Do Teams Wait Too Long to Pull the Goalie?

[TheDevilMadeMe]

Perhaps. But I sincerely doubt the attacking team will put out 3rd and 4th liners. It's not as if they'll be rolling lines.

1st and 2nd line players will dominate the ice for the attacking team.

With 4 forwards and 2 defensemen:
1st wave = 3 first liners + 1 second liner
2nd wave = 2 second liners + 2 third liners

With 5 forwards and 1 defenseman on the 1st wave:
1st wave = 3 first liners + 2 second liners
2nd wave = 1 second liner + 3 third liners + 1 4th liner or a 2nd defenseman

*I've seen the Devils go with 6 forwards back when the skill on their blueline was pitiful, but it was never very effective

 

[Czech Your Math]

If the GK is pulled around 2 minutes, a player or two might be double shifted (similar to players playing an entire power play).

If the GK is pulled around 3 minutes, the top unit would be sent out... if/when there's a stoppage in play, the GK could be put back in with lesser players... then pull goalie again as one normally would.

As far as the line change w/o a goalie, just rotate players to change every few seconds, not all at once. If the team can't maintain possession enough to do this, with an extra skater, then they probably aren't going to win anyway. Also, remember the defense will have to make line change too, which could result in a good opportunity with an extra skater.

 

[Feed Me A Stray Cat]

With 4 forwards and 2 defensemen:
1st wave = 3 first liners + 1 second liner
2nd wave = 2 second liners + 2 third liners

With 5 forwards and 1 defenseman on the 1st wave:
1st wave = 3 first liners + 2 second liners
2nd wave = 1 second liner + 3 third liners + 1 4th liner or a 2nd defenseman

*I've seen the Devils go with 6 forwards back when the skill on their blueline was pitiful, but it was never very effective

I don't think those line structures are very optimal. Assuming two defenseman, something sensible would be:

First wave: 2 first liners + 1 second liner + 1 third liner (presumably a guy who can make space down low)
Second wave: 1 first liner + 2 second liners + 1 third liner

And if the team has puck possession the entire time, some players might even be able to stay out there the whole time.

Using a team like the Devils, this would equate to:

1st: Kovalchuk, Zajac, Zubrus, Sykora + Greene, Zidlicky

2nd: Henrique, Elias, Clarkson, Josefson + Larsson, Fayne

There's no significant drop off there.

 

[Feed Me A Stray Cat]

If the GK is pulled around 2 minutes, a player or two might be double shifted (similar to players playing an entire power play).

If the GK is pulled around 3 minutes, the top unit would be sent out... if/when there's a stoppage in play, the GK could be put back in with lesser players... then pull goalie again as one normally would.

As far as the line change w/o a goalie, just rotate players to change every few seconds, not all at once. If the team can't maintain possession enough to do this, with an extra skater, then they probably aren't going to win anyway. Also, remember the defense will have to make line change too, which could result in a good opportunity with an extra skater.

That's a good point too. Nothing says the goalie has to stay pulled the entire time.

 

[TheDevilMadeMe]

I don't think those line structures are very optimal. Assuming two defenseman, something sensible would be:

First wave: 2 first liners + 1 second liner + 1 third liner (presumably a guy who can make space down low)
Second wave: 1 first liner + 2 second liners + 1 third liner

And if the team has puck possession the entire time, some players might even be able to stay out there the whole time.

Using a team like the Devils, this would equate to:

1st: Kovalchuk, Zajac, Zubrus, Sykora + Greene, Zidlicky

2nd: Henrique, Elias, Clarkson, Josefson + Larsson, Fayne

There's no significant drop off there.

I agree that if you're going to pull the goalie for 3 minutes, it would be ideal to spread the wealth more.

But then you can't assume that the 6-on-5 scoring rates we observe when teams do put out all their best players for 1 minute can be extrapolated over 2-3 minutes!

 

[Canadiens1958]

Practicality

If the GK is pulled around 2 minutes, a player or two might be double shifted (similar to players playing an entire power play).

If the GK is pulled around 3 minutes, the top unit would be sent out... if/when there's a stoppage in play, the GK could be put back in with lesser players... then pull goalie again as one normally would.

As far as the line change w/o a goalie, just rotate players to change every few seconds, not all at once. If the team can't maintain possession enough to do this, with an extra skater, then they probably aren't going to win anyway. Also, remember the defense will have to make line change too, which could result in a good opportunity with an extra skater.

The offensive teams bench is located outside the defensive zone so your rotation scenario is not practical and is an invitation for "too many men" penalty calls.

Defensive changes happen when the puck is safe, down ice away from the defenders net.

 

[Iain Fyffe]

2011-12 NHL season PPG team average 47. Atlantic Division top 4 forwards produced the following PPG totals per team:

NYR top 4 forwards scored 34 of the teams 44 PPG, Pitt - 41 of 57 PPG, Phil - 41 of 66 PPG, NJ - 33 of 46 PPG, NYI - 32 of 45 PPG.

You'll probably be shocked to discover that for four of these five teams, the forwards ranked 1st-4th in PP goals also ranked 1st-4th in PP minutes by forwards on their teams. Philly's the sole exception, whose 1st-4th PP goal-scorers ranked 1st, 2nd, 3rd and 5th in PP time by forwards.

So what you're saying is: the players who get the most PP time score the most PP goals. I think that's fairly self-evident.

But of course, this has nothing to do with my question. I didn't say "do you have any evidence that there's a dropoff in offence from the first unit to the second". I said "do you have any evidence that the negative effect of 'disrupting your lines' outweighs the positive effect of a longer man-advantage situation, when trailing by a goal?"

 

[Canadiens1958]

Meritocracy

You'll probably be shocked to discover that for four of these five teams, the forwards ranked 1st-4th in PP goals also ranked 1st-4th in PP minutes by forwards on their teams. Philly's the sole exception, whose 1st-4th PP goal-scorers ranked 1st, 2nd, 3rd and 5th in PP time by forwards.

So what you're saying is: the players who get the most PP time score the most PP goals. I think that's fairly self-evident.

But of course, this has nothing to do with my question. I didn't say "do you have any evidence that there's a dropoff in offence from the first unit to the second". I said "do you have any evidence that the negative effect of 'disrupting your lines' outweighs the positive effect of a longer man-advantage situation, when trailing by a goal?"

Not shocked at all since NHL hockey is a meritocracy. No advantage to having a team's least qualified performers on the ice at the most critical time in unfamiliar circumstances.

Your question has nothing to do with the situation being discussed.
Disrupting lines was a Fred Shero strategy with the pre 1976 Flyers. Other coaches had their own variations going back to the early O6 era. With off setting penalties the opposition had to play makeshift lines against the Flyers who were used to playing without their role players, thus gaining an advantage.

Same basic strategy applies to a 6 on 5 situations. Time-outs in hockey are a recent phenomena. A team pulling their goalie wants their elite offensive top 6 skaters to strike quickly before the defensive team can adjust their 5 skater defensive set(time-out) or replace tired skaters post icing.

 

[Yossarian54]

It is estimated that NHL teams would on average gain one point over an 82-game season if they simply decided to pull their goalies at the 3:00 mark instead of 1:00, regardless of situation.

This is where it falls down for me. That's simply not a large enough effect to consider it as an effective strategy. Natural variability would mask any noticeable effect, and any incorrect assumptions in methodology on behalf of the analysis could potentially mean the real effect is, in fact, negative.

 

[RandV]

With 4 forwards and 2 defensemen:
1st wave = 3 first liners + 1 second liner
2nd wave = 2 second liners + 2 third liners

With 5 forwards and 1 defenseman on the 1st wave:
1st wave = 3 first liners + 2 second liners
2nd wave = 1 second liner + 3 third liners + 1 4th liner or a 2nd defenseman

*I've seen the Devils go with 6 forwards back when the skill on their blueline was pitiful, but it was never very effective

It's a valid point that teams use the last minute because they want their best six out there for the push, but if a team were to experiment with pulling their goalie early they don't have to stick to the above formula. While taking your 4 best skilled forwards will give you the best chance to score in a single shift, taking your 3 best skilled forwards and adding a utility player to forecheck/screen/crash the net could still be effective.

Using last years Vancouver team as an example it would be the standard:

Sedin-Sedin-Burrows-Kesler-Edler-Salo

for 1-1.5 minutes vs

Sedin-Sedin-Burrows-Hansen-Edler-Salo
Booth-Kesler-Higgins-Raymond-Bieksa-Hamhuis

for 2-3 minutes.

Also keep in mind in the standard situation if you want your 1st line out for the last minute then you're going to want your 2nd line out for the second last minute, but if you're taking your best 2nd line forward to stack the 1st line in the last minute then you're weakening your chances for that second last shift, goalie pulled or not. Also, prolonged pressure in the offensive zone is a bit easier on the attacking teams endurance. If you have a stud forward or dman like Ilya Kovalchuk or Shea Weber you can probably leave them out there for both shifts if there's sustained pressure.

 

[Iain Fyffe]

Your question has nothing to do with the situation being discussed.

Really? The question is whether teams should be pulling their goalies ealier than what they're doing now.

You seem to be saying no, because the timing is optimized. This means you believe teams are getting the maximum net benefit from the timing currently used.

I'm asking if you have any evidence that this is true. That seems pretty bloody relevant to me.

If the "second unit" scores a significant amount less than the first, it can still be worthwhile if the second unit's rate is sufficiently high to merit the tactic. Can you offer anything to demonstrate that it's not?

 

[Iain Fyffe]

Not shocked at all since NHL hockey is a meritocracy. No advantage to having a team's least qualified performers on the ice at the most critical time in unfamiliar circumstances.

No, you see my point was that you used counting stats to argue that certain players score at a greater rate. That's not valid. You need to use rate stats for that.

Now, it is surely true that the first PP unit on a team will score at a greater rate than the second (the evidence for which would be the rate they score on the PP, not the total amount they score on the PP). But that by itself is not evidence that teams currently pull the goalie at the right time. As I said in the last post, the second unit might still score enough for it to be worthwhile. It doesn't have to be as much as the first unit, it just has to meet a certain threshold.

 

[TheDevilMadeMe]

It's a valid point that teams use the last minute because they want their best six out there for the push, but if a team were to experiment with pulling their goalie early they don't have to stick to the above formula. While taking your 4 best skilled forwards will give you the best chance to score in a single shift, taking your 3 best skilled forwards and adding a utility player to forecheck/screen/crash the net could still be effective.

Using last years Vancouver team as an example it would be the standard:

Sedin-Sedin-Burrows-Kesler-Edler-Salo

for 1-1.5 minutes vs

Sedin-Sedin-Burrows-Hansen-Edler-Salo
Booth-Kesler-Higgins-Raymond-Bieksa-Hamhuis

for 2-3 minutes.

Also keep in mind in the standard situation if you want your 1st line out for the last minute then you're going to want your 2nd line out for the second last minute, but if you're taking your best 2nd line forward to stack the 1st line in the last minute then you're weakening your chances for that second last shift, goalie pulled or not. Also, prolonged pressure in the offensive zone is a bit easier on the attacking teams endurance. If you have a stud forward or dman like Ilya Kovalchuk or Shea Weber you can probably leave them out there for both shifts if there's sustained pressure.

This is all true, but doesn't answer the objection to the study that you can't just extrapolate scoring rates from a minute of your best players over 2-3 minutes of 6 on 5

 

[Canadiens1958]

Your Point

No, you see my point was that you used counting stats to argue that certain players score at a greater rate. That's not valid. You need to use rate stats for that.

Now, it is surely true that the first PP unit on a team will score at a greater rate than the second (the evidence for which would be the rate they score on the PP, not the total amount they score on the PP). But that by itself is not evidence that teams currently pull the goalie at the right time. As I said in the last post, the second unit might still score enough for it to be worthwhile. It doesn't have to be as much as the first unit, it just has to meet a certain threshold.

Then it is up to you to define the worthwhile level and the certain threshold that you conjure up.

As is, the team originating the 6 on 5 situation does so in a fashion to optimize benefits while minimizing risks. NHL teams allow < 30 SOG per game or < 1 every 2 minutes. Giving up an empty net goals requires 1 SOG by the defending team so the offensive team wishes to minimize its risk when playing its 6 best offensive players.

 

[tarheelhockey]

Tangent: it infuriates me to see a team gain an offensive-zone faceoff with less than 2 seconds left in any period and not pull the goalie.

 

[Hivemind]

Rather than the pull the goalie, I think teams would be better off leaving him in, playing with 4 up front and 1 D, who would pinch to make a play. You could do that for a good 3 minutes and at least have a chance to go the whole the 3 without allowing a goal.

I agree.

I think they pull the goalies too early if anything.

You're entitled to believe what you want, but this is the "By the Numbers" board. Please cite some sort of statistical evidence to support your claim.

 

[Feed Me A Stray Cat]

Then it is up to you to define the worthwhile level and the certain threshold that you conjure up.

As is, the team originating the 6 on 5 situation does so in a fashion to optimize benefits while minimizing risks. NHL teams allow < 30 SOG per game or < 1 every 2 minutes. Giving up an empty net goals requires 1 SOG by the defending team so the offensive team wishes to minimize its risk when playing its 6 best offensive players.

How are average shots per minute relevant here? On 6-5 situations the defending team's shot rate goes down significantly.

 

[Iain Fyffe]

Then it is up to you to define the worthwhile level and the certain threshold that you conjure up.

In the context of this discussion, you have stated that what is currently done is optimal. I'm saying it might not be, but that we might be able to work it out if it is.

If you want to assert something, you need to provide evidence for it. You've made an assertion, but have provided no evidence. I will make no assertion at this point, because I haven't studied the issue yet.

As is, the team originating the 6 on 5 situation does so in a fashion to optimize benefits while minimizing risks.

This is the goal. What we're discussing in this thread is whether or not this goal is being achieved with mainstream strategy.

This is your assertion, right here: "the team originating the 6 on 5 situation does so in a fashion to optimize benefits while minimizing risks". I'd say that's the intent, but you're phrasing it as if this is what is actually happening, although you haven't provided any evidence to demonstrate that the intent is actually being achieved.

Tangent: it infuriates me to see a team gain an offensive-zone faceoff with less than 2 seconds left in any period and not pull the goalie.

Agreed. Although I wouldn't say it's infuriating. It's surely a tiny benefit, but you might as well take any little benefit you can get.

 

[Iain Fyffe]

How are average shots per minute relevant here? On 6-5 situations the defending team's shot rate goes down significantly.

Spot on, and the other important consideration is the cost/benefit. Allowing a goal, when you're already down with little time left, has a small marginal cost. That's why pulling the goalie makes sense at all.

 

[Choralone]

When Andy Murray coached the Kings, he was very aggressive pulling the goalie if his team was trailing. I'm pretty sure he pulled the goalie once with close to 5 minutes remaining. He may have continued the practice when he later coached in St. Louis.

You might want to check games he coached for relevant - albeit limited - data.

 

[Canadiens1958]

No Benefit

Tangent: it infuriates me to see a team gain an offensive-zone faceoff with less than 2 seconds left in any period and not pull the goalie.

No benefit is gained from the sixth skater. Regardless of how talented the sixth skater is he will either clutter the available shooting lanes or be out of the way of the shooting lanes and of no benefit.

The move that used to work was spreading the shooters to the wide side, getting the goalie to anticipate wide then have the offensive center snap a short side shot at the net of the draw. Easy to counter.

 

[Iain Fyffe]

No benefit is gained from the sixth skater. Regardless of how talented the sixth skater is he will either clutter the available shooting lanes or be out of the way of the shooting lanes and of no benefit.

I believe you mean "little benefit", which when compared to the cost (effectively nil) is still a benefit.

 

[Canadiens1958]

Negative Benefit

I believe you mean "little benefit", which when compared to the cost (effectively nil) is still a benefit.

No cost or benefit if the sixrh skater is positioned away from the shooting lanes. Negative benefit if the sixth skater clutters or gets collapsed into a shooting lane.

The goalie does not clutter nor will he get collapsed into the shooting lanes.

 

[MadLuke]

No cost or benefit if the sixrh skater is positioned away from the shooting lanes. Negative benefit if the sixth skater clutters or gets collapsed into a shooting lane.

The goalie does not clutter nor will he get collapsed into the shooting lanes.

Is this true for the fith skater, wath is the ideal number of skater ?

 

[Canadiens1958]

Coincidental Minors

Is this true for the fith skater, wath is the ideal number of skater ?

When coincidental minors had to be served teams like the Oilers with Gretzky etc preferred 5 on 4 or 4 on 3 late to tie or win. Likewise mid seventies Flyers.

Look at the regular season 5 minute OT after regulation ties. 4 on 4 creates more open ice and wider/easier shooting lanes.

6 on 5, the ideal situation is getting the puck to the open skater. The more skaters on the ice the harder it is to find open ice and the open skater.