On Ice Goals Against 5x5 chart

[rakish]

I haven't been watching the games, though I've seem most of the highlights, so I haven't been chatty too much about the Sabres, but this week I made some charts, the data's a couple games behind.  The chart below displays goals against/60 5x5 vs. the difficulty of opponent (also 5x5), trying to add some contextuality (difficulty of opponent) to a non-contextual stat (goals against).  I measure the difficulty of opponent by how many minutes of PP ice time the opponent forwards average. So if you are on the ice against Marchand (who might get 5 min of PP per game), and Pasternk (lets say 4 min), and Bergeron (lets say 3 min), for that time period, your opponent forward's PP ice time would be 12 minutes.

Dea, Smith, and Borgan were removed to make the chart prettier.

So a few thoughts from the chart.

I expected the chart to be a line of dots from top right to bottom left. That the players more often playing against Marchand, would have the most goals scored against them, but the chart, more or less, is a line of dots from top left to bottom right. Meaning that even though Bryson play against the toughest assignments, he's still on the ice for fewer goals than most everyone else. Even though Dahlen is on the ice against the weakest of competition, he still is on the ice for the most goals.

The Sabres have given Risto the toughest assignment for years, though this year Bryson gets the toughest assignments.

It's not Eakin killing the Sabres, it's Mittlestadt.

Why was Montour traded?

Lazar doesn't have the usage I thought he would have.

I would have to count empty net goals against by hand.

[Marvin]

When there are more data at the end of the season, I would like to see this again, but split between the coaches.  Numerous players are having radically different seasons than before, Mittlestadt and Ristolainen being 2 of them.

[Taro T]

3 hours ago, rakish said:

I haven't been watching the games, though I've seem most of the highlights, so I haven't been chatty too much about the Sabres, but this week I made some charts, the data's a couple games behind.  The chart below displays goals against/60 5x5 vs. the difficulty of opponent (also 5x5), trying to add some contextuality (difficulty of opponent) to a non-contextual stat (goals against).  I measure the difficulty of opponent by how many minutes of PP ice time the opponent forwards average. So if you are on the ice against Marchand (who might get 5 min of PP per game), and Pasternk (lets say 4 min), and Bergeron (lets say 3 min), for that time period, your opponent forward's PP ice time would be 12 minutes.

Dea, Smith, and Borgan were removed to make the chart prettier.

So a few thoughts from the chart.

I expected the chart to be a line of dots from top right to bottom left. That the players more often playing against Marchand, would have the most goals scored against them, but the chart, more or less, is a line of dots from top left to bottom right. Meaning that even though Bryson play against the toughest assignments, he's still on the ice for fewer goals than most everyone else. Even though Dahlen is on the ice against the weakest of competition, he still is on the ice for the most goals.

The Sabres have given Risto the toughest assignment for years, though this year Bryson gets the toughest assignments.

It's not Eakin killing the Sabres, it's Mittlestadt.

Why was Montour traded?

Lazar doesn't have the usage I thought he would have.

I would have to count empty net goals against by hand.

Are EN goals removed from your data set?

The Sabres seemed to give up a ton of them prior to the losing streak ending, and the worst offenders have gotten a lot of time out with the EN.  Even though it is only ~1 minute/ game that the goalie has been pulled that they're on the ice for, that might really skew the results.

[rakish]

6 hours ago, rakish said:

I would have to count empty net goals against by hand.

 

 

[Taro T]

1 hour ago, rakish said:

 

 

Not looking to create work for you & do appreciate the effort you put into this.

But that said, am curious how much EN GA affect the metrics of guys like Dahlin who are always on the ice in those situations, especially considering how often the Sabres have given up ENGs this season.

[rakish]

        goal against team: team 9,20,23,68,4,26 opponent 26,21,62,3,33,30
goal against team: team 13,20,24,26,62,35 opponent 92,43,13,3,33,41
goal against team: team 9,23,4,26,33,34 opponent 13,93,86,56,61,37
        goal against team: team 9,23,68,4,26,62 opponent 13,93,86,6,9,37
goal against team: team 12,72,68,26,33,35 opponent 19,43,13,4,74,41
goal against team: team 20,72,68,26,62,35 opponent 20,77,13,2,33,41
goal against team: team 9,23,4,26,33,35 opponent 26,21,62,2,33,41
goal against team: team 9,23,4,26,33,35 opponent 20,93,89,8,79,40
goal against team: team 9,23,4,26,33,35 opponent 16,43,48,77,79,40
goal against team: team 13,20,21,26,33,35 opponent 20,59,14,24,70,41
goal against team: team 15,27,37,26,33,40 opponent 86,11,63,24,28,1
      goal against team: team 9,37,68,4,26,55 opponent 17,37,44,28,70,1
goal against team: team 9,23,53,26,33,35 opponent 27,44,26,2,3,40
goal against team: team 12,23,4,26,33,40 opponent 44,26,18,2,6,30
      goal against team: team 9,23,68,4,26,33 opponent 12,27,44,3,24,30
goal against team: team 12,72,68,26,62,35 opponent 20,14,73,2,33,41
goal against team: team 20,21,4,26,33,35 opponent 53,15,17,3,4,40
goal against team: team 37,53,4,26,33,40 opponent 14,28,25,9,53,37
goal against team: team 9,23,68,26,55,34 opponent 14,25,86,9,53,79
goal against team: team 12,21,4,26,55,40 opponent 93,13,89,42,79,31
goal against team: team 20,24,4,26,55,34 opponent 7,13,27,4,8,30
goal against team: team 9,23,68,26,55,34 opponent 53,15,17,3,6,30
goal against team: team 9,24,68,26,55,34 opponent 12,29,18,3,4,30
goal against team: team 15,27,53,26,55,34 opponent 7,13,27,4,8,30
goal against team: team 9,24,26,55,62,34 opponent 53,15,17,3,4,30
goal against team: team 9,24,4,26,33,40 opponent 12,29,18,3,6,40
goal against team: team 9,24,4,26,33,40 opponent 53,15,17,3,6,40
goal against team: team 9,24,4,26,33,34 opponent 44,53,15,3,6,30
goal against team: team 23,24,4,26,62,34 opponent 11,13,21,6,61,37
goal against team: team 23,37,4,26,55,34 opponent 7,14,57,2,28,35
          goal against team: team 12,23,68,4,26,55 opponent 9,71,17,8,58,35
          goal against team: team 12,23,68,4,26,62 opponent 59,87,17,5,58,1
goal against team: team 15,27,26,62,78,40 opponent 92,10,8,73,2,41
goal against team: team 12,23,68,26,33,40 opponent 19,77,8,3,9,41
goal against team: team 12,23,68,26,33,40 opponent 92,10,13,3,33,41
goal against team: team 13,20,21,26,62,40 opponent 12,46,21,67,75,41
         goal against team: team 12,23,68,4,26,33 opponent 10,16,24,23,55,71
goal against team: team 12,15,68,26,33,31 opponent 9,14,18,4,5,35
goal against team: team 27,37,68,26,33,31 opponent 59,87,17,8,58,35
goal against team: team 27,37,53,10,26,31 opponent 11,57,67,4,5,1
        goal against team: team 15,37,53,68,10,26 opponent 59,87,17,4,6,1
goal against team: team 17,23,4,10,26,35 opponent 13,37,88,10,48,73
goal against team: team 15,23,68,10,26,35 opponent 11,13,25,5,6,37
goal against team: team 15,53,74,10,26,31 opponent 10,16,43,8,42,31
goal against team: team 37,72,4,10,26,35 opponent 10,16,24,23,55,31

 

These are Dahlen's 45 goals against 5x5. I tabbed the 8 I think are empty net, but I may have missed some. So his 4.6 can be reduced to maybe 3.8 doing rough math in my head, since, as you say, he's not on the ice for a long time during empty net.

[rakish]

So plugging the 3.8 back into the chart, is still not good.

[GASabresIUFAN]

Rakish, sorry but I don't think your chart passes the eye test. 

[rakish]

1 minute ago, GASabresIUFAN said:

Rakish, sorry but I don't think your chart passes the eye test. 

I was think of painting a scene of the 2021 Sabres, the goal light would be on, the goaltender would be flat on the ice, and 26 would be breaking his stick on the crossbar

[rakish]

There are arguments against this type of analysis. @Wildcard doesn't like the difficulty of opponent metric because some forwards on the powerplay aren't that fearsome in 5x5. I think Olofsson would come under this heading for Buffalo. The GED guy argues that in reality there is no difference between who Bryson plays against vs. who Dahlen plays against.

That said, I generally look at the defensemen on the ice each time the puck goes in the net, but I'm drinking way too much to keep track of the forwards. The camera shows the other team celebrating the goal, not so much the skate of shame for the team scored against.

My current theory is that it's the turning over of the puck at each blue line that leads to goals against. The forwards who get scored upon are those who are trying to skate past the defense, which is Mittlestadts game.

[thewookie1]

Something here doesn't seem kosher.

 

I'm unsure as to exactly what but these numbers give a very different look than both the regular analytics we all see and just the basic eye test. 

My only guess is the the whole opponent metric is flawed and thus creating some unusual results. Trying to base skill level 5v5 on PP time is a flawed concept in of itself. @WildCard is certainly on the right path seeing as some players are more specialists than others. How I personally see it is that PP time and 5v5 skill don't really correlate all that well.  For starters most teams have PP units that play 4F-1D thus skewing the numbers from the onset. As such, lower lined players will comparatively receive higher weight that they really are worth. Just picture a roster, 2 PP units usually equal 8F and 2D. 8 Forwards means that you have 2/3 of the 3rd line as well as the full 1st and 2nd lines involved in effect. However this is rarely the case as there is almost always one or two more "defensive" forwards in a top 6 whom aren't as suited for PP play. Specialists are typically stored in the Bot 6 thus creating a abnormality between the 12 Fs on the roster's order and the PP player order. Therefore, players like Eakin will end up playing supposedly far greater opponents when he's really just playing 3 or 4th liners with PP upsides. I highly doubt one person here would say Eakin has been effective this year at anything outside of faceoffs yet this chart makes him out to be doing great comparatively. The other catch in this is how the 1st PP can sometimes play 75% of a Powerplay and rarely does the 1st PP line up with the 1st Line + 1 . 

[rakish]

I've updated the data.

Here are the PP times of forwards, I think there's a good correlation between the offensive skill level and the PP time, at least for the Sabres
 

9 EICHEL3.94
4 HALL3.45
23 REINHART3.38
68 OLOFSSON3.38
37 MITTELSTADT1.99
24 COZENS1.48
12 STAAL1.28
72 THOMPSON1.27
53 SKINNER1.25
21 OKPOSO0.54
25 RUOTSALAINEN0.11
13 RIEDER0.07
74 ASPLUND0.07
27 LAZAR0.06
20 EAKIN0.05
15 SHEAHAN0.05
17 FOGARTY0.05

As to the playing 4 forwards vs. playing 3 forwards. It does affect the goaltenders, since a goaltender will get an entire game against 1 team, the players it should average out, at least past the players who played just a few games.

[rakish]

Boston:

[WildCard]

2 hours ago, thewookie1 said:

Something here doesn't seem kosher.

 

I'm unsure as to exactly what but these numbers give a very different look than both the regular analytics we all see and just the basic eye test. 

My only guess is the the whole opponent metric is flawed and thus creating some unusual results. Trying to base skill level 5v5 on PP time is a flawed concept in of itself. @WildCard is certainly on the right path seeing as some players are more specialists than others. 

Have I been doing this? Tbh I haven't been posting too much about the actual on-ice product this season lol