Jump to content

Playing above expectations


tomh009

Recommended Posts

We've talked before about the Habs playing above expectations this season, so I pulled data from naturalstattrick.com. The interesting view is to actually look at GF/xGF and GA/xGA by coach. For the Habs, since 2017 (regular season only):

  • Julien (253 games): 0.97 and 1.01
  • Ducharme (83): 0.88 and 1.14
  • St-Louis (55): 1.12 and 1.01

Based on this, Ducharme's numbers look quite grim, the Habs couldn't score and couldn't keep the puck out of their own net, even given the injuries. Julien's record pretty much met expectations. St-Louis, though, seems to have found some magic to enable the team to score above expectations.

 

1.12 isn't outrageously high, though, as some other coaches have had similar numbers over larger number of games -- for example, Jon Cooper is at 1.16 over 390 games, Jared Bednar at 1.14 over 388, Peter Laviolette at 1.12 over 363 (well above xGA in both Nashville and Washington)  and Craig Berube at 1.11 over 289.

 

On the defensive side, Barry Trotz is at 0.89 GA/xGA over 370 games ...

 

(Incidentally, Sheldon Keefe is at 1.06 GF/xGF, almost the same as Mike Babcock, but his GA/xGA is 1.07 where Babcock's was 0.96.)

Link to comment
Share on other sites

Play right within my expectations and only going to get better. Tankers be damned! We will go to the dance and Carey will be back for his final run. I believe! He would have a better shot with this defense in front of him than he had 2 years back and we have way more goal scoring. If that team could come up just short what makes you think this team can't?

Link to comment
Share on other sites

1 hour ago, TurdBurglar said:

My biggest issue with how they are playing is they aren’t a playoff team.  So they will finish with a pick in the mid teens in a year when they want to pick where they were projected to pick, top-5.  

 

Will they?  I think it's way too early to make any definitive predictions.  They haven't played well lately and one line can only carry a team so far, especially now with the goaltending starting to struggle.  Are they a last place team?  Probably not at this point but I don't think they're going to just come up short of a Wild Card spot either which is how you get a pick in the mid-teens.

Link to comment
Share on other sites

5 minutes ago, dlbalr said:

 

Will they?  I think it's way too early to make any definitive predictions.  They haven't played well lately and one line can only carry a team so far, especially now with the goaltending starting to struggle.  Are they a last place team?  Probably not at this point but I don't think they're going to just come up short of a Wild Card spot either which is how you get a pick in the mid-teens.

 

Montreal and Columbus seem to have switched places. Heading into last season, the Jackets were expected to be in contention for the top pick and Montreal was hoping Price would be back early enough to get them in wild card contention. Instead, Montreal went off the rails immediately and never got back on, while Columbus surprised and while they were never truly in playoff contention, they scored enough goals to give people a sliver of hope through mid-season, before ultimately finishing in the top half of the non-playoff teams.

 

This year, Montreal was expecting more of last year, while Columbus thought that Johnny Gaudreau would put them into the playoff hunt. Instead, Columbus went off the rails immediately and it is Montreal scoring goals left and right, giving that slight hope of sneaking into the playoffs.  The most likely scenario in my mind is that Montreal does exactly what Columbus did last year and finishes in the top half of the non-playoff teams. 

Link to comment
Share on other sites

1 hour ago, johnnyhasbeen said:

... We will go to the dance and Carey will be back for his final run. I believe! He would have a better shot with this defense in front of him than he had 2 years back and we have way more goal scoring. If that team could come up just short what makes you think this team can't?

 

 

And I'll watch it all ...
image.png

  • Like 1
  • Upvote 1
Link to comment
Share on other sites

I dont know how much a coach can do to have his team have their actual goals for be higher than than their xG.

 

To me, a coach can raise his teams xG by playing a better system, by line matching, by all the things a coach does. A coach can raise the number of quality chances a team produces or lower the number it gives up with good coaching.  Thats the xG and xGF column.

 

With a G:xG ratio.... we should expect volatilty on small samples, but on large samples it should eventually revert to 1:1

 

A team scoring more goals than expected... ie a ratio higher than 1, suggests that they either have Luck on a small sample size, or their shooting talent is elite.  Ie as an individual player, ovechkin consistently scores more goals than his individual xGF suggests he should.  For a Habs example, I bet Cole Caufield will probably be above a 1.0 for his career when all is said and done.

 

On the other end, being below 1.0, like putting up a 0.8 or a 0.9 or something suggests bad luck or that the people taking the shots are not talented shooters at the NHL level. See a guy like Victor Mete.  He just doesnt have an NHL level shot, so when you give him good opportunities, the number of actual goals he scores is lower than the expected goals of all his chances.

 

Good coaching puts the player in a position to succeed.  Talent is needed though to actually finish the play and score the goals.  If you have average shooting talent (for an NHLer) over a large sample size you'll be 1:1, if you are better you can be over 1, and worse, you can be under.

 

On the defensive end, its mostly goaltending.  A strong goalie will make extra saves... ie his Goals saved above expected will be good..  which means a number less than 1 on goals given up.  A bad goalie gives up goals where he should make the save and the ratio on goals against goes above 1

 

Link to comment
Share on other sites

The biggest thing is sample size though.

 

Julien is almost 1:1 on both ends.  Hes got about 3x the games of DD and 5x the games of MSL.  Youd suspect theyd all end up close to 1:1 with more games.  The difference in the coaching isnt the ratio of G:xG.... its the volume of xG they can create and xGA that they can prevent as coaches.

  • Upvote 1
Link to comment
Share on other sites

1 hour ago, Commandant said:

The biggest thing is sample size though.

 

Julien is almost 1:1 on both ends.  Hes got about 3x the games of DD and 5x the games of MSL.  Youd suspect theyd all end up close to 1:1 with more games.  The difference in the coaching isnt the ratio of G:xG.... its the volume of xG they can create and xGA that they can prevent as coaches.

Yes, for the Habs, the Ducharme and St-Laurent samples are small -- but others, like Cooper or Laviolette, have much larger samples and are still well away from 1.00. The xG model is not all-knowing, though, it takes into account maybe 20-25 parameters that are available in the NHL data. But, just like an individual's shooting accuracy or deception skill, there may be things that a coach can teach his team that impact the scoring.

 

Suppose, for example, that the mobile PP that St-Laurent has implemented is more effective (no, I don't have any data on whether it really is), this is not one of the parameters of the xG model, so it could result in the team's scoring being above xGF. Similarly, shot blocking is not included, due to data limitations, so that could impact GA but not xGA.

 

My mini-investigation was really driven by the discussion about the Habs winning more games than predicted by the underlying stats. I may look into this some more, one particularly interesting aspect (for me) is the impact of coaches that have coached multiple teams, and whether that is consistent.

Link to comment
Share on other sites

1 hour ago, tomh009 said:

Yes, for the Habs, the Ducharme and St-Laurent samples are small -- but others, like Cooper or Laviolette, have much larger samples and are still well away from 1.00. The xG model is not all-knowing, though, it takes into account maybe 20-25 parameters that are available in the NHL data. But, just like an individual's shooting accuracy or deception skill, there may be things that a coach can teach his team that impact the scoring.

 

Suppose, for example, that the mobile PP that St-Laurent has implemented is more effective (no, I don't have any data on whether it really is), this is not one of the parameters of the xG model, so it could result in the team's scoring being above xGF. Similarly, shot blocking is not included, due to data limitations, so that could impact GA but not xGA.

 

My mini-investigation was really driven by the discussion about the Habs winning more games than predicted by the underlying stats. I may look into this some more, one particularly interesting aspect (for me) is the impact of coaches that have coached multiple teams, and whether that is consistent.

 

John cooper has coached the Lightning... a team with Kucherov, Stamkos, Point, etc.... etc....

 

As i said in the ovechkin example.  If a team is constantly shooting above their xG over a short sample its probably luck but if the team or player does it consistently, then it shows that the team has a level of elite shooting talent.  I think Tampa certainly qualifies for elite.shooting talent in those three players.

Link to comment
Share on other sites

As for.the PP and shot block.examples.

 

They are in xG.  The mobile pp (if it's better) is likely better because it either creates more volume of.chances or.higher quality chances.  Both improve xG.

 

Shot blocks are also part of xG.  The number of opponents between you and the goalie is a factor of xG.  

Link to comment
Share on other sites

As I understand it, the NHL data doesn't provide the location of the shots that were blocked, only where they were blocked, so it's incomplete. Unless this has changed recently.

 

Mobile PP could just provide better shot locations or more shots. Or it could have additional impact that's not covered by those data points.

 

xG is very useful, but it doesn't know everything. Shooter skill is one thing that it doesn't account for, as you said, but what is skill? Accuracy, choice of shot, speed of release, deception, and others? Also some shooters (Suzuki!) will be more likely to pass than take a low-percentage shot.

 

So, can a coach impact scoring beyond xG? I don't know, but I do want to look into it more because I am intrigued.

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...