, such as those you helpfully explained again, are just fine.  But it's easy to mix metaphor into reality when describing them, and that's confusing IMO.  The term "puck luck" being a prime example.  There is no such thing.

What sort of precision can you possibly guarantee with a biomechanical movement? How many times can you throw a paper ball into the trash can from a certain distance? How many times can you do it while running laterally? Diving to the side?

I think the confusion lies in the belief that players are 100% in control of what they do with the puck, when there are so many external and internal variables at play in a game of inches at high speeds. How in control of the ice surface are the players? How in control of the skate angle of a defenseman is a shooter? How in control are they of the sticks hooking their sweaters?

It is unreasonable to believe these players, even the elitest of players, are infallibly accurate, let alone deliberately missing as you've suggested. Physics itself has probabilistic events encoded into its very fabric at the quantum level, so yeah it's uniformly going to have variance.

Just stepping back from the physical/biomechanical/metaphorical(??) sense of variance and 'luck', I came across a pretty good explainer video on why hockey results are more luck-driven than most sports on a statistical level from an unusually reasonable r/leafs.

 

[CarltonTheBear]

Shot charts for last night's game, via @IneffectiveMath:

I will say, that's a lot of missed shots from the slot. But this was still a great game by the Leafs.

 

[ZBBM]

herman, thanks for another long-ass answer but I am talking about real probability, not metaphorical hockey probability.  What I mean is that every potential event in the world can be assigned a probability, so saying "shots are probabilistic events" doesn't really tell you anything.

Then, when you write things like

A good portion of what determines if a puck goes in or not is outside the control of the players.

I have to assume, again, that you mean this in some kind of metaphorical. not physical, sense.  Players shooting pucks are 100% in control of what's happening (since we can assume the laws of physics are uniform for all players on the ice at the time of the shot event, given the scale we are talking about).  The reason Tavares clanked it, and Marleau too, is because they and they alone caused the puck to be XXX mm too far to the outside of the post.

I think advanced stats, such as those you helpfully explained again, are just fine.  But it's easy to mix metaphor into reality when describing them, and that's confusing IMO.  The term "puck luck" being a prime example.  There is no such thing.

What sort of precision can you possibly guarantee with a biomechanical movement? How many times can you throw a paper ball into the trash can from a certain distance? How many times can you do it while running laterally? Diving to the side?

I think the confusion lies in the belief that players are 100% in control of what they do with the puck, when there are so many external and internal variables at play in a game of inches at high speeds. How in control of the ice surface are the players? How in control of the skate angle of a defenseman is a shooter? How in control are they of the sticks hooking their sweaters?

It is unreasonable to believe these players, even the elitest of players, are infallibly accurate, let alone deliberately missing as you've suggested. Physics itself has probabilistic events encoded into its very fabric at the quantum level, so yeah it's uniformly going to have variance.

Just stepping back from the physical/biomechanical/metaphorical(??) sense of variance and 'luck', I came across a pretty good explainer video on why hockey results are more luck-driven than most sports on a statistical level from an unusually reasonable r/leafs.

I'm pretty sure we're talking past each other here so not much point in continuing, but I would have hoped my saying that players are 100% in control of their actions is not understood to mean they "are infallibly accurate."

I guess my one-line argument is that talking about "luck" in any form just obscures things.  There is no such thing as luck, only shifting probabilities.

 

[herman]

Gauthier was also benched for most of the latter half of the game, not because of anything he did wrong (he played well), but because similar to Zaitsev, he was not the right tool for the task at hand. Zaitsev played well too.

Gauthier's SF% was 66.7%! His HDCF% was 100%!

It's baffling based on how he looks when he plays. But he's the Marincin/McElhinney of the centres. Awkwardly effective in limited minutes.

 

[CarltonTheBear]

It's baffling based on how he looks when he plays. But he's the Marincin/McElhinney of the centres. Awkwardly effective in limited minutes.

Oh, you missed the sarcasm there based off my last discussion about him. The raw numbers for those are 2-1, and 1-0 respectively. The furthest my praise for Gauthier can go is to say that nothing happens on the ice when he's out there. Which is better than getting caved in sure but still not exactly something to be proud of.

 

[herman]

I'm pretty sure we're talking past each other here so not much point in continuing, but I would have hoped my saying that players are 100% in control of their actions is not understood to mean they "are infallibly accurate."

I guess my one-line argument is that talking about "luck" in any form just obscures things.  There is no such thing as luck, only shifting probabilities.

There's no Leafs game today, so I have lots of time to digest what you're trying to explain. To me, luck, skill, and probability are all part of the same mesh. Skill is the individual's ability to shift probabilities in his or her favour. Luck is the rest that the individual cannot control.

Auston Matthews, for example, has the ability to add pre-shot movement to his shot (on top of any movement that came prior) that shifts the chances of scoring closer to success than most players. That's a skill and not luck. Once it leaves his stick, the puck is still subject to the bodies in between the puck and the net, and those individuals' respective decisions.

I don't think we're talking about different things, but I also don't know what you're talking about yet and I hope you have the time to enlighten me.

 

[herman]

It's baffling based on how he looks when he plays. But he's the Marincin/McElhinney of the centres. Awkwardly effective in limited minutes.

Oh, you missed the sarcasm there based off my last discussion about him. The raw numbers for those are 2-1, and 1-0 respectively. The furthest my praise for Gauthier can go is to say that nothing happens on the ice when he's out there. Which is better than getting caved in sure but still not exactly something to be proud of.

Nah, I got the sarcasm. But I'm still baffled by how he is fairly consistently (not)affecting the game and I think that's a pretty difficult thing to accomplish. The raw numbers are small and thus super noisy/spiky in the small sample size sense, but there are things that Gauthier is doing that aren't publicly statted (yet) that is consistently producing this type of result and I'm curious about that and how to measure it going forward.

 

[OldTimeHockey]

I actually really like Gauthier out there. I am less nervous with that fourth line on the ice than I was when Lindholm was playing c on the 4th

 

[bustaheims]

Was at the game last night, and I really have to echo the sentiments of the Leafs being the better team. Even my dad, who really doesn't have any team allegiances remarked that he felt the score went against the way the run of play on the ice.

Obviously, there were some execution issues, but, on top of that, the ice looked bad. The puck was bouncing and rolling, guys were falling down all over the place. A smoother ice surface, and a good number of flubbed plays become much more dangerous scoring chances, and possibly even goals.

Also, man, do Hainsey and Zaitsev both read the play poorly. A few times, they each had obvious, clear, direct outlet pass opportunities right in front of them, and the opted for the lower percentage play.

 

[Highlander]

Was at the game last night, and I really have to echo the sentiments of the Leafs being the better team. Even my dad, who really doesn't have any team allegiances remarked that he felt the score went against the way the run of play on the ice.

Obviously, there were some execution issues, but, on top of that, the ice looked bad. The puck was bouncing and rolling, guys were falling down all over the place. A smoother ice surface, and a good number of flubbed plays become much more dangerous scoring chances, and possibly even goals.

Also, man, do Hainsey and Zaitsev both read the play poorly. A few times, they each had obvious, clear, direct outlet pass opportunities right in front of them, and the opted for the lower percentage play.

I still can't believe why the Air Canada Centre (sorry will not call it Scotiabank for a while), doesn't have great ice.  It is not the best ice in the league, that is for certain.
Also wanted to say I thought Johnsson had his best game by far and lets hope he steps ups and does what we all believe he can do now.  Also the Goat is doing well in limited minutes, I really hope he pops in a few soon.