VIVA LAS VEGAS

by Lowetide

The Golden Knights begin their inaugural season tomorrow night and the club is still out of sync. Capfriendly has the team’s alignment as 11 forwards, nine defenders and two goalies. I imagine Brandon Pirri will be forward No. 12, but the nine defenders need to be culled.

Edmonton is playing Eric Gryba as No. 6 and I’m not sure Yohann Auvitu has earned the trust of the coaching staff based on performance. Is there an opportunity to make a deal?

  • Bob McKenzie: No guarantee it’s imminent by any means, but keep an eye on VGK as expansion team is in trade mode. May move a goalie (Pickard?) and a D. VGK expected to hold onto vet D Engelland, Garrison, Sbisa and won’t part with kids Theodore, Schmidt and C Miller but any other D possible. Source

That leaves Clayton Stoner (who is on IR), Brayden McNabb, Jon Merrill, Griffin Reinhart and Brad Hunt. Hmmm. I don’t know that there’s a lot there for Edmonton, but let’s have a last at the entire group to identify attractive options.

GOLDEN KNIGHTS BLUE (alpha)

  1. LD Brayden McNabb. $1.7 MILLION, UFA summer 2018. Missed a bunch of time last season with the Kings due to an arm injury. McNabb would be a fine addition (his possession numbers are excellent) and he has had success in the NHL (he is 26).
  2. LD Jon Merrill. $1.1375 MILLION, RFA SUMMER 2018. He’s a big defender with some two-way ability. Fancy stats don’t boast much but do suggest he could play successfully on a depth pair.
  3. RD Colin Miller. $1.0 MILLION, RFA summer 2018. He’s pretty close to ideal, young enough to grow with the group and experienced enough to be a plug and play option. Good two-way player, Puck IQ had him playing 45 percent of his season against mid-level competition and delivering good results. The Oilers have some extra forwards hanging around, Miller is under control moving forward. Would you trade Jujhar Khaira for him?
  4. LD Griffin Reinhart. $800,000, RFA summer 2019. Reinhart has draft pedigree and the organization valued him before the expansion draft. No reason to think they value him less now.
  5. LD Nate Schmidt. $2.25 MILLION, UFA summer 2019. This is a very interesting player, under contract and with successful possession numbers a year ago. He played mostly mid-level opposition (source: Puck IQ) and sawed off the competition. He is a good puck mover and can play either side according to The Hockey News.

Who would you choose?

119 comments
0

You may also like

0 0 vote
Article Rating
119 Comments
Newest
Oldest Most Voted
Inline Feedbacks
View all comments
Pescador

VEGAS WINS BABY!
Hahahaha

Dominoiler

kgo,

Oh dang, that post was it’s own mess of hot garbage, i must have been really up on marincin.. LT, you do too good of a sell job on prospects, i was obviously stoned on marincin pipe dreams.. as a guest on your show a couple days ago pointed out, marincin -> gryba was a meh trade and never of the calibre that warranted my hyperbolistic rant.. shamed hard, in the end, seems a fitting outcome..

VOR

Georges,

It was always fun arguing with you. I hope you will be back!

Woodguy v2.0

Georges,

– I went and checked on the individual games I had from last season. At a team level, there’s no relationship. Winning or losing on Fenwick doesn’t tell you much about winning or losing on goals.

Yes.

My contention is that the score doesn’t always tell the story of the game.

Let’s look at the score of 3-0.

Game 1

Edmonton 3 – Calgary 0

EDM FF% 60.4% vs CGY FF% 39.6%

The score and FF% tells a similar story on the face of it.

Edmonton owned the game both via score and FF%

How about Jan 29/2014?

Edmonton 3 San Jose 0

EDM FF% 24.1% vs SJS FF% 75.9%

Not exactly what you’d expect eh?

I know this is an extreme example, but I’m trying to show you that that score doesn’t tell the story of the game nearly as well as the FF% does.

If you remember that game there is no way on Gord’s Green Earth you would agree that the score told the story of the game.

If you watched you knew that Edmonton was beaten like a rented mule, but Scrivens wouldn’t break while Niemi couldn’t stop a beachball. (somethings never change…….)

You’re right that there is no relationship between winning and losing at the game level with FF%.

We were talking about describing the game, not predicting it.


– I think there is a relationship at a player level, however. I did the work quickly so I could’ve made a mistake. But it makes sense to me…
– Anyway, life stuff calls so I’m leaving for an indefinite long weekend.
– It’s been fun chatting with you during my time here, WG. Best of luck with PuckIQ. I hope it rocks for you guys!

Always fun to talk to you too Georges.

One day you’ll accept how much luck there is in hockey and we’ll sing and dance and drink beer.

Georges

Woodguy v2.0,

– I went and checked on the individual games I had from last season. At a team level, there’s no relationship. Winning or losing on Fenwick doesn’t tell you much about winning or losing on goals.

– I think there is a relationship at a player level, however. I did the work quickly so I could’ve made a mistake. But it makes sense to me…

– Anyway, life stuff calls so I’m leaving for an indefinite long weekend.

– It’s been fun chatting with you during my time here, WG. Best of luck with PuckIQ. I hope it rocks for you guys!

Woodguy v2.0

Georges,

– Flotsam doesn’t sink. It floats.

It does when you tie a brick around it.


– I asked LT to replace a redundant column with a useful one. His call, obviously.

Ok.


– 5v5 goals are a pretty good description of what happened. Games are mostly decided at 5v5.

I think 5v5 goals are a good description of goals, but not of the play overall. This is pretty well established.


– Last season, players with positive (negative) 5v5 goal differentials in games tended to have positive (negative) 5v5 Fenwick differentials. They’re correlated.

Yes they are, but Fenwick will give us way more information.

– This is what I meant by setting context. Winning Fenwick generally goes hand in hand with winning goals. When you win both, you’re winning the right way… outscoring the opponent while controlling the play.

Not in individual games. Over a season yes, in a game, no.


– The LAK had positive 5v5 Fenwick differential even when they had negative goal differential. If you’re right and Fenwick better describes what happened than goals, a Kings fan who read a season’s worth of Fenwick recaps should’ve had no complaints.

It would also lead them to ask questions like “why is our Fenwick dominance not turning into goal dominance?” LAK changed coaches to change that equation. All the verbal around Sutter —> Stevens is about “scoring more”


– If the criteria for more descriptive is more events, we should prefer Corsi; we should also consider including faceoffs.

I don’t like to limit data at all. The more the merrier.


– If the criteria is more information = more descriptive, at a season level, 5v5 goal differential provides the most information on overall goal differential, which provides the most information on points in the standings.

At the end of the season yes. In an individual game, no.

– LT includes goals, assists, and points which relate to the GF part of goals; I take it you want him to remove those; I want him to add GF and GA.

At no point did I or would I ever ask LT to remove information. This is a very poor assumption. Its like you’ve never conversed with me here! 🙂


– How many sites can I check DFF on? If it’s just one, how do I know that the number I’m seeing is correct? Can I calculate DFF myself?

Only one. You can match it up to the other public xGF and see what you get. I think you downloaded G’s data today so you’ll know soon.


– I’m thrilled with everything that you write. I started posting because I saw G talking about math and hockey and I wanted to join in. He left. Now you’re the only poster on here who regularly uses numbers (other than ricki). I missed the window where there were more mathish people involved in these conversations. My loss.

I miss it too. I learned a ton and I’m not qualified when the conversation gets very mathy.

G left for the same reason I don’t post as much. The portion of the comment section has turned hostile about counting stuff in a hockey game. It used to be the one place you didn’t have people telling you to watch the game. Things change.

I still post because there are still lots of great people here, but the conversations sprouting out of the posts (like the genesis of WoodMoney) are very few and far between.


– Twitter is gross.

It has a mute and block buttons so you control your feed. You can make it non-gross. I chat with G, Wheatnoil, FRJohn, CodexRex, Tyler Dellow, Jon Willis, Kent Wilson, Hbomb, Showerhead, Pat McLean and many others all the time.

It can be really really good.

Georges

Woodguy v2.0,

– Flotsam doesn’t sink. It floats.

– I asked LT to replace a redundant column with a useful one. His call, obviously.

– 5v5 goals are a pretty good description of what happened. Games are mostly decided at 5v5.

– Last season, players with positive (negative) 5v5 goal differentials in games tended to have positive (negative) 5v5 Fenwick differentials. They’re correlated.

– This is what I meant by setting context. Winning Fenwick generally goes hand in hand with winning goals. When you win both, you’re winning the right way… outscoring the opponent while controlling the play.

– The LAK had positive 5v5 Fenwick differential even when they had negative goal differential. If you’re right and Fenwick better describes what happened than goals, a Kings fan who read a season’s worth of Fenwick recaps should’ve had no complaints.

– If the criteria for more descriptive is more events, we should prefer Corsi; we should also consider including faceoffs.

– If the criteria is more information = more descriptive, at a season level, 5v5 goal differential provides the most information on overall goal differential, which provides the most information on points in the standings.

– LT includes goals, assists, and points which relate to the GF part of goals; I take it you want him to remove those; I want him to add GF and GA.

– How many sites can I check DFF on? If it’s just one, how do I know that the number I’m seeing is correct? Can I calculate DFF myself?

– I’m thrilled with everything that you write. I started posting because I saw G talking about math and hockey and I wanted to join in. He left. Now you’re the only poster on here who regularly uses numbers (other than ricki). I missed the window where there were more mathish people involved in these conversations. My loss.

– Twitter is gross.

HT Joe

rickithebear: She says i have been a hammer when presenting ideas.
That my apraoch guarantees them to be defensive.
But she tells me I am ethically bound to take the position i have.

Ricki: Don’t ever change… you’re an absolute gem on this website! Thank you for your regular posts over the years.

Woodguy v2.0

Georges,

It’s like we’ve never met before. You keep trying to educate me on things I think you think I know nothing about. It’s fun.

Well, when you write this:

The FF% column in the player table is redundant once you’ve listed 5v5 FF-FA. You should consider replacing it with 5v5 GF-GA.
5v5 GF-GA is really, really important. It’s what happened. Here’s where winning happened, here’s where losing happened. Without that context, Fenwick is just Fenwick. It’s not winning and losing. But when you include GF-GA, you can get into did a player win or lose in a way we can be happy with, i.e., how did possession look?

My take away from it is that you think goals described what happened in a hockey game the best.

I wrote why that’s not true.

You wrote that 3 events describe what happened in a hockey game better than ~70 events. I disagree with that.


“Small samples and volatile” aren’t relevant to describing what happened. Describing is different from inferring or predicting. Goals are what happened. Fenwick is a bunch of things that happened that didn’t lead to goals and a few that did.
You say that Fenwick describes the game better than Corsi. Why? How did you arrive at this?

Corsi describes every shot attempt.

Fenwick desribes every shot attempt that wasn’t’ blocked.

It describes who got the puck to the net more and therefore describes what actually happened better.


And then why is DFF even better at describing than Fenwick?

DFF further describes each unblocked shot by assigning a “danger value” to each shot based on location and type of shot.

More information = more descriptive.

DFF is pretty obscure and has a trust me, I know what I’m doing feel to it. It’s not simple. It’s not in the open. To my understanding, it (along with Manny’s xGF) can’t be recreated. So it can’t be tested. You’re an advocate but I’m not sure you yourself know or understand all the details of the calculations G is using. How do you know he did them right? (If you think the question is offside or impertinent, you’ve never built a calculating/reporting system before.) Why not shots? Everyone understands shots. Nothing to calculate there. I can check shots on a bunch of sites to see if a number is right. How many sites can I check DFF on?

The danger value was figured out via shot location, timing, and type data vis a vi SH% from 2007 on .

G’s jumping off point was Mike Parkatti’s work on one of the first “expected goals” metrics that Mike worked on in 2013: http://www.boysonthebus.com/2013/11/12/piecing-the-shot-puzzle-together/

G’s method was very different from Manny’s so when it turned out that DFF had a correlation to Manny’s xGF of .954 it was a good validation of both metrics. Two different methods of danger weighting shots essentially came to the same conclusion.

G’s work is here:https://oilersnerdalert.wordpress.com/2015/10/30/explaining-dangerous-fenwick/

and here: https://oilersnerdalert.wordpress.com/2016/07/17/a-brief-statistical-look-at-dangerous-fenwick/

Manny’s work is here: http://www.corsica.hockey/blog/2016/03/03/shot-quality-and-expected-goals-part-i/

and here: http://www.corsica.hockey/blog/2016/08/13/shot-quality-and-expected-goals-part-1-5/

As for “can’t be re-created and tested” I don’t know what you mean.

Did someone try to re-create it and fail? Link?


Not sure why you think I asked LT to only include goals. Maybe read my post that suggested replacing the FF% column in his table. Having FF-FA and GF-GA in the same table

When I read “replace FF with GF” it told me you wanted FF removed from the table.

If I erred, that’s on me.


By the way, what’s the relationship between winning on 5v5 Fenwick and winning on 5v5 GF%? Do you think the question is important? If so, how would you go about trying to suss out an answer?
Getting back to DFF… I’d love for G to drop by. I’ve looked at his methodology and I remember having lots of questions. As an example, what data does G use to calculate the probabilities for different types of shots? Does he then use that same data to calculate DFF? Some of my first posts on here were on SF% and PDO as two independent factors that correlate with GF%. I mentioned that any shot-based measure that has a higher correlation with GF% would have to take some information from PDO. I can go back and look at G’s methodology post again. I know this kind of thing takes a lot of thought and work.

Now you’re getting onto a different topic, that of prediction which should have its own thread.

You’re probably talking about how the relationship between points and shot metrics has waned in the last few seasons.

The answer to “why?” that most come up with, and which makes the most sense imo, is that the standard deviation of the shot metrics across teams has shrunk significantly over the last 10 years.

It used to be there were 5 or so “elite” teams and 7-8 “bad” teams. Up to half the league very good or very bad in regards to the shot metrics. It was easier for the cream to rise to the top and the flotsam to sink.

Now the standard deviation has shrunk and consequently their predictive value has shrunk as well.

Instead of getting lots of matchups where teams’ CF% are ~5% apart, you have many more where the spread is under 3%.

It actually introduces more luck into the outcomes of each game. I know you’re thrilled that I wrote that, but its true.

I also think it significantly increases the value of above average goalies.

Also,

If you want to ask G some questions that I cannot answer (I did not and could not do the work on DFF) its best to ask him on twitter. He’ll generally post there.

OriginalPouzar

I will be interested to see the line rushes at practice today to get a sense if Yamamoto is going to play again tomorrow night or if Pak will draw in.

Georges

VOR,

1. What do you suppose PC and TMac were talking about in their first season in Edmonton, when they weren’t on their way to 103 points? Mostly process, process, process – very little about outcomes? It seems so straightforward to me. It’s a competitive league. Every coach is preaching process. How do they know they’re making progress? What motivates them to try different things?… Anyway, I believe you’ve said you’ve coached, so maybe this is personal for you. No problem. I’ll let it go.

2. When I say luck evens out, I’m not invoking the gambler’s fallacy, i.e., believing a string of bad luck will be followed by a string of good luck. I mean the effect of a string of good/bad luck diminishes as you make more observations (i.e., increase the size of the sample) of a random variable. The mean of the sample (the typical performance) approaches the expected value of the random variable as sample size increases. That is a simple statistical reality that I’m aware of.

As for the statistical reality you’re referring to above, I believe you referenced a master’s thesis on this topic. That thesis applied the work of Tom Tango to hockey. Here’s a link describing Tango’s method and offering a sort of proof for why it’s valid.

http://blog.philbirnbaum.com/2011/08/tango-method-of-regression-to-mean-kind.html

It’s an interesting read and pretty straightforward to follow. The comments section is fun; Sunny Mehta makes some excellent points. I”m not sure you’re motivated to discuss it, but that simple statistical reality you cite isn’t quite as real as you may believe it to be.

if you meant something else, then I guess I need elaboration. Your description of noise in the signal from the first scrimmage to the winning of the Cup is poetic. I’m not sure I understand the math involved. I pointed out to you before that your argument for parity and the decisive role of luck didn’t jive with the reality that 4 teams have won the past 9 Cups. Not sure if you worked out the odds of that happening under the assumption of absolute parity. They’re not good.

I guess what I really wish is that you would sometimes use statistics or math in your arguments, instead of just arguments and anecdotes in your arguments. I think it would make us both feel a little better. You would be vindicated. And I would be convinced.

3. I have a habit of doing the math myself. Because I often don’t understand it. And that bothers me. The PDO stuff is very basic, not sure what you’ve gleaned from my posts. Happy to recap. I do have trouble excepting that luck is luck. But I accept it just fine.

This, tho…

“Outcomes follow inputs (innate ability, cohesive team play”) to quote you. The thing is that statement isn’t true about hockey. In hockey your proposed correlations are severly eroded by luck. This makes it particularly important in hockey to coach process.

… just hurts my head. You’re saying the same thing I’m saying, except you’re holding process up as a fetish: luck severely erodes the connection between outcomes and inputs, boys, so let’s stick to process and pray to luck. I’m saying it’s more like: we can’t control luck but we can control what we do and if we control what we do and we’re good enough as individuals and as a team, we can expect to see good results… but, hey, no harm in praying to luck.

Gayfish

I adore you Ricki. I will follow you blindly. Shove whatever you want down my throat.

GMB3

kgo: If you reread the comments from the day Marincin got traded, (https://lowetide.ca/2015/06/27/oilers-trade-martin-marincin-for-no-107-and-brad-ross/)

A dozen of our brethren made unambiguously negative statements about trading Marincin…my favourite was by DOMINOILER:

“I like the phrase ‘hot garbage’.. Oilers deal Martin Marincin for steaming pile of hot garbage.. It has a certain quality that rings in your ears, while boiling your blood.. How the F! do you divorce yourself from this shit show of a sports team right as they draft a generational talent.. FFS.. Locked in the ever cycling ‘because oilers’ hampster wheel..”

So the eye test guys have never been wrong? Are you playing dumb?

jm363561

Keep doing what you are doing LT. In an ideal world you might provide:

1. Sophia Loren imagery.
2. Free form comments highlighting key stats, which will vary from game to game (which you provide).
3. Cult of Hockey’s contributions to chances for / against. These give me the most useful information for an individual game (although most posters seem to condescendingly dismiss these).

Most stats do not give me too much on a game by game basis. The cumulative stats for a reasonable sample are more interesting. (Goal difference is much better than Corsi.If I knew where to get GF% I would look at it, but +/- is okay as a substitute.)

Georges

v4ance:
Georges,

The issue is that luck, while it evens out in the long run, doesn’t always even out during one full 82 game season.It actually can still be too small of a sample size.

The 2012-13 Leafs and the 2013-14 Avsare both examples of this.Despite being badly outshot throughout those seasons, Randy Carlyle and Patrick Roy were heralded as “great” coaches by many in the mainstream media for their teams’ results.Analytics concluded that the teams were being badly out possessed based on corsi numbers and that an unsustainable amount of luck was involved.

The flipside argument is the 2011-12 LA Kings where their underlying possession numbers were stellar but their luck and their results were sub-par in comparison.They appeared to sneak into the playoffs, but once there, showed their true capability.

The tactics and strategies employed by those coaches in the subsequent years led to results commensurate with their possession numbers.

I’m not sure we’re on the same topic.

But on the topic that you’re on, the 2011-12 Kings went from a dominant possession team to a meh possession team in the playoffs. They won with spectacular goaltending from Quick and a super high PDO. That completely reversed their horrible shooting % and super low PDO during the regular season. So luck did even out in their case. But that team isn’t an unqualified success for possession-leaning analytics types.

As for results being commensurate to possession numbers, now that corsica is back up you can check for yourself how that proposition has held up recently.

LoDog

v4ance,

Regular old stats concluded that as well.

VOR

Hi Georges,

At the elite levels the only difference between pro athletes and amateur athletes is the amount of money they have in their ass pocket. Ego is ego, drive is drive, professionalism isn’t about a paycheck, it is about a world view. In that world view you control every thing you can and let every thing else go. If something bad happens it is somebody else’s fault, if something good happens it is your doing. The calculus is identical.

Using hockey as the example, every GM knows that it takes time before you see a long term impact from a coaching change. So they are looking for process changes to tell them progress is being made (do the players feel safer, are less cheap shots being taken, do they win more puck battles). Don’t believe me, go back and read Peter Chiarelli’s comments from last year as the season went along. It is mostly process, process, process – very little about outcomes.

I don’t mean to be difficult but there is no evidence of luck evening out in the sense you mean. Luck creates noise in the signal in the first minute of the first scrimmage of rookie camp and that continues until the final minute of the final game of the Stanley Cup. The plus and minus values of those effects on a single player or a single team over a single season do not have to, and frequently do not combine to equal zero. This is true in all sports to some extent but luck determines more of the outcome of hockey games, seasons, and careers than in the case in most professional sports. That is a simple statistical reality.

Based on everything you post here you seem to have trouble excepting that luck is luck (for example your arguments about PDO’s predictive ability. You continue to display this world view in your response to me. “Outcomes follow inputs (innate ability, cohesive team play”) to quote you. The thing is that statement isn’t true about hockey. In hockey your proposed correlations are severly eroded by luck. This makes it particularly important in hockey to coach process. It also explains why professional hockey players are the most superstitious atheletes in sports.

kgo

rickithebear,

Ethical…you keep using that word….I do not think it means what you think in means.

v4ance

Georges,

The issue is that luck, while it evens out in the long run, doesn’t always even out during one full 82 game season. It actually can still be too small of a sample size.

The 2012-13 Leafs and the 2013-14 Avs are both examples of this. Despite being badly outshot throughout those seasons, Randy Carlyle and Patrick Roy were heralded as “great” coaches by many in the mainstream media for their teams’ results. Analytics concluded that the teams were being badly out possessed based on corsi numbers and that an unsustainable amount of luck was involved.

The flipside argument is the 2011-12 LA Kings where their underlying possession numbers were stellar but their luck and their results were sub-par in comparison. They appeared to sneak into the playoffs, but once there, showed their true capability.

The tactics and strategies employed by those coaches in the subsequent years led to results commensurate with their possession numbers.

Jay Arrr

Lowetide: Derek Zona hates me, too. So we have that in common. I didn’t mention Reinhart and I’ll thank you to be respectful.

So nicely handled. Lowetide, you are truly a gem.

dustrock

rickithebear: My wife was raised by parents and family from ethical careers.
Mom was a seat of flesh eating and infecious dicease WHO,
cousin is (FIS)!technical director for NA and oversees womens world cup races.
Wife director of racing and sailing program in baddeck.
Editor in post Media. You know a persons charecter when they maintain ethical standards when recieving death threats against here and her family ( My children).

She went thru much of our discourse.
Gave me a list of individuals on he she would fire on ethical grounds.

Quite funny actually.
She is the driving force around many of the letters she helped send to media sources and academia.

She says i have been a hammer when presenting ideas.
Thatmy apraoch guarantees them to be defensive.
But she tells me I am ethically bound to take the position i have.

You’re arguing an ethical position on a hockey blog comments board? I’m pretty curious to see the list of people your wife would fire for being “unethical”.

Georges

VOR:
Georges,

As a coach you have nothing but process. You can’t control outcomes and neither can your athletes. You can only control inputs. More to the point if you start trying to control outcomes almost invariably you screw up the quality or quantity of your inputs.

“The archer ceases to be conscious of himself as the one who is engaged in hitting the bull’s-eye which confronts him. This state of unconscious is realized only when, completely empty and rid of the self, he becomes one with the perfecting of his technical skill, though there is in it something of a quite different order which cannot be attained by any progressive study of the art.” Eugen Herrigel.

Outcomes in hockey have a lot to do, as I have explained previously, with luck. This creates new problems (beyond how destructive goal orientation is to performance excellence) for any coach who tries to tie outcome to process as a feedback loop. The player messes up but the team scores by fluke- you don’t want the play duplicating that behaviour because over time bad practices will become bad outcome. The player does everything right but his team gives up a goal because of bad luck – you don’t want them to give up doing everything right because over time good practices will produce good outcomes.

So why would we as fans want to be obsessed with outcomes when what the ocach and athletes can control is inputs and inputs are the only fair way of evaluating their play or their coaching?

Hi VOR.

First off, I’m not talking about lower levels where athletes are trained, only about the pros. I think the calculus is different in the pros. Also, I agree with you on the first example and any capable coach would take the player aside and do some coaching.

As for the second example, luck evens out. There’s we did everything right and the puck went in off of a skate, a stick, a post and the goalie’s butt. But I’m guessing, most coaches don’t conclude that they did everything right on most goals against. There’s coaching to do here as well. Something broke down in our process and something bad happened.

Why do you have faith in a process? Because you have some empirical or theoretical evidence that the process leads to intended results. But then you also need to see the process working, i.e., results. Your faith has to be incredibly strong for you to continue doing things in exactly the same way despite a streak of unfavorable results. Poker gives you a long time horizon. People can be disciplined as long as they don’t go broke. Pro sports has a much shorter time horizon.

Slepyshev may do everything right. If he doesn’t score above AHL-replacement player level, he’ll have a short career. He can blame his luck. He can also question whether he was maybe doing everything right but just not doing the right things. Process matters because it leads to results. If it doesn’t lead to results, then it better lead to results soon. If not, at the NHL level, someone is out of a job. That’s just how it is.

As for only controlling inputs, there are a certain number of inputs an average player can control. A skilled player can control more. CMD, even more. The ability to make a play is not the same thing as the ability to make the same play the same way every time. So individual players aren’t exactly following the same process. Coaches have to understand this. There isn’t one process.

How do NHL coaches respond to losing? And how do they respond to winning? Are they more open to change when they’re losing or when they’re winning? Do they question themselves more and are they questioned more when they’re losing or when they’re winning?

Again, the reason we focus on outcomes is because outcomes contain information. In the long run, in multiple trials, luck evens out. Outcomes follow inputs (innate ability, cohesive team play). In the short run, there’s luck involved. But the long run is built up from a bunch of short runs. And most of the short runs have to follow the long run direction. So, in the present, the short run contains information as well.

Munny

LT,

In place of the Fenwick percentage number, which is already built into the FF and FA numbers you provide, maybe a year-to-date Fenwick number to contrast with the in-game number?

Or a high danger number of some sort would be sweet instead too… for eg HDSC against for Dmen and HDSC For for Forwards.

But really, whatever turns your crank. And I have no idea what is out there I terms of stats for you after the game. I was a bit disappointed that GMoney didn’t have his usual post-game up last night at Nerd Alert.

Gayfish

Hfs. Pittsburgh got smacked.

season not played

I wonder what’s more important process during a game, A well executed chip in to the zone that leads to a cycle in which heavy physical contact is initiated on the d which has a lasting effect but does not generate a meaningful shot metric or creating a shot metric event which unless there is a deflection or a lapse in nhl level goaltending has a very low chance of generating a goal and a far greater chance of leading to a turnover. I would dare to venture those that could be included in the completely subjective high danger chances.

Also, whatever strategy McPhee used to assemble his blue line confounds me. As a believer that a marginal to slightly more than marginal difference in player ability on the blue line is offset by proper side handedness, how does he end up with Miller and Engelland as his only viable right shooting options?

No math, more like commonsense/60

This is all meant to be rhetorical.

godot10

Include only stuff that cannot be easily gleaned from watching the game, the highlights, or browsing the NHL.com or oilers.com game reports.

Including plus/minus or goal differential is redundant and pointless in this context. Watch the game. Watch the hightlghts. Just scan the NHL.com game stats. You get plus/minus or goal differential from all of them.

Side

A running tally of how many times Taylor Hall is mentioned on this website.

Gayfish

Lowetide:
Okay, I’ll open this up to the group.What WOULD you like to see in the table each game? Currently I have

Boxcars
Fenwick
Fenwick percentage
EV TOI
Comments

Which is followed by something like this:

Klefbom—Larsson went 30-9 together Fenwick (all numbers Fenwick and all numbers 5×5), including a tidy 23-5 with the McDavid line and 7-5 with the Nuge. The Swedes played most against Gaudreau-Monahan-Ferland (12-9) and were 18-0 against the rest. Klefbom is going to get a lot of attention this season, on merit. Larsson is less flashy, but very effective. He’s a rocknrolla.

What would you like to see?

I’d like a disclaimer of small sample sizes (even on descriptive statistics)

Jordan

My preference would be to see a number of stats moving from the highest number of events (Corsi) through to the lowest number of events (Goals).

Something like 5×5 TOI, Corsi for/against/percentage, Fenwick for/against/percentage, Dangerous Fenwick for/against/percentage, Goals for/against/percentage

I recognize this is a lot of information, but for my mind, it clearly takes us from the most general of information through to the most specific, and ultimately what wins or loses games.

Ideally, you’d have other stats to look at other aspects of the game, like passing completion tied into interception rates, hits, board battles, and puck retreival rates and others. I don’t beleive we’re there yet, as to the best of my knowledge these activities aren’t being tracked, but they would start to give us metrics for skills other than scoring.

After all, scoring is only one part of the game, right? And while I think most of us would agree it might be the most important, a deeper understanding of those other parts would probably still help enhance our understanding of the game as a whole, no?

JD_Wry

asshairs/in2

VOR

Georges,

As a coach you have nothing but process. You can’t control outcomes and neither can your athletes. You can only control inputs. More to the point if you start trying to control outcomes almost invariably you screw up the quality or quantity of your inputs.

“The archer ceases to be conscious of himself as the one who is engaged in hitting the bull’s-eye which confronts him. This state of unconscious is realized only when, completely empty and rid of the self, he becomes one with the perfecting of his technical skill, though there is in it something of a quite different order which cannot be attained by any progressive study of the art.” Eugen Herrigel.

Outcomes in hockey have a lot to do, as I have explained previously, with luck. This creates new problems (beyond how destructive goal orientation is to performance excellence) for any coach who tries to tie outcome to process as a feedback loop. The player messes up but the team scores by fluke- you don’t want the play duplicating that behaviour because over time bad practices will become bad outcomes. The player does everything right but his team gives up a goal because of bad luck – you don’t want them to give up doing everything right because over time good practices will produce good outcomes.

So why would we as fans want to be obsessed with outcomes when what the athletes and coaches can control is inputs and inputs are the only fair way of evaluating their play or their coaching?

Lucinius

Hawks smoke the Pens 10-1. Geeez. Ugly, ugly game from the Pens.

jonrmcleod

Number of comments relative to the previous season’s gameday post.

Jethro Tull

Lowetide:
Okay, I’ll open this up to the group

Errr, what are the most common last words used by a blogger, Alex?

Jethro Tull

Kgs of wood chopped/Liters of water carried. Times that by VO2 max/donuts not eaten (inverse broccoli function).

JD_Wry

Lowetide: What would you like to see?

Nailsstraightened/60

treevojo

Lowetide:
Okay, I’ll open this up to the group.What WOULD you like to see in the table each game? Currently I have

Boxcars
Fenwick
Fenwick percentage
EV TOI
Comments

Which is followed by something like this:

Klefbom—Larsson went 30-9 together Fenwick (all numbers Fenwick and all numbers 5×5), including a tidy 23-5 with the McDavid line and 7-5 with the Nuge. The Swedes played most against Gaudreau-Monahan-Ferland (12-9) and were 18-0 against the rest. Klefbom is going to get a lot of attention this season, on merit. Larsson is less flashy, but very effective. He’s a rocknrolla.

What would you like to see?

I would like to see a true open hole shot chart for each forward, dman and goalie.

HT Joe

Lowetide:
Okay, I’ll open this up to the group.What WOULD you like to see in the table each game? Currently I have

Boxcars
Fenwick
Fenwick percentage
EV TOI
Comments

What would you like to see?

I used to like Eyeglow/60, but I thought the group had moved onto the more critical AssHair/60. Either of those would be great. Please and thanks.

Chachi

Lowetide:
Okay, I’ll open this up to the group.What WOULD you like to see in the table each game? Currently I have

Boxcars
Fenwick
Fenwick percentage
EV TOI
Comments

Which is followed by something like this:

Klefbom—Larsson went 30-9 together Fenwick (all numbers Fenwick and all numbers 5×5), including a tidy 23-5 with the McDavid line and 7-5 with the Nuge. The Swedes played most against Gaudreau-Monahan-Ferland (12-9) and were 18-0 against the rest. Klefbom is going to get a lot of attention this season, on merit. Larsson is less flashy, but very effective. He’s a rocknrolla.

What would you like to see?

Life is short. Do what makes you happy.

gogliano

I’m on board for including Goal Differential. For the reasons Georges cites. It’s nice to have the the ticker that decides the games alongside the other stats.

Georges

Woodguy v2.0:
Georges,

5v5 GF-GA is really, really important. It’s what happened. Here’s where winning happened, here’s where losing happened. Without that context, Fenwick is just Fenwick. It’s not winning and losing. But when you include GF-GA, you can get into did a player win or lose in a way we can be happy with, i.e., how did possession look?
The way you’re doing it, you’re setting winning at Fenwick as the goal. That is not the goal. Scoring more goals is the goal. And you’re leaving out score effects which takes you even further from what happened. So we can expect recaps where Russell (as an example) will get caved on Fenwick (and will get called out for it) in a game that we win by 2 and the Russell pairing comes out even (which won’t get mentioned in the present format,).


You seem committed to the idea we can understand the game through the numbers we record in the course of a game. 5v5 goals are the most important numbers on the record for tracking team and player performance.
Hoping you’ll respond. I’ll drop it if you don’t.

You know how volatile goals are in small samples and 1 game is uber small.

Very often the score doesn’t tell story of the game.

Fenwick describes the game better than corsi, Dangerous Fenwick (or xGF) describe the game better than Fenwick.We’ll have that available on a game level soon.

NHL coaches talk “process, process, process”that’s “winning or losing in way he can be happy with”

McLellan said after the game “last year we outscored our mistakes, tonight we didn’t make any”

So he wasn’t happy with the way they were winning earlier last year, even though they were winning.

Process.

The shot metrics describes the process waaaaaaaaaaaaaaaaaaay better than the score.

You can be interested in the final score and the process to get there equally, or one more than another, but its an individual’s choice on how they consume and enjoy hockey.

I like when LT posts multiple stats.

I bet many different people hook into the ones they prefer and ignore the rest.

That’s just fine.

To only post goals is huge mistake though.Doesn’t tell the story well and goals are most often the most at conflict with the eye test.

Also, most of LT’s readers are interesting in the process, in the score metrics etc.

It’s like we’ve never met before. You keep trying to educate me on things I think you think I know nothing about. It’s fun.

“Small samples and volatile” aren’t relevant to describing what happened. Describing is different from inferring or predicting. Goals are what happened. Fenwick is a bunch of things that happened that didn’t lead to goals and a few that did.

You say that Fenwick describes the game better than Corsi. Why? How did you arrive at this? And then why is DFF even better at describing than Fenwick? DFF is pretty obscure and has a trust me, I know what I’m doing feel to it. It’s not simple. It’s not in the open. To my understanding, it (along with Manny’s xGF) can’t be recreated. So it can’t be tested. You’re an advocate but I’m not sure you yourself know or understand all the details of the calculations G is using. How do you know he did them right? (If you think the question is offside or impertinent, you’ve never built a calculating/reporting system before.) Why not shots? Everyone understands shots. Nothing to calculate there. I can check shots on a bunch of sites to see if a number is right. How many sites can I check DFF on?

Not sure why you think I asked LT to only include goals. Maybe read my post that suggested replacing the FF% column in his table. Having FF-FA and GF-GA in the same table would help connect process and results. I mentioned winning or losing in a way we can be happy with, but you must have missed that too. Show me coaches with bad records who are happy with their process. (I already know about Eakins.) Show me players who are committed to a process that doesn’t produce results. Results are the feedback mechanism for process. Follow process. Track process. Display results.

By the way, what’s the relationship between winning on 5v5 Fenwick and winning on 5v5 GF%? Do you think the question is important? If so, how would you go about trying to suss out an answer?

Getting back to DFF… I’d love for G to drop by. I’ve looked at his methodology and I remember having lots of questions. As an example, what data does G use to calculate the probabilities for different types of shots? Does he then use that same data to calculate DFF? Some of my first posts on here were on SF% and PDO as two independent factors that correlate with GF%. I mentioned that any shot-based measure that has a higher correlation with GF% would have to take some information from PDO. I can go back and look at G’s methodology post again. I know this kind of thing takes a lot of thought and work.

rickithebear

WG:

1. CA60 is established by failings of offence. Weather nz can be trapped and blue line can be presed.
2. The puck is directed at the net. ga/corsi against has highest success rate of finding top 60 hd dmen.
3. X,y establishes the corsi hd density.
4. hD save % mean is established.
5. Goalies Save % performance can be measured versus the mean.
6. Goslies game by game performance ca be extrapolated versus avg graph shot density/ga
7. Elite open corsi supression starts. (Blocks, misses, closed shots)
8. The best supression dmen are identified.
9. Supression can be a huge affect on ga.
10. A true open hole shot chart for each forward, dman and goalie can be created.

OmJo

Do you think Pittsburgh is feeling fatigued after back to back Stanley’s?

8-1 Blackhawks with 5 minutes to go in the SECOND period.

BlueNoteNorth

LT

Your current stats layout works for me. The highlights in the last column provide interest as well. The follow-up summary of pairing and lines completes the picture.

I say stay the course.

digger50

Lowetide:
Okay, I’ll open this up to the group.What WOULD you like to see in the table each game? Currently I have

Boxcars
Fenwick
Fenwick percentage
EV TOI
Comments

Which is followed by something like this:

Klefbom—Larsson went 30-9 together Fenwick (all numbers Fenwick and all numbers 5×5), including a tidy 23-5 with the McDavid line and 7-5 with the Nuge. The Swedes played most against Gaudreau-Monahan-Ferland (12-9) and were 18-0 against the rest. Klefbom is going to get a lot of attention this season, on merit. Larsson is less flashy, but very effective. He’s a rocknrolla.

What would you like to see?

I liked to see the LT. version of the sunshine girl

In regards to the hockey, I enjoy the descriptors more and would prefer detailed stats every 6-8 games when they can tell us more.

8-1 for Chicago tonight. Wow.

OmJo

Off topic, but…

https://twitter.com/HKY_Tweets/status/916112452886155264

Nail Yakupov with a nice spin-o-rama pass to Duchene https://t.co/MckGsgeggh

And…

https://twitter.com/SCTrojan5280/status/916091833599524864

Matt Duchene buries a rebound after Yakupov uses great vision to find Nemeth trailing down the seam https://t.co/kKMaWPCbE2 https://t.co/Xv7PMcGdIm

Duchapov seems to be looking good so far. Like a poor mans Conyak, only it might last more than 10 games this season ‘just because.’

As an aside, has there ever been a more polarizing player than Yakupov. He seems to split up the fan base of every team he plays for into one group of die hard supporters and another group of die hard haters (for lack of a better word). Oilers AND Blues fans are still following his career, hoping he finds success lol.

Pescador

bendelson:
Apropos of nothing…

I saw RTB at an Oiler game awhile back.I wanted to approach and try to get a photo, then saw he was with cubs, so naturally thought better of it.

When Oil country meets Bear country:
https://youtu.be/R7bA37gE8Ko

wheatnoil

theDjdj: So if we trade for him the league investigates before its cleared? I swear they make decisions just to keep lawyers employed.

I suspect if the Oilers trade for Reinhart, the NHL won’t do anything unless some other club complains (which they won’t because, who cares, it’s Reinhart). Then they’ll have some internal, black-box investigation and the trade will stand.

That said, I don’t think Edmonton trades for Reinhart. Miller is the obvious target but I don’t see Vegas trading Miller. So my guess is nothing happens.

(Braces for Sbisa trade)

Professor Q

Yeti: Last time I checked, pretty much no-one liked things shoved down their throat. Just spitballing here, but maybe that’s part of the problem?

Except sword and fire swallowers.

dessert1111

I like the stats you use LT. If you wanted to include an extra one you could get rid of the Fenwick percentage and just have the raw numbers. Thanks for all that you do.